[ { "id": "databricks", "name": "Databricks", "company": "Databricks", "region": "global", "tagline": "Unified data and AI platform", "description": "Databricks Lakehouse Platform combines data warehousing with data lakes for unified analytics and AI including foundation model training.", "users": "10K+ companies", "userCount": 10000, "releaseDate": "2013-01", "technicalImpact": 84, "competitors": [ "huggingface", "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry" ], "features": [ "Unified data lakehouse", "Delta Lake", "MLflow", "Mosaic ML" ], "pricing": { "usage_based": "Pay-as-you-go", "enterprise": "Contact sales" }, "link": "https://databricks.com", "apiLink": "https://docs.databricks.com", "aiCoreShare": 73, "scoringRationale": "Databricks is the leading data and AI platform for enterprise ML workflows, with 10K+ enterprise customers and multi-billion-dollar ARR. The Databricks Lakehouse and Mosaic AI stack (MLflow, Unity Catalog, Model Serving, DBRX, Mosaic AI Gateway) represent deep infrastructure control across the enterprise AI development pipeline. Score reflects genuine platform lock-in and enterprise scale. Kept at 84 rather than higher because Databricks-specific AI-product user counts are not separately disclosed from the broader data platform, and the AI layer competes with AWS Bedrock, Azure OpenAI, and Vertex AI for model serving." }, { "id": "fal-ai", "name": "fal.ai", "company": "fal.ai", "region": "global", "tagline": "Serverless GPU inference platform for generative media — image, video, audio, and 3D at scale.", "description": "Serverless inference API for 1,000+ generative media models including FLUX, Stable Diffusion, Kling, and Recraft. Developers integrate via REST API or Python/JavaScript SDK with no GPU provisioning; billing is per-inference with a free tier. Sub-second cold starts differentiate it from VM-based or self-hosted inference. Includes LoRA fine-tuning, a visual Workflow builder, and sandboxed code execution. Enterprise customers include Adobe, Canva, Shopify, Perplexity, and Amazon MGM Studios. Adoption evidence: 2.5M+ active developers, 50M+ AI-generated creations daily, $400M ARR (February 2026 est.), Series D at $4.5B valuation. Does not own the underlying models; long-term defensibility depends on speed and cost margins vs. cloud inference offerings from AWS, GCP, and Azure.", "users": "2.5M+ active developers", "userCount": 2500000, "releaseDate": "2021", "technicalImpact": 84, "competitors": [ "replicate", "modal", "fireworks-ai", "together-ai", "aws-bedrock" ], "features": [ "Serverless GPU inference", "1,000+ hosted generative media models", "LoRA fine-tuning", "Visual workflow builder", "Python and JavaScript SDKs", "Pay-per-inference billing" ], "pricing": { "api": "Pay-per-inference with free tier", "enterprise": "Contact sales" }, "link": "https://fal.ai", "apiLink": "https://fal.ai/docs", "aiCoreShare": 100, "scoringRationale": "84 reflects cleared adoption threshold (2.5M+ developers, $400M ARR est.) and enterprise distribution at Adobe, Canva, and Shopify. No first-party models and no proprietary model control prevent scoring above 84 despite strong revenue growth.", "sourceUrls": [ "https://techcrunch.com/2025/12/09/fal-nabs-140m-in-fresh-funding-led-by-sequoia-tripling-valuation-to-4-5b/", "https://fortune.com/2025/02/12/exclusive-fal-generative-media-platform-for-developers-raises-49-million-series-b/" ] }, { "id": "azure-openai", "name": "Azure OpenAI", "company": "Azure", "region": "global", "tagline": "Enterprise AI services", "description": "Azure OpenAI Service provides enterprise access to OpenAI models including the GPT-5 generation and image-generation models with Azure security, compliance, and regional availability. 80% of Fortune 500 use it.", "users": "Fortune 500", "userCount": 0, "releaseDate": "2023-01", "technicalImpact": 80, "competitors": [ "aws-bedrock", "gemini-enterprise-agent-platform", "mosaic-ai", "snowflake-cortex-ai" ], "features": [ "GPT-5 generation access", "Enterprise security", "Regional deployment", "Content filtering", "Fine-tuning" ], "pricing": { "usage_based": "Pay per token" }, "link": "https://azure.microsoft.com/products/ai-services/openai-service", "apiLink": "https://learn.microsoft.com/azure/ai-services/openai/", "aiCoreShare": 77 }, { "id": "aws-bedrock", "name": "AWS Bedrock", "company": "AWS", "region": "global", "tagline": "Enterprise foundation models", "description": "AWS Bedrock provides access to foundation models from AI21 Labs, Anthropic, Cohere, Meta, Stability AI, and Amazon via a single API for enterprise AI applications.", "users": "100K+ enterprises", "userCount": 100000, "releaseDate": "2023-09", "technicalImpact": 84, "competitors": [ "azure-openai", "gemini-enterprise-agent-platform", "mosaic-ai", "snowflake-cortex-ai", "Azure AI Foundry", "Databricks" ], "features": [ "Managed foundation-model access", "Serverless inference APIs", "Agents and knowledge bases", "Fine-tuning and RAG workflows", "Suite modules: Amazon Bedrock, Bedrock Agents, Bedrock Knowledge Bases" ], "pricing": { "usage_based": "Pay per token" }, "link": "https://aws.amazon.com/bedrock", "apiLink": "https://docs.aws.amazon.com/bedrock", "aiCoreShare": 99, "sourceUrls": [], "scoringRationale": "AWS's managed multi-model AI API serving foundation models (Anthropic Claude, Meta Llama, Amazon Nova, Mistral, Cohere) to enterprise customers. Score reflects cloud infrastructure distribution, enterprise contract depth, and AWS ecosystem lock-in rather than direct end-user count. 100K+ enterprise customers is a catalog estimate; AWS does not disclose Bedrock-specific customer numbers." }, { "id": "amazon-sagemaker-ai", "name": "Amazon SageMaker AI", "company": "AWS", "region": "global", "tagline": "Managed machine learning and model operations platform", "description": "Amazon SageMaker AI is AWS' managed machine learning and model operations platform.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://aws.amazon.com/machine-learning/", "apiLink": "https://docs.aws.amazon.com/machine-learning/", "aiCoreShare": 84, "contentReview": "needs-review" }, { "id": "speech-to-text", "name": "Speech-to-Text", "company": "Google Cloud", "region": "global", "tagline": "Speech recognition API", "description": "Speech-to-Text is Google Cloud's speech recognition API.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "Deepgram", "AssemblyAI", "Whisper API", "Amazon Transcribe" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://cloud.google.com/products/ai", "apiLink": "https://cloud.google.com/products/ai", "aiCoreShare": 86, "contentReview": "needs-review" }, { "id": "vector-embeddings-and-search", "name": "Vector Embeddings and Search", "company": "Google Cloud", "region": "global", "tagline": "Managed embeddings and vector search across Google data platforms", "description": "Vector Embeddings and Search is Google Cloud's managed embeddings and vector search across Google data platforms.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "OpenAI Embeddings", "Voyage AI", "Cohere Embed", "BGE" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://cloud.google.com/products/ai", "apiLink": "https://cloud.google.com/products/ai", "aiCoreShare": 76, "contentReview": "needs-review" }, { "id": "microsoft-cobalt", "name": "Microsoft Cobalt", "company": "Azure", "region": "global", "tagline": "Microsoft custom cloud CPU for AI infrastructure", "description": "Microsoft Cobalt is Microsoft's custom cloud CPU for AI infrastructure.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://learn.microsoft.com/en-us/azure/foundry/what-is-foundry", "apiLink": "https://learn.microsoft.com/en-us/azure/ai-services/", "aiCoreShare": 77, "contentReview": "needs-review" }, { "id": "microsoft-maia", "name": "Microsoft Maia", "company": "Azure", "region": "global", "tagline": "Microsoft custom AI accelerator", "description": "Microsoft Maia is Microsoft's custom AI accelerator.", "users": "Deployed in Microsoft Azure datacenters; unit counts not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "NVIDIA Hopper", "NVIDIA Blackwell", "AMD Instinct", "Cloud TPU v5" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://learn.microsoft.com/en-us/azure/foundry/what-is-foundry", "apiLink": "https://learn.microsoft.com/en-us/azure/ai-services/", "aiCoreShare": 74, "adoptionSignal": "Maia is Microsoft's own datacenter AI accelerator, deployed inside Azure to serve Microsoft and Azure OpenAI workloads rather than sold to third parties. Microsoft does not disclose deployed unit counts.", "contentReview": "needs-review" }, { "id": "openai-evals", "name": "OpenAI Evals", "company": "OpenAI", "region": "global", "tagline": "Evaluation framework for model behavior and quality", "description": "OpenAI Evals is OpenAI's evaluation framework for model behavior and quality.", "users": "Not disclosed", "userCount": 0, "releaseDate": "2026-02", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://github.com/openai/evals", "apiLink": "https://platform.openai.com/docs", "aiCoreShare": 78, "contentReview": "needs-review" }, { "id": "openai-realtime-api", "name": "OpenAI Realtime API", "company": "OpenAI", "region": "global", "tagline": "Low-latency voice and multimodal API", "description": "OpenAI Realtime API is OpenAI's low-latency voice and multimodal API.", "users": "Not disclosed", "userCount": 0, "releaseDate": "2026-02", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://platform.openai.com/docs/guides/realtime", "apiLink": "https://platform.openai.com/docs", "aiCoreShare": 85, "contentReview": "needs-review" }, { "id": "sendbird", "name": "Sendbird", "company": "Sendbird", "region": "korea", "tagline": "Chat API platform (7B convos/month)", "description": "Korean-founded unicorn messaging API platform processing 7 billion conversations per month. YC W16 graduate. Pivoting to AI agents with Delight.ai — branded AI concierge with long-term memory.", "users": "4,000+ global customers", "userCount": 4000, "adoptionSignal": "7B+ monthly conversations across global customer apps; named customers include Redfin, Teladoc, Noom, Match Group, Hinge, Yahoo, and Lotte.", "releaseDate": "2013-01", "technicalImpact": 78, "competitors": [ "intercom", "zendesk", "twilio" ], "features": [ "7B conversations/month", "Chat and messaging APIs", "AI agent integration", "Voice and video" ], "pricing": { "starter": "Free tier", "pro": "Contact", "enterprise": "Contact Sendbird" }, "link": "https://sendbird.com", "apiLink": "https://sendbird.com/docs", "aiCoreShare": 72, "scoringRationale": "Sendbird is a globally proven communications API business, but this row is not a frontier AI product. Its AI relevance comes through agent integration and Delight.ai rather than standalone model, developer, or agent-platform leadership.", "sourceUrls": [ "https://sendbird.com/customers", "https://claude.com/customers/sendbird" ] }, { "id": "amazon-comprehend", "name": "Amazon Comprehend", "company": "AWS", "region": "global", "tagline": "Natural language processing service", "description": "Amazon Comprehend is AWS' natural language processing service.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://aws.amazon.com/machine-learning/", "apiLink": "https://docs.aws.amazon.com/machine-learning/", "aiCoreShare": 78, "contentReview": "needs-review" }, { "id": "amazon-kendra", "name": "Amazon Kendra", "company": "AWS", "region": "global", "tagline": "Enterprise search service", "description": "Amazon Kendra is AWS' enterprise search service.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "Glean", "Cohere RAG", "Vectara", "Mendable" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://aws.amazon.com/machine-learning/", "apiLink": "https://docs.aws.amazon.com/machine-learning/", "aiCoreShare": 77, "contentReview": "needs-review" }, { "id": "aws-neuron", "name": "AWS Neuron", "company": "AWS", "region": "global", "tagline": "SDK and compiler stack for Trainium and Inferentia", "description": "AWS Neuron is AWS' sDK and compiler stack for Trainium and Inferentia.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://aws.amazon.com/machine-learning/", "apiLink": "https://docs.aws.amazon.com/machine-learning/", "aiCoreShare": 81, "contentReview": "needs-review" }, { "id": "mlflow", "name": "MLflow", "company": "Databricks", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Snowflake Cortex" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://mlflow.org/", "apiLink": null, "aiCoreShare": 79, "sourceUrls": [ "https://mlflow.org/" ], "contentReview": "needs-review" }, { "id": "text-to-speech", "name": "Text-to-Speech", "company": "Google Cloud", "region": "global", "tagline": "Voice synthesis API", "description": "Text-to-Speech is Google Cloud's voice synthesis API.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "ElevenLabs", "Murf", "LOVO", "Resemble AI" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://cloud.google.com/products/ai", "apiLink": "https://cloud.google.com/products/ai", "aiCoreShare": 86, "contentReview": "needs-review" }, { "id": "azure-ai-search", "name": "Azure AI Search", "company": "Azure", "region": "global", "tagline": "AI-powered enterprise and application search", "description": "Azure AI Search is Microsoft's AI-powered enterprise and application search.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://learn.microsoft.com/en-us/azure/foundry/what-is-foundry", "apiLink": "https://learn.microsoft.com/en-us/azure/ai-services/", "aiCoreShare": 74, "contentReview": "needs-review" }, { "id": "openai-agents-sdk", "name": "OpenAI Agents SDK", "company": "OpenAI", "region": "global", "tagline": "SDK for building tool-using AI agents", "description": "OpenAI Agents SDK is OpenAI's sDK for building tool-using AI agents.", "users": "Not disclosed", "userCount": 0, "releaseDate": "2026-02", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://openai.github.io/openai-agents-python/", "apiLink": "https://platform.openai.com/docs", "aiCoreShare": 82, "contentReview": "needs-review" }, { "id": "amazon-healthscribe", "name": "Amazon HealthScribe", "company": "AWS", "region": "global", "tagline": "Clinical documentation service", "description": "Amazon HealthScribe is AWS' clinical documentation service.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "Suki", "Abridge", "DeepScribe", "Nuance DAX" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://aws.amazon.com/machine-learning/", "apiLink": "https://docs.aws.amazon.com/machine-learning/", "aiCoreShare": 84, "contentReview": "needs-review" }, { "id": "amazon-polly", "name": "Amazon Polly", "company": "AWS", "region": "global", "tagline": "Text-to-speech service", "description": "Amazon Polly is AWS' text-to-speech service.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "ElevenLabs", "Murf", "LOVO", "Resemble AI" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://aws.amazon.com/machine-learning/", "apiLink": "https://docs.aws.amazon.com/machine-learning/", "aiCoreShare": 82, "contentReview": "needs-review" }, { "id": "aws-inferentia", "name": "AWS Inferentia", "company": "AWS", "region": "global", "tagline": "Custom inference accelerator", "description": "AWS Inferentia is AWS' custom inference accelerator.", "users": "Available as EC2 Inf instances across AWS regions worldwide; unit counts not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "NVIDIA Hopper", "NVIDIA Blackwell", "AMD Instinct", "Cloud TPU v5" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://aws.amazon.com/machine-learning/", "apiLink": "https://docs.aws.amazon.com/machine-learning/", "aiCoreShare": 79, "adoptionSignal": "Inferentia is AWS's own inference accelerator, rentable as EC2 Inf instances across AWS regions worldwide. AWS does not disclose deployed unit counts or customer numbers.", "contentReview": "needs-review" }, { "id": "aws-trainium2", "name": "AWS Trainium2", "company": "AWS", "region": "global", "tagline": "Second-generation AWS training accelerator", "description": "AWS Trainium2 is AWS' second-generation AWS training accelerator.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://aws.amazon.com/machine-learning/", "apiLink": "https://docs.aws.amazon.com/machine-learning/", "aiCoreShare": 79, "contentReview": "needs-review" }, { "id": "agent-search", "name": "Agent Search", "company": "Google Cloud", "region": "global", "tagline": "Managed agentic search for sites and applications", "description": "Agent Search is Google Cloud's managed agentic search for sites and applications.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://cloud.google.com/products/ai", "apiLink": "https://cloud.google.com/products/ai", "aiCoreShare": 74, "contentReview": "needs-review" }, { "id": "cortex-analyst", "name": "Cortex Analyst", "company": "Snowflake", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-analyst", "apiLink": null, "aiCoreShare": 88, "sourceUrls": [ "https://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-analyst" ], "contentReview": "needs-review" }, { "id": "amazon-devops-guru", "name": "Amazon DevOps Guru", "company": "AWS", "region": "global", "tagline": "ML-powered operations insights", "description": "Amazon DevOps Guru is AWS' mL-powered operations insights.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "GitHub Copilot", "Cursor", "Windsurf Editor", "Tabnine" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://aws.amazon.com/machine-learning/", "apiLink": "https://docs.aws.amazon.com/machine-learning/", "aiCoreShare": 78, "contentReview": "needs-review" }, { "id": "amazon-fraud-detector", "name": "Amazon Fraud Detector", "company": "AWS", "region": "global", "tagline": "Fraud detection service", "description": "Amazon Fraud Detector is AWS' fraud detection service.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "Sardine", "Sift", "Stripe Radar", "Featurespace" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://aws.amazon.com/machine-learning/", "apiLink": "https://docs.aws.amazon.com/machine-learning/", "aiCoreShare": 85, "contentReview": "needs-review" }, { "id": "amazon-translate", "name": "Amazon Translate", "company": "AWS", "region": "global", "tagline": "Machine translation service", "description": "Amazon Translate is AWS' machine translation service.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://aws.amazon.com/machine-learning/", "apiLink": "https://docs.aws.amazon.com/machine-learning/", "aiCoreShare": 87, "contentReview": "needs-review" }, { "id": "friendli-engine", "name": "Friendli Inference Cloud", "company": "FriendliAI", "region": "korea", "tagline": "Managed, engine-optimized inference for generative AI models.", "description": "Friendli Inference Cloud is FriendliAI's managed inference platform — serverless Model APIs (deploy from 570K+ Hugging Face models with no setup) and dedicated endpoints for proprietary/fine-tuned models — all powered by the proprietary Friendli Engine (custom GPU kernels, continuous batching, speculative decoding). The value is tokens/sec per GPU: customers run the same load on fewer GPUs rather than just renting cheaper GPUs. Backed by a 99.99% uptime SLA and SOC 2 Type II / HIPAA compliance.", "descriptionKo": "Friendli Inference Cloud는 FriendliAI의 관리형 추론 플랫폼으로, 57만 개 이상의 Hugging Face 모델을 별도 설정 없이 배포할 수 있는 서버리스 Model API와 사내 파인튜닝 모델을 위한 전용 엔드포인트를 함께 제공합니다. 핵심 경쟁력은 자체 개발 추론 엔진(Friendli Engine)으로, 연속 배칭(continuous batching)·투기적 디코딩(speculative decoding)·커스텀 GPU 커널을 활용해 GPU당 토큰 처리 속도를 높입니다. 즉, 같은 부하를 더 적은 GPU로 처리함으로써 운영 비용을 줄이는 것이 핵심 가치입니다. SK텔레콤, LG AI 리서치, 업스테이지가 고객사로 이름을 올리고 있으며, 99.99% 업타임 SLA와 SOC 2 Type II·HIPAA 컴플라이언스를 갖추고 있습니다. 2025년 $2,000만 시드 익스텐션 라운드를 완료했으며, 전년 대비 약 6~7배 매출 성장을 보고하고 있습니다.", "users": "Named enterprise customers (count undisclosed)", "userCount": 0, "releaseDate": "2021-01", "technicalImpact": 79, "competitors": [ "together-ai", "Fireworks AI", "Baseten", "modal", "anyscale", "replicate" ], "features": [ "Serverless Model APIs (570K+ models)", "Dedicated endpoints for custom models", "Friendli Engine: continuous batching, speculative decoding", "99.99% SLA, SOC 2 Type II, HIPAA" ], "pricing": { "managed": "Per-token usage-based", "enterprise": "Dedicated endpoints / contact sales" }, "link": "https://friendli.ai/products/serverless-endpoints", "apiLink": "https://docs.friendli.ai", "aiCoreShare": 83, "adoptionSignal": "FriendliAI reports ~6–7x year-over-year revenue growth in 2025 and names SK Telecom, LG AI Research, and Upstage among customers; NextDay AI cites serving trillions of tokens with 50% fewer GPUs on Friendli. Backed by a $20M seed extension (2025, Capstone Partners with Sierra Ventures, Alumni Ventures, KDB, KB).", "scoringRationale": "Reduced from 82 to 79 in the 2026-07-29 fairness pass. The engine IP is genuinely differentiated (the team's continuous-batching research defined industry-standard LLM serving) and the named enterprise customers (SK Telecom, LG AI Research, Upstage) are specific — but they are self-published on friendli.ai, the company is seed-stage, and per-product usage is undisclosed. 79 places it above unnamed-'Enterprise' evidence (Delight.ai 78) and below ATOM-Max (82), whose named deployment is third-party sourced and CSAP-certified. Clears higher with quantified usage or independently sourced deployments.", "sourceUrls": [ "https://friendli.ai/products/serverless-endpoints", "https://friendli.ai/news/friendliai-raises-20m-in-seed-extension-round", "https://news.crunchbase.com/ai/inference-platform-friendliai-raises-seed-extension-chun/" ] }, { "id": "gemini-enterprise-agent-platform", "name": "Gemini Enterprise Agent Platform", "company": "Google Cloud", "region": "global", "tagline": "Google Cloud model and agent development platform, formerly Vertex AI", "description": "Gemini Enterprise Agent Platform succeeds Vertex AI for building, deploying and governing models and agents. Google announced the transition in April 2026. Model Garden, Agent Studio and Agent Garden are capabilities of this suite, not separate products. This developer platform is distinct from the employee-facing Gemini Enterprise app.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 80, "competitors": [ "aws-bedrock", "microsoft-foundry", "mosaic-ai" ], "features": [ "Model Garden and managed model access", "Agent Studio and Agent Garden", "Agent runtime, memory and orchestration", "Model training, deployment and evaluation", "Agent identity, security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://cloud.google.com/products/gemini-enterprise-agent-platform", "apiLink": "https://cloud.google.com/products/gemini-enterprise-agent-platform", "aiCoreShare": 79, "sourceUrls": [ "https://cloud.google.com/products/gemini-enterprise-agent-platform", "https://cloud.google.com/blog/products/ai-machine-learning/introducing-gemini-enterprise-agent-platform" ], "adoptionSignal": "Google names Comcast, Color Health and PayPal among Agent Platform users. Product-specific active users and revenue are not disclosed.", "scoringRationale": "Reviewed 2026-09-29: carries forward Vertex AI at 80 without counting its successor and suite modules as additional products. Named production workflows support relevance; Google Cloud revenue and Gemini app users are not attributed to this platform.", "versions": [ { "name": "Gemini Enterprise Agent Platform", "released": "2026-04", "note": "Successor to Vertex AI; distinct from the Gemini Enterprise workplace app." }, { "name": "Vertex AI", "note": "Former name; services continue under Agent Platform." } ], "versionsUpdated": "2026-09-29" }, { "id": "translation-ai", "name": "Translation AI", "company": "Google Cloud", "region": "global", "tagline": "Machine translation API", "description": "Translation AI is Google Cloud's machine translation API.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "amazon-translate", "DeepL", "Papago", "Lilt" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://cloud.google.com/products/ai", "apiLink": "https://cloud.google.com/products/ai", "aiCoreShare": 84, "contentReview": "needs-review" }, { "id": "azure-machine-learning", "name": "Azure Machine Learning", "company": "Azure", "region": "global", "tagline": "Managed platform for ML engineering and MLOps", "description": "Azure Machine Learning is Microsoft's managed platform for ML engineering and MLOps.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://learn.microsoft.com/en-us/azure/foundry/what-is-foundry", "apiLink": "https://learn.microsoft.com/en-us/azure/ai-services/", "aiCoreShare": 77, "contentReview": "needs-review" }, { "id": "prompt-flow", "name": "Prompt Flow", "company": "Azure", "region": "global", "tagline": "Development workflow for LLM applications", "description": "Prompt Flow is Microsoft's development workflow for LLM applications.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "GPT-5", "claude", "Gemini 2.5", "Grok 4" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://learn.microsoft.com/en-us/azure/foundry/what-is-foundry", "apiLink": "https://learn.microsoft.com/en-us/azure/ai-services/", "aiCoreShare": 73, "contentReview": "needs-review" }, { "id": "openai-responses-api", "name": "OpenAI Responses API", "company": "OpenAI", "region": "global", "tagline": "Unified API for model, tool, and agent workflows", "description": "OpenAI Responses API is OpenAI's unified API for model, tool, and agent workflows.", "users": "Not disclosed", "userCount": 0, "releaseDate": "2026-02", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://platform.openai.com/docs/api-reference/responses", "apiLink": "https://platform.openai.com/docs", "aiCoreShare": 75, "contentReview": "needs-review" }, { "id": "amazon-textract", "name": "Amazon Textract", "company": "AWS", "region": "global", "tagline": "Document extraction service", "description": "Amazon Textract is AWS' document extraction service.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://aws.amazon.com/machine-learning/", "apiLink": "https://docs.aws.amazon.com/machine-learning/", "aiCoreShare": 87, "contentReview": "needs-review" }, { "id": "lakehouse-monitoring", "name": "Lakehouse Monitoring", "company": "Databricks", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Snowflake Cortex" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://docs.databricks.com/aws/en/data-governance/unity-catalog/data-quality-monitoring/data-profiling", "apiLink": null, "aiCoreShare": 79, "sourceUrls": [ "https://docs.databricks.com/aws/en/data-governance/unity-catalog/data-quality-monitoring/data-profiling" ], "contentReview": "needs-review" }, { "id": "github-advanced-security-ai", "name": "GitHub Advanced Security AI", "company": "GitHub", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "CrowdStrike", "Splunk", "Palo Alto Networks", "SentinelOne" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://github.com/features/copilot", "apiLink": "https://docs.github.com/copilot", "aiCoreShare": 76, "contentReview": "needs-review" }, { "id": "vision-ai", "name": "Vision AI", "company": "Google Cloud", "region": "global", "tagline": "Image and video understanding APIs", "description": "Vision AI is Google Cloud's image and video understanding APIs.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://cloud.google.com/products/ai", "apiLink": "https://cloud.google.com/products/ai", "aiCoreShare": 80, "contentReview": "needs-review" }, { "id": "groqcloud", "name": "GroqCloud", "company": "Groq", "region": "global", "tagline": "Developer inference cloud powered by Groq LPU silicon", "description": "GroqCloud is Groq's self-serve and enterprise inference platform for serving LLM, speech, and vision models on Groq LPU hardware. It deserves a global-infrastructure score because Groq reports 3M developers and teams, named customer logos, public/private/co-cloud options, and four global data-center regions.", "users": "3M developers and teams reported by Groq; active users not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 85, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks", "cerebras-inference", "Together AI", "Anyscale" ], "features": [ "Developer workflow automation", "Model operations", "Observability", "Secure deployment", "Edge or embodied AI", "Inference acceleration", "Autonomous operation", "Industrial deployment" ], "pricing": { "free": "Varies by plan", "pro": "Subscription or usage-based", "enterprise": "Contact sales", "api": "Usage-based or enterprise pricing" }, "link": "https://groq.com/groqcloud", "apiLink": null, "aiCoreShare": 87, "adoptionSignal": "GroqCloud reports 3M developers and teams, popular-model serving, public/private/co-cloud deployment options, and four global data-center regions.", "scoringRationale": "GroqCloud is Groq's developer API and inference platform delivering the fastest publicly benchmarked LLM inference (sub-100ms TTFT on leading models) via custom LPU hardware. Groq reports 3M developers and teams, but this figure is company-reported and not independently verified. Reduced from 88 to 85: at 88 it sat above Midjourney (87, 21M independently reported users), which is not defensible on unverified developer counts alone. 85 reflects a genuinely differentiated inference platform with real traction but without the scale evidence that would justify the 87–90 tier. Re-evaluates upward if Groq discloses verified API call volume or revenue.", "sourceUrls": [ "https://groq.com/groqcloud", "https://groq.com/lpu-architecture" ] }, { "id": "backend-ai", "name": "Backend.AI", "company": "Lablup", "region": "korea", "tagline": "Lablup's GPU cluster orchestration platform, self-hosted or managed", "description": "Backend.AI is Lablup's platform for orchestrating GPU and NPU clusters for AI training and inference, deployed on-premises or consumed as Backend.AI Cloud. Deployment mode is packaging, not a separate product.", "descriptionKo": "Backend.AI는 랩업(Lablup)이 개발한 오픈소스 GPU 클러스터 관리 플랫폼으로, GPU 가상화·멀티테넌트 클러스터 운영·컨테이너 오케스트레이션·Jupyter 통합을 제공합니다. 국내 주요 대학 및 연구 기관에서 AI 연구 인프라로 활용하고 있으며, 업스테이지와 함께 한국 국가 AI 이니셔티브의 소버린 AI 컨소시엄에 참여하고 있습니다. 오픈소스로 무료 사용이 가능하며, 기업용 지원은 Lablup과 직접 계약합니다. GPU 자원을 여러 연구팀이 공유하는 환경에서의 스케줄링·격리·모니터링에 특화되어 있습니다.", "users": "Research institutions", "userCount": 0, "releaseDate": "2015-01", "technicalImpact": 78, "competitors": [ "run-ai", "anyscale", "coreweave", "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "Open-source GPU management", "Multi-tenant clusters", "Jupyter integration", "Container orchestration", "Suite modules: Backend.AI Cloud", "Developer workflow automation", "Model operations", "Observability", "Secure deployment" ], "pricing": { "open_source": "Free", "enterprise": "Contact Lablup" }, "link": "https://backend.ai", "apiLink": "https://docs.backend.ai", "aiCoreShare": 84, "nativeName": "백엔드.AI", "sourceUrls": [ "https://www.backend.ai/" ], "versions": [ { "name": "Backend.AI", "released": "2015-01", "variants": [ "Backend.AI Cloud" ] } ], "versionsUpdated": "2026-07-29", "scoringRationale": "2026-07-29 Korean top-tier stress test: reduced 80 -> 78. Open-source project with visible public traction and research-institution use — third-party-visible ecosystem evidence clears the unnamed tier — but no named production deployments or quantified adoption." }, { "id": "microsoft-autogen", "name": "Microsoft AutoGen", "company": "Azure", "region": "global", "tagline": "Framework for multi-agent AI applications", "description": "Microsoft AutoGen is Microsoft's framework for multi-agent AI applications.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "LangChain", "LlamaIndex", "CrewAI", "Haystack" ], "features": [ "Suite modules: AutoGen Studio" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://learn.microsoft.com/en-us/azure/foundry/what-is-foundry", "apiLink": "https://learn.microsoft.com/en-us/azure/ai-services/", "aiCoreShare": 86, "sourceUrls": [], "contentReview": "needs-review" }, { "id": "microsoft-foundry", "name": "Microsoft Foundry", "company": "Azure", "region": "global", "tagline": "Unified Azure platform for agents, models, tools, and AI operations", "description": "Microsoft Foundry is Microsoft's unified Azure platform for agents, models, tools, and AI operations.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks", "ElevenLabs", "Murf", "LOVO", "Resemble AI", "Cognigy", "Botpress" ], "features": [ "Model catalog and deployment", "Agent-service orchestration", "Document, speech, language, and vision tools", "Content safety and governance", "Suite modules: Foundry Tools, Microsoft Foundry Agents, Foundry Agent Service" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://learn.microsoft.com/en-us/azure/foundry/what-is-foundry", "apiLink": "https://learn.microsoft.com/en-us/azure/ai-services/", "aiCoreShare": 85, "sourceUrls": [], "contentReview": "needs-review" }, { "id": "netspresso", "name": "NetsPresso", "company": "Nota AI", "region": "korea", "tagline": "AI model compression and optimization", "description": "NetsPresso automates AI model compression, pruning, and quantization for on-device deployment. Clients include NVIDIA, Samsung, Qualcomm, Sony. First Korean AI optimization company to IPO on KOSDAQ (Nov 2025).", "descriptionKo": "NetsPresso는 Nota AI가 개발한 AI 모델 경량화·최적화 플랫폼으로, 온디바이스 배포를 위한 모델 압축(pruning), 양자화(quantization), 하드웨어 인식 최적화(hardware-aware optimization), AutoML 압축을 자동화합니다. NetsPresso Compressor, Modeler, Search 세 가지 모듈로 구성되며, NVIDIA·삼성·Qualcomm·Sony가 고객사로 이름을 올리고 있습니다. Nota AI는 2025년 11월 코스닥 IPO를 완료한 한국 최초의 AI 최적화 전문 상장 기업입니다. 엣지 기기에서 대형 모델을 실행해야 하는 기업의 추론 비용·전력 소비를 줄이는 데 특화되어 있습니다.", "users": "Enterprise", "userCount": 0, "releaseDate": "2015-01", "technicalImpact": 77, "competitors": [ "octoml", "neural-magic", "deci-ai", "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "Model pruning and quantization", "Edge deployment", "Hardware-aware optimization", "AutoML compression", "Suite modules: NetsPresso Compressor, NetsPresso Modeler, NetsPresso Search", "Developer workflow automation", "Model operations", "Observability", "Secure deployment" ], "pricing": { "free": "Community tier", "enterprise": "Contact Nota AI" }, "link": "https://netspresso.ai", "apiLink": "https://www.nota.ai/netspresso", "aiCoreShare": 89, "sourceUrls": [], "scoringRationale": "2026-07-29 Korean top-tier stress test: reduced 80 -> 77. Real edge-optimization niche, but the row's own rationale concedes no customer count, license volume or named production program. Unnamed relationships are 77-tier evidence." }, { "id": "saip-s2w", "name": "S2W", "company": "S2W", "region": "korea", "tagline": "Dark-web threat intelligence and security analytics", "description": "S2W collects and analyses dark-web, deep-web and threat-actor data, delivering intelligence products and an analytics platform to Korean government agencies, financial institutions and enterprises. The proprietary collection corpus is the asset rather than the model. S2W listed on KOSDAQ in September 2025; per-product customer counts are not disclosed.", "users": "Enterprise", "userCount": 0, "releaseDate": "2022-01", "technicalImpact": 78, "competitors": [ "crowdstrike", "splunk", "palo-alto", "Palo Alto Networks", "SentinelOne" ], "features": [ "Real-time threat detection", "AI-driven response", "Dark web monitoring", "Enterprise integration", "Suite modules: S2W XARVIS, S2W Threat Intelligence, S2W Quaxar", "Risk analytics", "Research automation", "Compliance workflow", "Decision support" ], "pricing": { "enterprise": "Contact S2W" }, "link": "https://s2w.inc", "apiLink": null, "aiCoreShare": 73, "sourceUrls": [ "https://s2w.inc/en" ], "versions": [ { "name": "S2W", "released": "2022-01", "variants": [ "S2W XARVIS", "S2W Threat Intelligence", "S2W Quaxar" ] } ], "versionsUpdated": "2026-07-29", "scoringRationale": "2026-07-29 Korean top-tier stress test: reduced 80 -> 78. Public-company disclosure discipline (KOSDAQ 488280) and documented intelligence-sector partnerships clear the unnamed tier; product-specific customer counts and usage remain undisclosed." }, { "id": "amazon-lex", "name": "Amazon Lex", "company": "AWS", "region": "global", "tagline": "Conversational interface service", "description": "Amazon Lex is AWS' conversational interface service.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://aws.amazon.com/machine-learning/", "apiLink": "https://docs.aws.amazon.com/machine-learning/", "aiCoreShare": 85, "contentReview": "needs-review" }, { "id": "amazon-sagemaker-studio", "name": "Amazon SageMaker Studio", "company": "AWS", "region": "global", "tagline": "Integrated ML development environment", "description": "Amazon SageMaker Studio is AWS' integrated ML development environment.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://aws.amazon.com/machine-learning/", "apiLink": "https://docs.aws.amazon.com/machine-learning/", "aiCoreShare": 86, "contentReview": "needs-review" }, { "id": "amazon-transcribe", "name": "Amazon Transcribe", "company": "AWS", "region": "global", "tagline": "Speech-to-text service", "description": "Amazon Transcribe is AWS' speech-to-text service.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "Suki", "Abridge", "DeepScribe", "Nuance DAX" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://aws.amazon.com/machine-learning/", "apiLink": "https://docs.aws.amazon.com/machine-learning/", "aiCoreShare": 82, "contentReview": "needs-review" }, { "id": "document-ai", "name": "Document AI", "company": "Google Cloud", "region": "global", "tagline": "Document processing and structured extraction platform", "description": "Document AI is Google Cloud's document processing and structured extraction platform.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://cloud.google.com/products/ai", "apiLink": "https://cloud.google.com/products/ai", "aiCoreShare": 83, "contentReview": "needs-review" }, { "id": "natural-language-ai", "name": "Natural Language AI", "company": "Google Cloud", "region": "global", "tagline": "Text understanding API for classification and sentiment", "description": "Natural Language AI is Google Cloud's text understanding API for classification and sentiment.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://cloud.google.com/products/ai", "apiLink": "https://cloud.google.com/products/ai", "aiCoreShare": 75, "contentReview": "needs-review" }, { "id": "video-ai", "name": "Video AI", "company": "Google Cloud", "region": "global", "tagline": "Video metadata and content intelligence APIs", "description": "Video AI is Google Cloud's video metadata and content intelligence APIs.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://cloud.google.com/products/ai", "apiLink": "https://cloud.google.com/products/ai", "aiCoreShare": 80, "contentReview": "needs-review" }, { "id": "llamaindex", "name": "LlamaIndex", "company": "LlamaIndex", "region": "global", "tagline": "Connect agents to documents and enterprise data", "description": "An open-source framework for building context-aware agents over documents and other data sources. LlamaIndex provides data connectors, indexing, retrieval and workflow components; its managed document services cover parsing and extraction. Evaluate retrieval quality and operational cost on your own documents; aggregate adoption is not disclosed here.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 79, "competitors": [ "LangChain", "CrewAI", "AutoGen", "Haystack" ], "features": [ "Data connectors and indexing", "Retrieval and agent workflows", "Document parsing and extraction services" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://www.llamaindex.ai/llamaindex", "apiLink": null, "aiCoreShare": 84, "sourceUrls": [ "https://www.llamaindex.ai/llamaindex" ], "taglineKo": "에이전트와 문서·기업 데이터 연결", "descriptionKo": "문서와 여러 데이터 소스를 활용하는 에이전트를 만드는 오픈소스 프레임워크입니다. 데이터 연결, 인덱싱, 검색, 워크플로 구성 요소와 관리형 문서 파싱·추출 서비스를 제공합니다. 검색 품질과 운영 비용은 실제 자료로 검증해야 하며, 전체 사용 규모는 이 기록에서 확인되지 않았습니다." }, { "id": "semantic-kernel", "name": "Semantic Kernel", "company": "Azure", "region": "global", "tagline": "Open-source SDK for agentic applications", "description": "Semantic Kernel is Microsoft's open-source SDK for agentic applications.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks", "Cognigy", "Botpress", "Kore.ai", "Yellow.ai" ], "features": [ "Suite modules: Semantic Kernel Agents" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://learn.microsoft.com/en-us/azure/foundry/what-is-foundry", "apiLink": "https://learn.microsoft.com/en-us/azure/ai-services/", "aiCoreShare": 82, "sourceUrls": [], "contentReview": "needs-review" }, { "id": "cortex-search", "name": "Cortex Search", "company": "Snowflake", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-search/cortex-search-overview", "apiLink": null, "aiCoreShare": 87, "sourceUrls": [ "https://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-search/cortex-search-overview" ], "contentReview": "needs-review" }, { "id": "snowpark-ml", "name": "Snowpark ML", "company": "Snowflake", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://docs.snowflake.com/en/developer-guide/snowflake-ml/overview", "apiLink": null, "aiCoreShare": 80, "sourceUrls": [ "https://docs.snowflake.com/en/developer-guide/snowflake-ml/overview" ], "contentReview": "needs-review" }, { "id": "amazon-codeguru", "name": "Amazon CodeGuru", "company": "AWS", "region": "global", "tagline": "AI code review and profiling service", "description": "Amazon CodeGuru is AWS' AI code review and profiling service.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "GitHub Copilot", "Cursor", "Windsurf Editor", "Tabnine" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://aws.amazon.com/machine-learning/", "apiLink": "https://docs.aws.amazon.com/machine-learning/", "aiCoreShare": 78, "contentReview": "needs-review" }, { "id": "amazon-personalize", "name": "Amazon Personalize", "company": "AWS", "region": "global", "tagline": "Recommendation service", "description": "Amazon Personalize is AWS' recommendation service.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks", "Dynamic Yield", "Bloomreach", "Constructor", "Algolia" ], "features": [ "Suite modules: Amazon Personalize" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://aws.amazon.com/machine-learning/", "apiLink": "https://docs.aws.amazon.com/machine-learning/", "aiCoreShare": 81, "sourceUrls": [], "contentReview": "needs-review" }, { "id": "amazon-q-business", "name": "Amazon Q Business", "company": "AWS", "region": "global", "tagline": "Enterprise AI assistant connected to work data", "description": "Amazon Q Business is AWS' enterprise AI assistant connected to work data.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks", "Cognigy", "Botpress", "Kore.ai", "Yellow.ai" ], "features": [ "Suite modules: Amazon Q Business Plugins" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://aws.amazon.com/machine-learning/", "apiLink": "https://docs.aws.amazon.com/machine-learning/", "aiCoreShare": 92, "sourceUrls": [], "contentReview": "needs-review" }, { "id": "amazon-sagemaker-hyperpod", "name": "Amazon SageMaker HyperPod", "company": "AWS", "region": "global", "tagline": "Distributed training infrastructure", "description": "Amazon SageMaker HyperPod is AWS' distributed training infrastructure.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://aws.amazon.com/machine-learning/", "apiLink": "https://docs.aws.amazon.com/machine-learning/", "aiCoreShare": 81, "contentReview": "needs-review" }, { "id": "friendli-container", "name": "Friendli Container", "company": "FriendliAI", "region": "korea", "tagline": "Self-hostable Friendli Engine for inference in your own environment.", "description": "Friendli Container packages the proprietary Friendli Engine as a container that teams run in their own VPC, on-prem, or any cloud (including on rented GPUs from a neocloud), keeping the inference-efficiency gains — continuous batching, speculative decoding, custom kernels — while retaining full data control. It targets enterprises with compliance, residency, or existing-GPU constraints that rule out fully managed endpoints.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 77, "competitors": [ "vLLM", "NVIDIA NIM", "TensorRT-LLM", "together-ai" ], "features": [ "Self-hosted Friendli Engine", "Continuous batching and speculative decoding", "Runs in your VPC, on-prem, or any cloud", "Full data control for regulated workloads" ], "pricing": { "enterprise": "Container license / contact sales" }, "link": "https://github.com/friendliai/container-resource", "apiLink": "https://docs.friendli.ai", "aiCoreShare": 87, "adoptionSignal": "FriendliAI names Scatter Lab as a production Container customer serving Zeta across Korea, Japan and the US. Reported interactions are not unique users.", "scoringRationale": "Reviewed 2026-09-29: retains 77. Self-hosted packaging of Friendli Engine, with a named production case study; do not double-count the same deployment as additional platform adoption.", "sourceUrls": [ "https://friendli.ai/news/friendliai-raises-20m-in-seed-extension-round", "https://github.com/friendliai/container-resource", "https://friendli.ai/customers/scatter-lab" ] }, { "id": "github-models", "name": "GitHub Models", "company": "GitHub", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://github.com/features/copilot", "apiLink": "https://docs.github.com/copilot", "aiCoreShare": 82, "contentReview": "needs-review" }, { "id": "bigquery-ai", "name": "BigQuery AI", "company": "Google Cloud", "region": "global", "tagline": "SQL-native machine learning and generative AI in BigQuery", "description": "BigQuery AI is Google Cloud's sQL-native machine learning and generative AI in BigQuery.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://cloud.google.com/products/ai", "apiLink": "https://cloud.google.com/products/ai", "aiCoreShare": 86, "contentReview": "needs-review" }, { "id": "colab-enterprise", "name": "Colab Enterprise", "company": "Google Cloud", "region": "global", "tagline": "Enterprise notebook environment for data-to-AI work", "description": "Colab Enterprise is Google Cloud's enterprise notebook environment for data-to-AI work.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://cloud.google.com/products/ai", "apiLink": "https://cloud.google.com/products/ai", "aiCoreShare": 84, "contentReview": "needs-review" }, { "id": "snowflake-cortex-ai", "name": "Snowflake Cortex AI", "company": "Snowflake", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://snowflake.com", "apiLink": null, "aiCoreShare": 76, "contentReview": "needs-review" }, { "id": "twelve-labs-api", "name": "Twelve Labs API", "company": "Twelve Labs", "region": "korea", "tagline": "Video understanding API", "description": "Twelve Labs builds multimodal AI APIs for semantic video search and analysis. Korean-founded startup based in SF and Seoul — one of the few Korean teams competing at the global AI infrastructure layer.", "descriptionKo": "Twelve Labs API는 Twelve Labs의 멀티모달 비디오 이해 API로, 동영상 내 의미적 검색(semantic search), 비디오 Q&A, 이벤트 감지 기능을 제공합니다. 자체 개발한 Marengo(임베딩)와 Pegasus(생성) 모델을 엔드투엔드로 학습시켜, 범용 모델 래퍼가 아닌 비디오 특화 AI 인프라를 구축했습니다. SF와 서울에 거점을 둔 한국계 창업팀의 글로벌 스타트업으로, 3만 명 이상의 개발자가 플랫폼을 사용하고 있으며 Databricks·Snowflake와의 통합도 공개되어 있습니다. 글로벌 AI 인프라 레이어에서 직접 경쟁하는 드문 한국계 팀 중 하나입니다.", "users": "30K+ developers", "userCount": 30000, "adoptionSignal": "30,000+ developers use the platform; Databricks and Snowflake have built integrations around Twelve Labs video embeddings, but most end-customer names are not publicly disclosed.", "releaseDate": "2021-01", "technicalImpact": 80, "competitors": [ "video-ai", "gemini-enterprise-agent-platform", "aws-bedrock", "azure-openai" ], "features": [ "Semantic video search", "Video Q&A", "Event detection", "Marengo and Pegasus models" ], "pricing": { "free": "Free tier", "api": "Usage-based" }, "link": "https://twelvelabs.io", "apiLink": "https://docs.twelvelabs.io", "aiCoreShare": 83, "sourceUrls": [ "https://www.twelvelabs.io/blog/twelve-labs-is-building-ai-that-can-analyze-and-search-through-videos", "https://www.snowflake.com/en/blog/snowflake-ventures-invests-in-twelve-labs-to-bring-advanced-video-understanding-to-ai-data-cloud/" ], "scoringRationale": "Differentiated multimodal video-understanding API with proprietary Marengo and Pegasus model families trained end-to-end on video — not a commodity wrapper. 30K+ developers across media, sports, and enterprise use cases with Snowflake and Databricks integrations. Recalibrated from 84 to 80: 30K developers and two named platform integrations are real traction but do not clear the embedding/model API ceiling (84), which requires 1M+ userCount or hyperscaler-class ecosystem leverage. Production revenue scale and enterprise contract depth are not disclosed." }, { "id": "vessl-ai", "name": "VESSL Cloud", "company": "VESSL AI", "region": "korea", "tagline": "VESSL's GPU Liquidity Layer for multi-cloud GPU capacity.", "description": "VESSL Cloud is VESSL AI's GPU infrastructure product — a Korean \"neocloud\" / \"GPU Liquidity Layer\" that accesses, routes, and guarantees GPU capacity (A100, H100, B200, B300, GB200) across multiple clouds with spot, on-demand, and reserved options, self-service in under three minutes and per-minute billing up to ~80% cheaper than AWS/GCP. It is VESSL's primary product after the company repositioned from MLOps toward GPU cloud. GPU-cloud-specific usage is not disclosed.", "descriptionKo": "VESSL Cloud은 VESSL AI의 GPU 인프라 제품으로, 여러 클라우드에 흩어진 GPU를 하나로 묶어 접근·라우팅·확보하는 '신클라우드(neocloud)' 겸 'GPU 유동성 레이어'입니다. A100, H100, B200, B300, GB200 등의 GPU를 스팟·온디맨드·예약 방식으로 3분 이내에 셀프서비스로 확보할 수 있고, 분 단위 과금으로 AWS·GCP 대비 최대 약 80% 저렴합니다. MLOps에서 GPU 클라우드로 사업을 재편한 이후 VESSL의 핵심 제품이 되었으며, GPU 클라우드 자체의 사용자 수는 공개되지 않았습니다.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "CoreWeave", "Lambda", "RunPod", "Together AI", "Vast.ai" ], "features": [ "Multi-cloud GPU provisioning", "Spot, on-demand, and reserved capacity", "Cross-cloud GPU routing and failover", "Transparent unified pricing" ], "pricing": { "free": "Free tier", "team": "Usage-based", "enterprise": "Contact VESSL AI" }, "link": "https://vessl.ai/en", "apiLink": "https://docs.vessl.ai", "aiCoreShare": 80, "sourceUrls": [ "https://vessl.ai/en", "https://vessl.ai/en/blog/vesslcloud-en", "https://vessl.ai/en/blog/what-is-a-neocloud" ] }, { "id": "amazon-rekognition", "name": "Amazon Rekognition", "company": "AWS", "region": "global", "tagline": "Image and video analysis service", "description": "Amazon Rekognition is AWS' image and video analysis service.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://aws.amazon.com/machine-learning/", "apiLink": "https://docs.aws.amazon.com/machine-learning/", "aiCoreShare": 77, "contentReview": "needs-review" }, { "id": "aws-trainium", "name": "AWS Trainium", "company": "AWS", "region": "global", "tagline": "Custom training accelerator", "description": "AWS Trainium is AWS' custom training accelerator.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://aws.amazon.com/machine-learning/", "apiLink": "https://docs.aws.amazon.com/machine-learning/", "aiCoreShare": 74, "contentReview": "needs-review" }, { "id": "visual-copilot", "name": "Visual Copilot", "company": "Builder.io", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "GitHub Copilot", "Cursor", "Windsurf Editor", "Tabnine" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://www.builder.io/m/design-to-code", "apiLink": null, "aiCoreShare": 74, "sourceUrls": [ "https://www.builder.io/m/design-to-code" ], "contentReview": "needs-review" }, { "id": "datumo-evaluation-platform", "name": "Datumo", "company": "Datumo", "region": "korea", "tagline": "Training-data production and LLM evaluation for Korean AI builders", "description": "Datumo (formerly SelectStar) covers both halves of the model-quality problem: producing labelled training and instruction data through a managed workforce, and evaluating LLM applications through benchmark construction, automated scoring and red-teaming. Korean model builders and enterprises fine-tuning on proprietary data are the buyers, and the company sits directly upstream of Korea's sovereign model programme. Customer counts and contract values are not disclosed.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks", "LangSmith", "LangFuse", "Arize", "Helicone" ], "features": [ "Training-data collection and labelling", "Instruction and preference data for fine-tuning", "LLM evaluation and benchmark construction", "Red-teaming for safety and hallucination", "Managed annotation workforce", "Suite modules: Datumo Data Engine, SelectStar Data Platform, SelectStar LLM Evaluation", "Developer workflow automation", "Model operations", "Observability", "Secure deployment" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://datumo.com", "apiLink": null, "aiCoreShare": 83, "sourceUrls": [], "versions": [ { "name": "Datumo", "variants": [ "Datumo Data Engine", "SelectStar Data Platform", "SelectStar LLM Evaluation" ] } ], "scoringRationale": "78 reflects a company sitting directly upstream of Korea's sovereign model programme, covering both training-data production and LLM evaluation. Held at the evidence cap because Datumo discloses no customer count, contract value or named model customers.", "versionsUpdated": "2026-07-29" }, { "id": "huggingface", "name": "Hugging Face Hub", "company": "Hugging Face", "region": "global", "tagline": "The GitHub of machine learning", "description": "Hugging Face Hub is the de-facto model registry for the ML community — 500K+ models, 100K+ datasets, and Spaces for interactive ML apps. The Transformers library's `from_pretrained()` API is embedded in millions of production codebases, and virtually every major open-weight model release (Llama, Mistral, Stable Diffusion, EXAONE) ships on HuggingFace first, creating infrastructure-level lock-in. Backed by Google, Amazon, and Nvidia.", "users": "5M+", "userCount": 5000000, "releaseDate": "2018-01", "technicalImpact": 90, "competitors": [ "Replicate", "Weights & Biases", "GitHub Models", "NVIDIA NGC" ], "features": [ "500K+ hosted models", "Transformers library", "Datasets library", "Spaces for ML apps" ], "pricing": { "free": "Open-source tools", "pro": "$9/month", "enterprise": "Contact sales" }, "link": "https://huggingface.co", "apiLink": "https://huggingface.co/docs", "aiCoreShare": 84, "adoptionSignal": "5M+ registered users; 500K+ public models with millions of monthly downloads. Nearly every major open-weight model release publishes on HuggingFace first.", "scoringRationale": "De-facto infrastructure standard for open-weight model distribution. 5M+ users clears the adoption cap. Developer ecosystem lock-in is exceptional: model IDs are hardcoded in production code globally, `from_pretrained()` is the universal API, and switching away would require coordinated migration across the entire open-source ML ecosystem. Backed by Google, Amazon, and Nvidia as strategic investors.", "sourceUrls": [ "https://huggingface.co/models", "https://huggingface.co/blog/the-hub" ] }, { "id": "microsoft-phi-models", "name": "Microsoft Phi models", "company": "Azure", "region": "global", "tagline": "Small language model family for efficient AI", "description": "Microsoft Phi models is Microsoft's small language model family for efficient AI.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://learn.microsoft.com/en-us/azure/foundry/what-is-foundry", "apiLink": "https://learn.microsoft.com/en-us/azure/ai-services/", "aiCoreShare": 76, "contentReview": "needs-review" }, { "id": "baseten-inference", "name": "Baseten Inference", "company": "Baseten", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://baseten.co", "apiLink": null, "aiCoreShare": 79, "contentReview": "needs-review" }, { "id": "crowdworks-data-labeling", "name": "Crowdworks", "company": "Crowdworks", "region": "korea", "tagline": "Korea's largest crowdsourced AI training-data platform", "description": "Crowdworks runs one of Korea's largest crowdsourced data-production platforms, covering vision, speech and text annotation plus the instruction, preference and evaluation datasets Korean model builders need for fine-tuning and alignment — including Korean-language RLHF data foreign vendors supply poorly. It is KOSDAQ-listed, making it one of the few Korean data-infrastructure vendors with public financials. Per-product contract values are not disclosed.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "Scale AI", "Labelbox", "V7", "Surge AI", "GPT-5", "claude", "Gemini 2.5", "Grok 4", "AWS Bedrock", "gemini-enterprise-agent-platform" ], "features": [ "Crowdsourced vision, speech and text annotation", "Korean-language instruction and preference data", "Enterprise quality control and review", "Distributed worker base", "KOSDAQ-listed with public financials", "Suite modules: Crowdworks LLM Data, Crowdworks AI Data Platform", "Developer workflow automation", "Model operations", "Observability", "Secure deployment" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://crowdworks.kr", "apiLink": null, "aiCoreShare": 89, "sourceUrls": [], "versions": [ { "name": "Crowdworks", "variants": [ "Crowdworks LLM Data", "Crowdworks AI Data Platform" ] } ], "scoringRationale": "78 reflects KOSDAQ-listed status with public financials — unusual for a Korean data vendor — plus Korean-language RLHF and instruction data that foreign vendors supply poorly. Held at the cap because per-product contract values and named model customers are not disclosed.", "versionsUpdated": "2026-07-29" }, { "id": "mosaic-ai", "name": "Mosaic AI", "company": "Databricks", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Snowflake Cortex", "LangChain", "LlamaIndex", "CrewAI", "AutoGen", "vLLM", "TensorRT-LLM" ], "features": [ "Suite modules: Mosaic AI Vector Search, Mosaic AI Agent Framework, Mosaic AI Model Serving" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://www.databricks.com/product/artificial-intelligence", "apiLink": null, "aiCoreShare": 87, "sourceUrls": [ "https://www.databricks.com/product/artificial-intelligence" ], "contentReview": "needs-review" }, { "id": "lakera-guard", "name": "Lakera", "company": "Lakera", "region": "global", "tagline": "Runtime guardrails and red-teaming for LLM applications", "description": "Lakera intercepts prompt injection, jailbreaks and data leakage at runtime, with automated red-teaming and a public benchmark for injection detection. It sits in the request path of LLM applications, which is both its value and its latency constraint.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance", "Suite modules: Lakera PINT Benchmark, Lakera Red, Lakera Runtime Protection" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://lakera.ai", "apiLink": null, "aiCoreShare": 75, "sourceUrls": [], "versions": [ { "name": "Lakera", "variants": [ "Lakera PINT Benchmark", "Lakera Red", "Lakera Runtime Protection" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "openrouter", "name": "OpenRouter", "company": "OpenRouter", "region": "global", "tagline": "Unified LLM routing and inference marketplace", "description": "API gateway routing developer requests across 400+ LLMs from OpenAI, Anthropic, Google, Meta and open-source providers through a single OpenAI-compatible endpoint, with model fallback, load balancing and cost-aware routing. TechCrunch reported 8M global users and about 100 trillion tokens processed per month in May 2026 — a 5x increase in six months — alongside a $113M Series B led by CapitalG at roughly $1.3B post-money. Chatroom (multi-model chat UI) and Rankings (leaderboard) are surfaces on the same platform. Bedrock, Vertex AI and Azure OpenAI are suppliers rather than competitors; lock-in comes from routing configuration depth and multi-model workflows built on the API.", "users": "8M+ global users", "userCount": 8000000, "releaseDate": "2023-01", "technicalImpact": 88, "competitors": [ "Portkey", "LiteLLM", "Helicone", "Not Diamond", "Martian" ], "features": [ "400+ LLM coverage via a single API endpoint", "Model fallback and load balancing", "Cost-aware routing and optimization", "OpenAI-compatible API", "Provider-neutral routing across competing model vendors" ], "pricing": { "api": "Pass-through per-token pricing", "enterprise": "Contact sales" }, "link": "https://openrouter.ai", "apiLink": "https://openrouter.ai/docs/quickstart", "aiCoreShare": 84, "scoringRationale": "Raised from 84 to 88. The previous 84 rested on a '$1B annualized revenue in 2025' claim that is wrong by roughly 20x — actual ARR is near $50M (Sacra, Mar 2026); the $1B figure conflates marketplace token spend with OpenRouter's own revenue. The row is nonetheless underscored on its real evidence: 8M product-specific users clears the horizontal-platform cap outright (1M threshold), and 100T tokens/month is quantified adoption at genuine scale. 88 places it above GroqCloud (85, 3M developers) and below Hugging Face Hub (90), whose ecosystem entrenchment is deeper — model IDs are hardcoded in production code globally. It stays under 90 because OpenRouter has no first-party model control and a take-rate marketplace is structurally exposed to provider disintermediation.", "adoptionSignal": "TechCrunch (May 2026): 8M global users, ~100T tokens processed per month (5x growth in six months), 400+ models, $113M Series B led by CapitalG at ~$1.3B post-money. Sacra puts annualized revenue near $50M as of March 2026 — the platform is a pass-through marketplace, so token spend flowing through it is far larger than OpenRouter's own take.", "sourceUrls": [ "https://techcrunch.com/2026/05/26/openrouter-more-than-doubles-valuation-to-1-3b-in-a-year/", "https://sacra.com/c/openrouter/" ] }, { "id": "superb-platform", "name": "Superb Platform", "company": "Superb AI", "region": "korea", "tagline": "Vision AI data management", "description": "One-stop computer vision AI platform for dataset management, model training, and deployment. Y Combinator graduate. Only Korean company in NVIDIA Inception's Physical AI ecosystem. Investors include Samsung, Hyundai, POSCO, Kakao. Planning IPO H1 2026.", "descriptionKo": "Superb Platform은 Superb AI가 제공하는 컴퓨터 비전 AI 원스톱 플랫폼으로, 데이터셋 관리(Superb Curate), 자동 레이블링(Superb Label), 모델 학습 및 배포(Superb Model) 세 모듈로 구성됩니다. YC(와이컴비네이터) 출신이며, NVIDIA Inception의 Physical AI 에코시스템에 참여한 유일한 한국 기업입니다. 삼성·현대차·포스코·카카오가 투자자로 참여하고 있으며, 2026년 상반기 IPO를 준비 중입니다. 컴퓨터 비전 AI 개발 전 주기(데이터 수집→레이블링→학습→배포)를 단일 플랫폼에서 처리할 수 있어, Scale AI·Labelbox 같은 글로벌 경쟁사에 맞서는 한국의 대표적인 데이터·모델 운영 플랫폼입니다.", "users": "Enterprise", "userCount": 0, "releaseDate": "2018-01", "technicalImpact": 77, "competitors": [ "scale-ai", "labelbox", "v7", "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "Dataset management", "Auto-labeling", "Model training", "Computer vision deployment", "Suite modules: Superb Curate, Superb Label, Superb Model", "Developer workflow automation", "Model operations", "Observability", "Secure deployment" ], "pricing": { "free": "Free starter", "pro": "Contact", "enterprise": "Contact Superb AI" }, "link": "https://superb-ai.com", "apiLink": "https://api.superb-ai.com", "aiCoreShare": 83, "sourceUrls": [], "scoringRationale": "2026-07-29 Korean top-tier stress test: reduced 80 -> 77. Full-platform breadth is real, but no customer count, seat volume or revenue is disclosed. Unnamed 'Enterprise' adoption sits at 77." }, { "id": "moreh-ai-framework", "name": "Moreh", "company": "Moreh", "region": "korea", "tagline": "GPU cluster software and virtualization for AI training", "description": "Moreh builds the software layer that makes large GPU and NPU clusters usable for AI training — virtualizing accelerators across nodes so a job can address pooled compute, with support for non-NVIDIA silicon including AMD. It matters to Korean sovereign-AI ambitions because it reduces dependence on a single hardware vendor. Deployment and customer figures are not disclosed.", "users": "Enterprise", "userCount": 0, "releaseDate": "2018-01", "technicalImpact": 77, "competitors": [ "run-ai", "anyscale", "nvidia", "NVIDIA Hopper", "NVIDIA Blackwell", "AMD Instinct", "Cloud TPU v5" ], "features": [ "Heterogeneous GPU support", "Zero code changes", "Large model training", "Korean cloud integration", "Suite modules: Moreh GPU Virtualization", "Developer workflow automation", "Model operations", "Observability", "Secure deployment" ], "pricing": { "enterprise": "Contact Moreh" }, "link": "https://moreh.io", "apiLink": "https://docs.moreh.io", "aiCoreShare": 84, "sourceUrls": [], "versions": [ { "name": "Moreh", "released": "2018-01", "variants": [ "Moreh GPU Virtualization" ] } ], "scoringRationale": "2026-07-29 Korean top-tier stress test: reduced 80 -> 77. Deep systems work and strategic non-NVIDIA support, but no cluster deployments, customer names or scale are disclosed; investor relationships (KT, AMD) are not deployments.", "versionsUpdated": "2026-07-29" }, { "id": "sona-persona", "name": "SONA", "company": "Persona AI", "region": "korea", "tagline": "On-device AI engine (no GPU needed)", "description": "Persona AI's SONA engine enables AI operations without internet or GPU — true edge AI. Two-time CES Innovation Award winner. 9.2% stake held by SK Telecom. Critical for defense, edge computing, and offline environments.", "users": "Enterprise", "userCount": 0, "releaseDate": "2019-01", "technicalImpact": 77, "competitors": [ "qualcomm-ai", "apple-core-ml", "mediatek" ], "features": [ "No GPU required", "Offline AI inference", "Defense-grade", "Ultra low power" ], "pricing": { "enterprise": "Contact Persona AI" }, "link": "https://personaai.co.kr", "apiLink": null, "aiCoreShare": 70, "scoringRationale": "2026-07-29 Korean top-tier stress test: reduced 80 -> 77. On-device niche is real, but the row's own rationale concedes no design wins, shipped-device counts or named OEM programs." }, { "id": "insightvm-remediation-hub", "name": "InsightVM Remediation Hub", "company": "Rapid7", "region": "global", "tagline": "Security analytics and remediation AI", "description": "InsightVM Remediation Hub is Rapid7's security analytics and remediation AI.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://www.rapid7.com/products/insightvm/", "apiLink": null, "aiCoreShare": 68, "sourceUrls": [ "https://www.rapid7.com/products/insightvm/" ], "contentReview": "needs-review" }, { "id": "rapid7-ai-engine", "name": "Rapid7 AI Engine", "company": "Rapid7", "region": "global", "tagline": "Security analytics and remediation AI", "description": "Rapid7 AI Engine is Rapid7's security analytics and remediation AI.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://rapid7.com", "apiLink": null, "aiCoreShare": 73, "contentReview": "needs-review" }, { "id": "cortex-agents", "name": "Cortex Agents", "company": "Snowflake", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-agents", "apiLink": null, "aiCoreShare": 83, "sourceUrls": [ "https://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-agents" ], "contentReview": "needs-review" }, { "id": "vercel-ai-sdk", "name": "Vercel AI SDK", "company": "Vercel", "region": "global", "tagline": "TypeScript toolkit and gateway for building AI applications", "description": "The Vercel AI SDK is a TypeScript library for streaming LLM responses, tool calling and building chat interfaces, paired with a gateway that routes across model providers and observability for AI routes. It is widely used in the Next.js ecosystem, which is Vercel's distribution.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "Developer workflow automation", "Model operations", "Observability", "Secure deployment", "Suite modules: Vercel MCP Adapter", "Suite modules: AI Gateway, Vercel Observability AI" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://vercel.com", "apiLink": "https://sdk.vercel.ai", "aiCoreShare": 86, "sourceUrls": [ "https://vercel.com/ai-gateway" ], "versions": [ { "name": "Vercel AI SDK", "variants": [ "AI Gateway", "Vercel Observability AI" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "cline-mcp", "name": "Cline MCP", "company": "Cline", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://cline.bot", "apiLink": null, "aiCoreShare": 77, "contentReview": "needs-review" }, { "id": "hiddenlayer-model-scanner", "name": "HiddenLayer Model Scanner", "company": "HiddenLayer", "region": "global", "tagline": "AI model security and threat detection", "description": "HiddenLayer Model Scanner is HiddenLayer's AI model security and threat detection.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://hiddenlayer.com", "apiLink": null, "aiCoreShare": 79, "contentReview": "needs-review" }, { "id": "proofpoint-nexusai", "name": "Proofpoint NexusAI", "company": "Proofpoint", "region": "global", "tagline": "AI detection across Proofpoint's threat protection platform", "description": "NexusAI is the detection layer across Proofpoint's email, web and cloud protection products, classifying phishing, business email compromise and malicious content. It ships as part of the platform rather than as a separate purchase.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "CrowdStrike", "Splunk", "Palo Alto Networks", "SentinelOne" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance", "Suite modules: Proofpoint Email Fraud Defense, Proofpoint ZenWeb, Proofpoint Aegis" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://proofpoint.com", "apiLink": null, "aiCoreShare": 84, "sourceUrls": [], "versions": [ { "name": "Proofpoint NexusAI", "variants": [ "Proofpoint Email Fraud Defense", "Proofpoint ZenWeb", "Proofpoint Aegis" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "runpod-serverless", "name": "RunPod Serverless", "company": "RunPod", "region": "global", "tagline": "Autoscaling serverless GPU endpoints for AI inference.", "description": "RunPod Serverless runs AI inference on autoscaling, pay-per-second GPU endpoints — developers package a model or handler, deploy it as an API endpoint, and RunPod scales workers up and down with demand (including scale-to-zero). It is one of RunPod's two core products alongside GPU Cloud and competes with serverless-inference platforms rather than horizontal model APIs. Usage is reported at the RunPod platform level (400K+ developers), not per product.", "users": "400K+ RunPod developers", "userCount": 400000, "releaseDate": "", "technicalImpact": 82, "competitors": [ "replicate", "Modal", "Baseten", "Together AI", "CoreWeave" ], "features": [ "Autoscaling GPU inference endpoints", "Per-second billing and scale-to-zero", "Custom handlers and container deploys", "Flashboot fast cold starts" ], "pricing": { "api": "Per-second usage-based pricing", "enterprise": "Contact sales" }, "link": "https://www.runpod.io/product/serverless", "apiLink": "https://docs.runpod.io", "aiCoreShare": 86, "adoptionSignal": "RunPod grew its developer community from 100K (May 2024) to 400K+ by late 2025 with roughly 10x year-over-year revenue growth, backed by a $20M seed round (May 2024) co-led by Intel Capital and Dell Technologies Capital.", "scoringRationale": "Impact reflects sourced developer-ecosystem leverage (400K+ developers, 10x revenue growth, tier-one infra investors) for a GPU serverless-inference platform. It stays in the low-80s rather than higher because RunPod resells third-party GPU capacity, does not own a frontier model, and per-product usage and revenue are only partially disclosed.", "sourceUrls": [ "https://www.runpod.io/product/serverless", "https://www.intelcapital.com/runpod-raises-20m-in-seed-funding-co-led-by-intel-capital-and-dell-technologies-capital/", "https://sacra.com/c/runpod/" ] }, { "id": "arize-ax", "name": "Arize", "company": "Arize AI", "region": "global", "tagline": "ML and LLM observability with open-source tracing", "description": "Arize monitors models in production — drift, performance and LLM traces — with evaluation tooling and the open-source Phoenix tracing library as the entry point into the commercial platform. Phoenix's OSS adoption is the funnel.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks", "GitHub Copilot", "Cursor", "Windsurf Editor", "Tabnine" ], "features": [ "Developer workflow automation", "Model operations", "Observability", "Secure deployment", "Suite modules: Arize Evaluations, Arize Copilot, Arize Phoenix" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://arize.com", "apiLink": null, "aiCoreShare": 87, "sourceUrls": [], "versions": [ { "name": "Arize", "variants": [ "Arize Evaluations", "Arize Copilot", "Arize Phoenix" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "coderabbit", "name": "CodeRabbit", "company": "CodeRabbit", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "GitHub Copilot", "Cursor", "Windsurf Editor", "Tabnine" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://coderabbit.ai", "apiLink": null, "aiCoreShare": 81, "contentReview": "needs-review" }, { "id": "makinarocks-runway", "name": "Runway (MakinaRocks)", "company": "MakinaRocks", "region": "korea", "tagline": "Enterprise AI operating system for industrial AI", "description": "Runway is MakinaRocks' enterprise AI operating system for scaling and operating specialized AI in industrial environments. It orchestrates AI workloads across enterprise systems with GPU resource optimization, standardized development and operations environments, governance, and closed-network deployment support.", "users": "Enterprise", "userCount": 0, "releaseDate": "2018-01", "technicalImpact": 78, "competitors": [ "datarobot-ai-platform", "databricks", "c3-ai-suite", "palantir-aip", "Dataiku" ], "features": [ "Industrial AI orchestration", "GPU resource optimization", "Standardized AI development and operations environment", "Enterprise governance and audit controls", "Closed-network deployment support", "Use cases across manufacturing, defense, public sector, and finance", "Suite modules: MakinaRocks Runway, MakinaRocks Industrial AI, MakinaRocks Manufacturing Optimization" ], "pricing": { "enterprise": "Contact MakinaRocks" }, "link": "https://www.makinarocks.ai/product/runway/", "apiLink": null, "aiCoreShare": 91, "adoptionSignal": "MakinaRocks' Runway product page cites named enterprise and public-sector references including Hyundai, Hanwha Systems, Samsung Electro-Mechanics, K-water, Korea Insurance Development Institute, and Korea Credit Bureau, but Runway-specific customer count is not disclosed.", "sourceUrls": [ "https://www.makinarocks.ai/product/runway/" ], "scoringRationale": "2026-07-29 Korean top-tier stress test: reduced 80 -> 78. Public company (KOSDAQ, May 2026) with press-documented manufacturing references; product-specific deployment counts undisclosed." }, { "id": "pan-os-ai-runtime-security", "name": "PAN-OS AI Runtime Security", "company": "Palo Alto Networks", "region": "global", "tagline": "Security operations and cloud security AI", "description": "PAN-OS AI Runtime Security is Palo Alto Networks' security operations and cloud security AI.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "CrowdStrike", "Splunk", "SentinelOne", "Darktrace" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://www.paloaltonetworks.com/prisma/prisma-ai-runtime-security", "apiLink": null, "aiCoreShare": 73, "sourceUrls": [ "https://www.paloaltonetworks.com/prisma/prisma-ai-runtime-security" ], "contentReview": "needs-review" }, { "id": "singularity-hyperautomation", "name": "Singularity Hyperautomation", "company": "SentinelOne", "region": "global", "tagline": "Autonomous cybersecurity operations AI", "description": "Singularity Hyperautomation is SentinelOne's autonomous cybersecurity operations AI.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "CrowdStrike", "Splunk", "Palo Alto Networks", "Darktrace" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://www.sentinelone.com/platform/singularity-hyperautomation/", "apiLink": null, "aiCoreShare": 75, "sourceUrls": [ "https://www.sentinelone.com/platform/singularity-hyperautomation/" ], "contentReview": "needs-review" }, { "id": "aimmo-data-labeling-platform", "name": "AIMMO", "company": "AIMMO", "region": "korea", "tagline": "Autonomous-driving and vision data labelling", "description": "AIMMO produces annotated training data for autonomous driving and industrial vision — 3D point-cloud, sensor-fusion and video labelling — offered as a platform and as a managed ground-truth service. Korean and Japanese mobility programmes are the buyers. Contract values and volumes are not disclosed.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 77, "competitors": [ "Scale AI", "Labelbox", "V7", "Surge AI", "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "Developer workflow automation", "Model operations", "Observability", "Secure deployment", "Suite modules: AIMMO GTaaS", "Suite modules: AIMMO Enterprise" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://aimmo.co.kr", "apiLink": null, "aiCoreShare": 86, "sourceUrls": [], "versions": [ { "name": "AIMMO", "variants": [ "AIMMO GTaaS" ] } ], "scoringRationale": "77 reflects specialisation in 3D point-cloud and sensor-fusion annotation for autonomous driving, which is harder than general labelling and has fewer credible suppliers. Held below 80 because contract values, volumes and named mobility programmes are not disclosed.", "versionsUpdated": "2026-07-29" }, { "id": "baseten-model-apis", "name": "Baseten Model APIs", "company": "Baseten", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://baseten.co", "apiLink": null, "aiCoreShare": 81, "contentReview": "needs-review" }, { "id": "helicone-gateway", "name": "Helicone Gateway", "company": "Helicone", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://helicone.ai", "apiLink": null, "aiCoreShare": 80, "contentReview": "needs-review" }, { "id": "hiddenlayer-aisec-platform", "name": "HiddenLayer AISec Platform", "company": "HiddenLayer", "region": "global", "tagline": "AI model security and threat detection", "description": "HiddenLayer AISec Platform is HiddenLayer's AI model security and threat detection.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://hiddenlayer.com", "apiLink": null, "aiCoreShare": 79, "contentReview": "needs-review" }, { "id": "modal", "name": "Modal", "company": "Modal", "region": "global", "tagline": "Serverless GPU compute for AI workloads", "description": "Modal runs Python functions on serverless GPUs with sub-second cold starts, persistent volumes and sandboxed execution, aimed at teams who want to run training and inference without managing clusters. Developer ergonomics are the differentiator.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 79, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "Developer workflow automation", "Model operations", "Observability", "Secure deployment", "Suite modules: Modal GPUs, Modal Volumes, Modal Sandboxes" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://modal.com", "apiLink": null, "aiCoreShare": 81, "sourceUrls": [], "versions": [ { "name": "Modal", "variants": [ "Modal GPUs", "Modal Volumes", "Modal Sandboxes" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "baseten-truss", "name": "Baseten Truss", "company": "Baseten", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://baseten.co", "apiLink": null, "aiCoreShare": 82, "contentReview": "needs-review" }, { "id": "datalore-ai", "name": "Datalore AI", "company": "JetBrains", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://jetbrains.com/ai", "apiLink": null, "aiCoreShare": 79, "contentReview": "needs-review" }, { "id": "cortex-copilot", "name": "Cortex Copilot", "company": "Palo Alto Networks", "region": "global", "tagline": "Security operations and cloud security AI", "description": "Cortex Copilot is Palo Alto Networks' security operations and cloud security AI.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "GitHub Copilot", "Cursor", "Windsurf Editor", "Tabnine" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://www.paloaltonetworks.com/cortex/cortex-agentic-assistant", "apiLink": null, "aiCoreShare": 73, "sourceUrls": [ "https://www.paloaltonetworks.com/cortex/cortex-agentic-assistant" ], "contentReview": "needs-review" }, { "id": "portkey-ai-gateway", "name": "Portkey AI Gateway", "company": "Portkey", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://portkey.ai", "apiLink": null, "aiCoreShare": 80, "contentReview": "needs-review" }, { "id": "qodo-gen", "name": "Qodo", "company": "Qodo", "region": "global", "tagline": "Test generation and code review agents", "description": "Qodo generates tests, reviews pull requests and runs code-quality agents, positioning on correctness and coverage rather than raw code generation — the part of the workflow Copilot-style completion leaves unaddressed.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "Developer workflow automation", "Model operations", "Observability", "Secure deployment", "Suite modules: Qodo Cover, Qodo Command, Qodo Merge" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://qodo.ai", "apiLink": null, "aiCoreShare": 87, "sourceUrls": [], "versions": [ { "name": "Qodo", "variants": [ "Qodo Cover", "Qodo Command", "Qodo Merge" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "tenable-cloud-security-ai", "name": "Tenable Cloud Security AI", "company": "Tenable", "region": "global", "tagline": "Exposure management AI", "description": "Tenable Cloud Security AI is Tenable's exposure management AI.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "CrowdStrike", "Splunk", "Palo Alto Networks", "SentinelOne" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://tenable.com", "apiLink": null, "aiCoreShare": 79, "contentReview": "needs-review" }, { "id": "galileo-observe", "name": "Galileo", "company": "Galileo", "region": "global", "tagline": "Evaluation, monitoring and guardrails for LLM applications", "description": "Galileo evaluates and monitors LLM applications in development and production, with its own Luna evaluation models scoring hallucination and quality without a frontier model in the loop, plus runtime guardrails.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "Developer workflow automation", "Model operations", "Observability", "Secure deployment", "Suite modules: Galileo Evaluate, Galileo Luna, Galileo Protect" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://galileo.ai", "apiLink": null, "aiCoreShare": 83, "sourceUrls": [], "versions": [ { "name": "Galileo", "variants": [ "Galileo Evaluate", "Galileo Luna", "Galileo Protect" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "runpod-pods", "name": "RunPod GPU Cloud", "company": "RunPod", "region": "global", "tagline": "On-demand and reserved GPU pods for AI training and inference.", "description": "RunPod GPU Cloud (Pods) provisions on-demand and reserved GPU instances — from consumer cards to H100/H200-class accelerators — that developers can spin up in seconds for training, fine-tuning, and inference, with persistent storage and per-second billing. It is one of RunPod's two core products alongside Serverless and competes with neocloud GPU providers. Usage is reported at the RunPod platform level (400K+ developers), not per product.", "users": "400K+ RunPod developers", "userCount": 400000, "releaseDate": "", "technicalImpact": 82, "competitors": [ "CoreWeave", "Lambda", "Vast.ai", "Together AI", "vessl-ai" ], "features": [ "On-demand and reserved GPU pods", "Per-second billing", "Persistent network storage", "Global multi-region GPU availability" ], "pricing": { "api": "Per-second usage-based pricing", "enterprise": "Contact sales" }, "link": "https://www.runpod.io/product/cloud-gpus", "apiLink": "https://docs.runpod.io", "aiCoreShare": 76, "adoptionSignal": "RunPod grew its developer community from 100K (May 2024) to 400K+ by late 2025 with roughly 10x year-over-year revenue growth, backed by a $20M seed round (May 2024) co-led by Intel Capital and Dell Technologies Capital.", "scoringRationale": "Impact reflects sourced developer-ecosystem leverage (400K+ developers, 10x revenue growth, tier-one infra investors) for a GPU cloud. It stays in the low-80s rather than higher because RunPod resells third-party GPU capacity, does not own a frontier model, and per-product usage and revenue are only partially disclosed.", "sourceUrls": [ "https://www.runpod.io/product/cloud-gpus", "https://www.intelcapital.com/runpod-raises-20m-in-seed-funding-co-led-by-intel-capital-and-dell-technologies-capital/", "https://sacra.com/c/runpod/" ] }, { "id": "cody", "name": "Cody", "company": "Sourcegraph", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://sourcegraph.com/cody", "apiLink": null, "aiCoreShare": 87, "sourceUrls": [ "https://sourcegraph.com/cody" ], "contentReview": "needs-review" }, { "id": "wiz-cloud-security-graph", "name": "Wiz", "company": "Wiz", "region": "global", "tagline": "Cloud security graph with AI posture management", "description": "Wiz maps cloud environments into a graph of resources, identities and vulnerabilities to prioritise real attack paths over raw alert volume, extending to AI model and pipeline posture. Acquired by Google in a deal that closed March 2026.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance", "Suite modules: Wiz Code AI, Wiz Runtime Sensor AI, Wiz AI-SPM" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://wiz.io", "apiLink": null, "aiCoreShare": 79, "sourceUrls": [], "versions": [ { "name": "Wiz", "variants": [ "Wiz Code AI", "Wiz Runtime Sensor AI", "Wiz AI-SPM" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "braintrust-data-platform", "name": "Braintrust", "company": "Braintrust", "region": "global", "tagline": "Evaluation and observability for LLM product teams", "description": "Braintrust runs evaluations against versioned datasets, logs production traces and proxies model calls so teams can compare prompts and models on real data. It targets the gap between prototype and reliable shipped LLM feature.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 79, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "Developer workflow automation", "Model operations", "Observability", "Secure deployment", "Suite modules: Braintrust, Braintrust Evals, Braintrust Proxy" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://braintrust.dev", "apiLink": null, "aiCoreShare": 87, "sourceUrls": [], "versions": [ { "name": "Braintrust", "variants": [ "Braintrust", "Braintrust Evals", "Braintrust Proxy" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "credo-ai-governance-platform", "name": "Credo AI", "company": "Credo AI", "region": "global", "tagline": "AI governance, policy and regulatory readiness", "description": "Credo AI tracks AI systems against internal policy and external regulation — EU AI Act, NIST AI RMF — with policy packs, risk registers and evidence collection for audit. Regulatory deadlines rather than model quality drive the purchase.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 79, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance", "Suite modules: Credo AI Regulatory Readiness, Credo AI Policy Packs, Credo AI Risk Center" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://credo.ai", "apiLink": null, "aiCoreShare": 77, "sourceUrls": [], "versions": [ { "name": "Credo AI", "variants": [ "Credo AI Regulatory Readiness", "Credo AI Policy Packs", "Credo AI Risk Center" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "activeai-security-platform", "name": "Darktrace", "company": "Darktrace", "region": "global", "tagline": "Self-learning network anomaly detection and response", "description": "Darktrace models normal behaviour on a customer network and flags deviations without signatures, extending into automated response and recovery. The unsupervised approach is both the differentiator and the source of long-running debate about false positives.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 79, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance", "Suite modules: Darktrace HEAL, Darktrace PREVENT, Darktrace DETECT" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://www.darktrace.com/platform", "apiLink": null, "aiCoreShare": 80, "sourceUrls": [ "https://www.darktrace.com/platform" ], "versions": [ { "name": "Darktrace", "variants": [ "Darktrace HEAL", "Darktrace PREVENT", "Darktrace DETECT" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "datumo-ai-qa", "name": "Datumo AI QA", "company": "Datumo", "region": "korea", "tagline": "AI model evaluation and monitoring", "description": "Datumo's platform helps enterprises test, monitor, and improve model behavior without deep engineering. Korea's answer to Scale AI and Arize. Raised $15.5M in 2025 for AI governance and validation.", "users": "Enterprise", "userCount": 0, "releaseDate": "2018-01", "technicalImpact": 79, "competitors": [ "scale-ai", "arize", "arthur-ai" ], "features": [ "Model evaluation", "Behavioral monitoring", "No-code testing", "AI governance compliance" ], "pricing": { "enterprise": "Contact Datumo" }, "link": "https://datumo.com", "apiLink": null, "aiCoreShare": 79, "scoringRationale": "79 reflects AI-assisted software QA — test generation and defect detection — which is a different buyer from Datumo's data and evaluation business and is kept separate for that reason. No customer count or deployment figure is disclosed." }, { "id": "giskard", "name": "Giskard", "company": "Giskard", "region": "global", "tagline": "Open-source testing and vulnerability scanning for AI models", "description": "Giskard scans ML and LLM systems for vulnerabilities — hallucination, bias, prompt injection, robustness — and generates test suites that run in CI, with a hub for collaborative review. Open-source distribution is the adoption path.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 79, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks", "GPT-5", "claude", "Gemini 2.5", "Grok 4" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance", "Suite modules: Giskard LLM Scan, Giskard Evaluation Tests, Giskard Hub" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://giskard.ai", "apiLink": null, "aiCoreShare": 86, "sourceUrls": [], "versions": [ { "name": "Giskard", "variants": [ "Giskard LLM Scan", "Giskard Evaluation Tests", "Giskard Hub" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "qodana-ai", "name": "Qodana AI", "company": "JetBrains", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 79, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://jetbrains.com/ai", "apiLink": null, "aiCoreShare": 80, "contentReview": "needs-review" }, { "id": "siliconflow-inference", "name": "SiliconFlow MaaS", "company": "SiliconFlow", "region": "china", "tagline": "Hosted model API and MaaS platform", "description": "SiliconFlow MaaS provides hosted model APIs for language, speech, image, and video models, plus reserved inference capacity and enterprise deployment options. Current site copy highlights DeepSeek-V4 and GLM-5.1 services.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 79, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://www.siliconflow.cn/", "apiLink": "https://api-docs.siliconflow.cn/", "aiCoreShare": 89, "sourceUrls": [ "https://www.siliconflow.cn/", "https://api-docs.siliconflow.cn/" ] }, { "id": "together-inference", "name": "Together AI", "company": "Together AI", "region": "global", "tagline": "Inference, fine-tuning and GPU clusters for open models", "description": "Together AI serves open-weight models through fast inference endpoints, with fine-tuning and dedicated GPU clusters for teams needing reserved capacity. It competes on price and throughput against the hyperscalers for open-model workloads.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 79, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks", "CoreWeave", "Lambda Labs", "RunPod", "Vast.ai" ], "features": [ "Developer workflow automation", "Model operations", "Observability", "Secure deployment", "Suite modules: Together Dedicated Endpoints, Together Fine-tuning, Together GPU Clusters" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://together.ai", "apiLink": null, "aiCoreShare": 87, "sourceUrls": [], "versions": [ { "name": "Together AI", "variants": [ "Together Dedicated Endpoints", "Together Fine-tuning", "Together GPU Clusters" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "whylabs-ai-observability", "name": "WhyLabs", "company": "WhyLabs", "region": "global", "tagline": "Data and ML monitoring with privacy-preserving telemetry", "description": "WhyLabs monitors data quality, drift and model performance using statistical profiles rather than raw data, which lets it operate where sending records to a vendor is not permitted. Guardrails extends the same approach to LLM inputs and outputs.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 79, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance", "Suite modules: WhyLabs Guardrails, WhyLabs Monitoring, LangKit" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://whylabs.ai", "apiLink": null, "aiCoreShare": 86, "sourceUrls": [ "https://whylabs.ai/langkit/" ], "versions": [ { "name": "WhyLabs", "variants": [ "WhyLabs Guardrails", "WhyLabs Monitoring", "LangKit" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "falcon-foundry", "name": "Falcon Foundry", "company": "CrowdStrike", "region": "global", "tagline": "Cybersecurity AI analyst and platform", "description": "Falcon Foundry is CrowdStrike's cybersecurity AI analyst and platform.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "Splunk", "Palo Alto Networks", "SentinelOne", "Darktrace" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://www.crowdstrike.com/en-us/platform/next-gen-siem/falcon-foundry/", "apiLink": null, "aiCoreShare": 75, "sourceUrls": [ "https://www.crowdstrike.com/en-us/platform/next-gen-siem/falcon-foundry/" ], "contentReview": "needs-review" }, { "id": "hiddenlayer-detection-and-response", "name": "HiddenLayer Detection and Response", "company": "HiddenLayer", "region": "global", "tagline": "AI model security and threat detection", "description": "HiddenLayer Detection and Response is HiddenLayer's AI model security and threat detection.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://hiddenlayer.com", "apiLink": null, "aiCoreShare": 84, "contentReview": "needs-review" }, { "id": "sentry-suspect-commits-ai", "name": "Sentry AI", "company": "Sentry", "region": "global", "tagline": "AI triage and autofix over application errors", "description": "Sentry's AI summarises error issues, identifies the commit likely responsible and proposes fixes, working from the stack traces and release data Sentry already collects. The error corpus is the asset.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance", "Suite modules: Sentry Issue Summary, Sentry Autofix, Seer" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://sentry.io", "apiLink": null, "aiCoreShare": 82, "sourceUrls": [ "https://sentry.io/product/seer/" ], "versions": [ { "name": "Sentry AI", "variants": [ "Sentry Issue Summary", "Sentry Autofix", "Seer" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "webcontainer-ai", "name": "WebContainer AI", "company": "StackBlitz", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://bolt.new", "apiLink": null, "aiCoreShare": 80, "contentReview": "needs-review" }, { "id": "vectra-mxdr", "name": "Vectra AI", "company": "Vectra AI", "region": "global", "tagline": "AI-driven network detection and response", "description": "Vectra detects attacker behaviour across network, identity and cloud telemetry, prioritising signals by attack progression rather than alert severity, delivered as a platform and as managed detection and response.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance", "Suite modules: Attack Signal Intelligence, Vectra MDR AI" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://vectra.ai", "apiLink": null, "aiCoreShare": 74, "sourceUrls": [ "https://www.vectra.ai/products/our-ai" ], "versions": [ { "name": "Vectra AI", "variants": [ "Attack Signal Intelligence", "Vectra MDR AI" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "falcon-exposure-management-ai", "name": "Falcon Exposure Management AI", "company": "CrowdStrike", "region": "global", "tagline": "Cybersecurity AI analyst and platform", "description": "Falcon Exposure Management AI is CrowdStrike's cybersecurity AI analyst and platform.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 77, "competitors": [ "Splunk", "Palo Alto Networks", "SentinelOne", "Darktrace" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://www.crowdstrike.com/en-us/platform/exposure-management/", "apiLink": null, "aiCoreShare": 75, "sourceUrls": [ "https://www.crowdstrike.com/en-us/platform/exposure-management/" ], "contentReview": "needs-review" }, { "id": "deepeval", "name": "DeepEval", "company": "DeepEval", "region": "global", "tagline": "Open-source LLM evaluation and red-teaming framework", "description": "DeepEval runs unit-test-style evaluations for LLM applications — hallucination, relevance, bias metrics plus red-teaming and standard benchmarks — with Confident AI as the hosted platform on top. Pytest-like ergonomics are the adoption hook.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 77, "competitors": [ "LangSmith", "LangFuse", "Arize", "Helicone" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance", "Suite modules: DeepEval Red Teaming, DeepEval Benchmarks, Confident AI" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://deepeval.com", "apiLink": null, "aiCoreShare": 87, "sourceUrls": [ "https://www.confident-ai.com/" ], "versions": [ { "name": "DeepEval", "variants": [ "DeepEval Red Teaming", "DeepEval Benchmarks", "Confident AI" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "fireworks-serverless-inference", "name": "Fireworks AI", "company": "Fireworks AI", "region": "global", "tagline": "Inference platform for open and fine-tuned models", "description": "Fireworks AI serves open-weight and custom models through serverless endpoints, dedicated deployments and a fine-tuning service. Those are pricing and deployment tiers of one inference platform, not separate products.", "users": "Enterprise developers", "userCount": 0, "releaseDate": "2023-01-01", "technicalImpact": 79, "competitors": [ "together-inference", "groqcloud", "replicate", "aws-bedrock", "gemini-enterprise-agent-platform", "modal" ], "features": [ "200+ hosted open-weight models", "FireAttention custom GPU inference kernels", "OpenAI-compatible REST API", "FireFunction structured output", "Suite modules: Fireworks Dedicated Deployments, Fireworks Fine-tuning", "Reserved GPU capacity", "Latency SLA guarantees", "Dedicated model isolation", "High-throughput enterprise inference", "LoRA and full fine-tuning", "Llama, Mistral, Mixtral base models", "Managed GPU training infrastructure", "One-click deploy to Fireworks endpoints" ], "pricing": { "api": "Pay-per-token with free tier", "enterprise": "Contact sales" }, "link": "https://fireworks.ai", "apiLink": "https://readme.fireworks.ai/docs/quickstart", "aiCoreShare": 86, "adoptionSignal": "Production deployments at LinkedIn, Cisco, and DoorDash; preferred open-model inference API for AI-native startups needing cost-efficient alternatives to proprietary APIs.", "scoringRationale": "Raised to 79 (matching Together Inference) reflecting named enterprise customers and differentiated FireAttention inference technology. No disclosed user/revenue counts prevent scoring above the peer tier. Commoditization risk from hyperscaler open-model hosting is the primary ceiling.", "sourceUrls": [ "https://fireworks.ai/models", "https://fireworks.ai/blog/fireworks-raises-52m-series-b", "https://fireworks.ai/dedicated-deployments", "https://fireworks.ai/fine-tuning" ], "versions": [ { "name": "Fireworks AI", "released": "2023-01-01", "variants": [ "Fireworks Dedicated Deployments", "Fireworks Fine-tuning" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "pinecone-serverless", "name": "Pinecone Serverless", "company": "Pinecone", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 77, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://pinecone.io", "apiLink": null, "aiCoreShare": 83, "contentReview": "needs-review" }, { "id": "helicone", "name": "Helicone", "company": "Helicone", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://helicone.ai", "apiLink": null, "aiCoreShare": 79, "contentReview": "needs-review" }, { "id": "jetbrains-ai-assistant", "name": "JetBrains AI", "company": "JetBrains", "region": "global", "tagline": "AI assistant and coding agent inside JetBrains IDEs", "description": "JetBrains AI provides code completion, explanation and refactoring inside IntelliJ, PyCharm and the rest of the IDE family, with Junie as the autonomous coding agent. Distribution is the installed IDE base, which is substantial among professional JVM and Python developers.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "Developer workflow automation", "Model operations", "Observability", "Secure deployment", "Suite modules: Junie" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://jetbrains.com/ai", "apiLink": null, "aiCoreShare": 79, "sourceUrls": [], "versions": [ { "name": "JetBrains AI", "variants": [ "Junie" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "litellm-router", "name": "LiteLLM", "company": "LiteLLM", "region": "global", "tagline": "Unified proxy and router across LLM providers", "description": "LiteLLM gives one OpenAI-compatible interface across 100+ model providers, with a proxy server, routing, fallbacks, spend tracking and budget limits per key. It is the piece teams reach for when they need to avoid provider lock-in.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "GPT-5", "claude", "Gemini 2.5", "Grok 4" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance", "Suite modules: LiteLLM Proxy, LiteLLM Gateway, LiteLLM Budget Manager" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://litellm.ai", "apiLink": null, "aiCoreShare": 90, "sourceUrls": [], "versions": [ { "name": "LiteLLM", "variants": [ "LiteLLM Proxy", "LiteLLM Gateway", "LiteLLM Budget Manager" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "ragas", "name": "Ragas", "company": "Ragas", "region": "global", "tagline": "Open-source evaluation framework for RAG systems", "description": "Ragas scores retrieval-augmented generation on faithfulness, answer relevance and context precision, and generates synthetic test sets so teams can evaluate without hand-labelled data. It became a default reference metric set for RAG evaluation.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "Glean", "Cohere RAG", "Vectara", "Mendable" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance", "Suite modules: Ragas RAG Evaluation, Ragas Metrics, Ragas Testset Generation" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://ragas.io", "apiLink": null, "aiCoreShare": 79, "sourceUrls": [], "versions": [ { "name": "Ragas", "variants": [ "Ragas RAG Evaluation", "Ragas Metrics", "Ragas Testset Generation" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "continue", "name": "Continue", "company": "Continue", "region": "global", "tagline": "Continue's open-source AI coding assistant for IDEs", "description": "Continue is an open-source AI coding assistant for VS Code and JetBrains, with a self-hosted server and a hub for sharing assistant configurations. Those are deployment and distribution surfaces of one product.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "Developer workflow automation", "Model operations", "Observability", "Secure deployment", "Suite modules: Continue Server, Continue Hub" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://continue.dev", "apiLink": null, "aiCoreShare": 85, "sourceUrls": [], "versions": [ { "name": "Continue", "variants": [ "Continue Server", "Continue Hub" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "fiddler-ai-observability", "name": "Fiddler AI", "company": "Fiddler AI", "region": "global", "tagline": "Model monitoring, explainability and LLM guardrails", "description": "Fiddler monitors models in production for drift and performance, explains individual predictions for regulated review, and scores LLM outputs for hallucination and safety. Explainability for audit is its distinguishing angle.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 75, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks", "LangSmith", "LangFuse", "Arize", "Helicone" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance", "Suite modules: Fiddler Trust Service, Fiddler LLM Monitoring, Fiddler Explainable AI" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://fiddler.ai", "apiLink": null, "aiCoreShare": 86, "sourceUrls": [], "versions": [ { "name": "Fiddler AI", "variants": [ "Fiddler Trust Service", "Fiddler LLM Monitoring", "Fiddler Explainable AI" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "gx-core", "name": "GX", "company": "Great Expectations", "region": "global", "tagline": "Great Expectations data-quality validation, open-source and cloud", "description": "GX is Great Expectations' data-quality validation framework, available as the open-source GX Core library and the managed GX Cloud service. The two are packaging of one product.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 75, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance", "Suite modules: GX Cloud" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://docs.greatexpectations.io/docs/core/introduction/", "apiLink": null, "aiCoreShare": 69, "sourceUrls": [ "https://docs.greatexpectations.io/docs/core/introduction/", "https://greatexpectations.io/gx-cloud/" ], "versions": [ { "name": "GX", "variants": [ "GX Cloud" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "langsmith", "name": "LangSmith", "company": "LangChain", "region": "global", "tagline": "LLM observability, tracing, and evaluation platform", "description": "LangSmith is LangChain's commercial platform for tracing, evaluating, and monitoring LLM applications in production. It captures full chain traces, supports human and automated evaluation, and integrates directly with LangChain and LangGraph workflows. Competes with Weights & Biases Weave, Arize AI, and Helicone for the LLM observability market.", "users": "Not disclosed", "userCount": 0, "releaseDate": "2023-06", "technicalImpact": 75, "competitors": [ "Weights & Biases", "Arize AI", "Helicone", "braintrust-data-platform" ], "features": [ "Full LLM trace capture", "Human and automated evaluation", "Dataset management", "Production monitoring" ], "pricing": { "free": "Free tier available", "enterprise": "Contact sales" }, "link": "https://www.langchain.com/langsmith", "apiLink": "https://docs.smith.langchain.com", "aiCoreShare": 85, "sourceUrls": [ "https://www.langchain.com/langsmith" ] }, { "id": "portkey-observability", "name": "Portkey Observability", "company": "Portkey", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 75, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://portkey.ai", "apiLink": null, "aiCoreShare": 84, "contentReview": "needs-review" }, { "id": "pydanticai", "name": "PydanticAI", "company": "Pydantic", "region": "global", "tagline": "Typed agent and validation tooling", "description": "PydanticAI is Pydantic's typed agent and validation tooling.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 75, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://pydantic.dev", "apiLink": null, "aiCoreShare": 82, "contentReview": "needs-review" }, { "id": "siliconflow-model-deployment", "name": "SiliconFlow Private Deployment", "company": "SiliconFlow", "region": "china", "tagline": "Enterprise private deployment for model-serving workloads", "description": "SiliconFlow Private Deployment covers enterprise model performance optimization, deployment, operations, and private/cloud-hybrid inference workflows.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 75, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://www.siliconflow.cn/", "apiLink": "https://api-docs.siliconflow.cn/", "aiCoreShare": 82, "sourceUrls": [ "https://www.siliconflow.cn/", "https://api-docs.siliconflow.cn/" ] }, { "id": "sourcegraph-batch-changes", "name": "Sourcegraph Batch Changes", "company": "Sourcegraph", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 75, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://sourcegraph.com", "apiLink": null, "aiCoreShare": 79, "contentReview": "needs-review" }, { "id": "sourcegraph-code-search", "name": "Sourcegraph Code Search", "company": "Sourcegraph", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 75, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://sourcegraph.com", "apiLink": null, "aiCoreShare": 79, "contentReview": "needs-review" }, { "id": "bits-ai", "name": "Datadog AI", "company": "Datadog", "region": "global", "tagline": "AI across Datadog's observability and security platform", "description": "Datadog's AI covers anomaly detection over metrics and logs, an assistant for investigating incidents, LLM application tracing and security alert triage — all inside its observability platform, on the telemetry customers already send.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "CrowdStrike", "Splunk", "Palo Alto Networks", "SentinelOne", "LangSmith", "LangFuse", "Arize", "Helicone", "Superhuman", "Shortwave" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance", "Suite modules: Watchdog, Datadog LLM Observability, Datadog Security Inbox AI" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://docs.datadoghq.com/bits_ai/", "apiLink": null, "aiCoreShare": 79, "sourceUrls": [ "https://docs.datadoghq.com/bits_ai/", "https://www.datadoghq.com/product/platform/watchdog/" ], "versions": [ { "name": "Datadog AI", "variants": [ "Watchdog", "Datadog LLM Observability", "Datadog Security Inbox AI" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "tenable-one", "name": "Tenable One", "company": "Tenable", "region": "global", "tagline": "Exposure management AI", "description": "Tenable One is Tenable's exposure management AI.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 74, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://tenable.com", "apiLink": null, "aiCoreShare": 78, "contentReview": "needs-review" }, { "id": "pinecone", "name": "Pinecone", "company": "Pinecone", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 73, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://pinecone.io", "apiLink": null, "aiCoreShare": 83, "contentReview": "needs-review" }, { "id": "pinecone-assistant", "name": "Pinecone Assistant", "company": "Pinecone", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 73, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://pinecone.io", "apiLink": null, "aiCoreShare": 81, "contentReview": "needs-review" }, { "id": "bolt-new", "name": "Bolt.new", "company": "StackBlitz", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 73, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://bolt.new", "apiLink": null, "aiCoreShare": 87, "contentReview": "needs-review" }, { "id": "aider", "name": "Aider", "company": "Aider", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 72, "competitors": [ "GitHub Copilot", "Cursor", "Windsurf Editor", "Tabnine" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://aider.chat", "apiLink": null, "aiCoreShare": 87, "contentReview": "needs-review" }, { "id": "aider-architect-mode", "name": "Aider Architect Mode", "company": "Aider", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 72, "competitors": [ "GitHub Copilot", "Cursor", "Windsurf Editor", "Tabnine" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://aider.chat", "apiLink": null, "aiCoreShare": 83, "contentReview": "needs-review" }, { "id": "colo-ai", "name": "COLO AI", "company": "Colosseum", "region": "korea", "tagline": "AI logistics management platform", "description": "Colosseum's 4PL AI logistics platform for order management, warehouse operations, and transportation optimization. 2025 Pre-Unicorn. Expanding to US, Japan, Southeast Asia.", "users": "Enterprise", "userCount": 0, "releaseDate": "2019-01", "technicalImpact": 72, "competitors": [ "flexport", "convoy", "fourkites" ], "features": [ "Order management", "Warehouse AI", "Transportation optimization", "Multi-country expansion" ], "pricing": { "enterprise": "Contact Colosseum" }, "link": "https://colosseum.global", "apiLink": null, "aiCoreShare": 72, "scoringRationale": "72 reflects a Korean AI product with a defined workflow but no disclosed adoption, customers or revenue. It sits below the default tier pending evidence of production use." }, { "id": "dynatrace-grail-ai", "name": "Dynatrace Grail AI", "company": "Dynatrace", "region": "global", "tagline": "Causal observability and automation AI", "description": "Dynatrace Grail AI is Dynatrace's causal observability and automation AI.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 72, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://dynatrace.com", "apiLink": null, "aiCoreShare": 79, "contentReview": "needs-review" }, { "id": "elastic-search-relevance-engine", "name": "Elastic AI", "company": "Elastic", "region": "global", "tagline": "AI across Elastic's search, observability and security stack", "description": "Elastic's AI layer adds vector search and semantic relevance to the search engine, anomaly detection to observability, threat classification to security, and an assistant across all three — running on data customers already index in Elasticsearch.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "CrowdStrike", "Splunk", "Palo Alto Networks", "SentinelOne" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance", "Suite modules: Elastic Observability AI, Elastic Security AI, Elastic AI Assistant" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://elastic.co", "apiLink": null, "aiCoreShare": 72, "sourceUrls": [], "versions": [ { "name": "Elastic AI", "variants": [ "Elastic Observability AI", "Elastic Security AI", "Elastic AI Assistant" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "insightcloudsec-ai", "name": "InsightCloudSec AI", "company": "Rapid7", "region": "global", "tagline": "Security analytics and remediation AI", "description": "InsightCloudSec AI is Rapid7's security analytics and remediation AI.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 72, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://www.rapid7.com/products/insightcloudsec/", "apiLink": null, "aiCoreShare": 71, "sourceUrls": [ "https://www.rapid7.com/products/insightcloudsec/" ], "contentReview": "needs-review" }, { "id": "vessl-ai-platform", "name": "VESSL MLOps Platform", "company": "VESSL AI", "region": "korea", "tagline": "VESSL's MLOps platform for training, serving, and pipelines.", "description": "VESSL MLOps Platform is VESSL AI's original end-to-end MLOps product for experiment tracking, model training orchestration, serving, and deployment pipelines across managed and multi-cloud GPU capacity. It bundles VESSL Run (jobs), VESSL Serve (model serving), and VESSL Pipelines (ML workflows). Named customers include Hyundai Motor's autonomous-driving team, Upstage, Oracle, Yanolja, KAIST, Seoul National University, and Scatter Lab.", "descriptionKo": "VESSL MLOps Platform은 VESSL AI의 초기 핵심 제품으로, 실험 추적부터 모델 학습 오케스트레이션, 서빙, 배포 파이프라인까지 관리형·멀티클라우드 GPU 위에서 엔드투엔드로 다루는 MLOps 플랫폼입니다. 작업 실행을 담당하는 VESSL Run, 모델 서빙을 위한 VESSL Serve, ML 워크플로를 위한 VESSL Pipelines를 하나로 묶습니다. 현대자동차 자율주행팀, 업스테이지, 오라클, 야놀자, KAIST, 서울대학교, 스캐터랩 등이 도입 고객으로 알려져 있으며, 창업자 안재만 대표는 포브스 2025 '주목할 AI 창업자'에 선정되었습니다. 전체 고객 수와 매출은 공개되지 않았습니다.", "users": "Named enterprise & research customers (count undisclosed)", "userCount": 0, "releaseDate": "2020-01", "technicalImpact": 78, "adoptionSignal": "VESSL's MLOps platform is used by named enterprises and research institutions including Hyundai Motor's autonomous-driving team, Upstage, Oracle, Yanolja, KAIST, Seoul National University, and Scatter Lab; CEO Jaeman Ahn was named to Forbes' 2025 list of AI founders to watch.", "scoringRationale": "Impact reflects sourced named enterprise and research adoption (Hyundai, Upstage, Oracle, KAIST) and a durable end-to-end MLOps workflow, not a disclosed user count. It stays mid-tier because total customer numbers and revenue are undisclosed and the platform depends on third-party clouds and models.", "competitors": [ "weights-biases", "mlflow", "neptune" ], "features": [ "Experiment tracking", "Training orchestration (VESSL Run)", "Model serving (VESSL Serve)", "ML pipelines (VESSL Pipelines)", "Deployment and monitoring", "Suite modules: VESSL Serve, VESSL Pipelines, VESSL Run" ], "pricing": { "free": "Free tier", "team": "Usage-based", "enterprise": "Contact sales" }, "link": "https://docs.vessl.ai", "apiLink": "https://docs.vessl.ai", "aiCoreShare": 85, "sourceUrls": [ "https://vessl.ai/en", "https://docs.vessl.ai/en", "https://www.samsungsds.com/ap/case-study/vessl-ai-scp-gpuaas.html", "https://vessl.ai/ko/blog/vessl-ai-series-a-funding" ] }, { "id": "dynatrace-automationengine", "name": "Dynatrace AutomationEngine", "company": "Dynatrace", "region": "global", "tagline": "Causal observability and automation AI", "description": "Dynatrace AutomationEngine is Dynatrace's causal observability and automation AI.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 71, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://dynatrace.com", "apiLink": null, "aiCoreShare": 73, "contentReview": "needs-review" }, { "id": "helicone-experiments", "name": "Helicone Experiments", "company": "Helicone", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 71, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://helicone.ai", "apiLink": null, "aiCoreShare": 84, "contentReview": "needs-review" }, { "id": "monte-carlo-data-reliability-copilot", "name": "Monte Carlo Data Reliability Copilot", "company": "Monte Carlo", "region": "global", "tagline": "Data observability with AI incident workflows", "description": "Monte Carlo Data Reliability Copilot is Monte Carlo's data observability with AI incident workflows.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 71, "competitors": [ "GitHub Copilot", "Cursor", "Windsurf Editor", "Tabnine" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://montecarlodata.com", "apiLink": null, "aiCoreShare": 68, "contentReview": "needs-review" }, { "id": "new-relic-ai", "name": "New Relic AI", "company": "New Relic", "region": "global", "tagline": "Observability assistant and analytics AI", "description": "New Relic AI is New Relic's observability assistant and analytics AI.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 71, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://newrelic.com", "apiLink": null, "aiCoreShare": 73, "contentReview": "needs-review" }, { "id": "snyk-ai-trust-platform", "name": "Snyk", "company": "Snyk", "region": "global", "tagline": "Developer security scanning with AI code analysis", "description": "Snyk scans code, dependencies, containers and IaC for vulnerabilities inside developer workflows, with DeepCode AI performing semantic analysis and AI-generated fixes. Its position is developer adoption ahead of security-team procurement.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 77, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "Developer workflow automation", "Model operations", "Observability", "Secure deployment", "Suite modules: Snyk AppRisk Pro, Snyk DeepCode AI, Snyk Code" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://snyk.io", "apiLink": null, "aiCoreShare": 88, "sourceUrls": [], "versions": [ { "name": "Snyk", "variants": [ "Snyk AppRisk Pro", "Snyk DeepCode AI", "Snyk Code" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "weaviate", "name": "Weaviate", "company": "Weaviate", "region": "global", "tagline": "Weaviate's open-source vector database", "description": "Weaviate is an open-source vector database for semantic search and retrieval-augmented generation, run self-hosted or as Weaviate Cloud, with agent capabilities built on the same engine. Self-hosted versus managed is packaging, not a different product.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 77, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "Developer workflow automation", "Model operations", "Observability", "Secure deployment", "Suite modules: Weaviate Cloud, Weaviate Agents" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://weaviate.io", "apiLink": null, "aiCoreShare": 86, "sourceUrls": [], "versions": [ { "name": "Weaviate", "variants": [ "Weaviate Cloud", "Weaviate Agents" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "abnormal-ai-platform", "name": "Abnormal Security", "company": "Abnormal Security", "region": "global", "tagline": "Behavioural AI against email compromise and account takeover", "description": "Abnormal models normal communication behaviour per organisation to catch business email compromise and account takeover that signature-based filters miss, extending into posture management across email and collaboration tools.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 70, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance", "Suite modules: Abnormal Inbound Email Security, Abnormal Security Posture Management, Abnormal Account Takeover Protection" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://abnormalsecurity.com", "apiLink": null, "aiCoreShare": 80, "sourceUrls": [], "versions": [ { "name": "Abnormal Security", "variants": [ "Abnormal Inbound Email Security", "Abnormal Security Posture Management", "Abnormal Account Takeover Protection" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "monte-carlo-data-observability", "name": "Monte Carlo Data Observability", "company": "Monte Carlo", "region": "global", "tagline": "Data observability with AI incident workflows", "description": "Monte Carlo Data Observability is Monte Carlo's data observability with AI incident workflows.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 70, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://montecarlodata.com", "apiLink": null, "aiCoreShare": 76, "contentReview": "needs-review" }, { "id": "cortex-xsiam", "name": "Cortex XSIAM", "company": "Palo Alto Networks", "region": "global", "tagline": "Security operations and cloud security AI", "description": "Cortex XSIAM is Palo Alto Networks' security operations and cloud security AI.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 70, "competitors": [ "CrowdStrike", "Splunk", "SentinelOne", "Darktrace" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://www.paloaltonetworks.com/cortex/cortex-xsiam", "apiLink": null, "aiCoreShare": 76, "sourceUrls": [ "https://www.paloaltonetworks.com/cortex/cortex-xsiam" ], "contentReview": "needs-review" }, { "id": "insightidr-ai", "name": "InsightIDR AI", "company": "Rapid7", "region": "global", "tagline": "Security analytics and remediation AI", "description": "InsightIDR AI is Rapid7's security analytics and remediation AI.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 70, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://www.rapid7.com/products/siem/", "apiLink": null, "aiCoreShare": 68, "sourceUrls": [ "https://www.rapid7.com/products/siem/" ], "contentReview": "needs-review" }, { "id": "tenable-identity-exposure-ai", "name": "Tenable Identity Exposure AI", "company": "Tenable", "region": "global", "tagline": "Exposure management AI", "description": "Tenable Identity Exposure AI is Tenable's exposure management AI.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 70, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://tenable.com", "apiLink": null, "aiCoreShare": 74, "contentReview": "needs-review" }, { "id": "trulens", "name": "TruLens", "company": "TruEra", "region": "global", "tagline": "LLM observability and evaluation tooling", "description": "TruLens is TruEra's lLM observability and evaluation tooling.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 70, "competitors": [ "LangSmith", "LangFuse", "Arize", "Helicone" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://www.trulens.org/", "apiLink": null, "aiCoreShare": 78, "sourceUrls": [ "https://www.trulens.org/" ], "contentReview": "needs-review" }, { "id": "h2o-ai-cloud", "name": "H2O.ai", "company": "H2O.ai", "region": "global", "tagline": "AutoML, document AI and open LLMs on one platform", "description": "H2O.ai spans automated machine learning, document extraction, computer vision and open-weight LLM serving, delivered as a managed cloud or self-hosted platform for enterprises with data-residency constraints.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 79, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "Developer workflow automation", "Model operations", "Observability", "Secure deployment", "Suite modules: h2oGPT, H2O Driverless AI, H2O Hydrogen Torch, H2O Document AI" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://h2o.ai", "apiLink": null, "aiCoreShare": 90, "sourceUrls": [], "versions": [ { "name": "H2O.ai", "variants": [ "h2oGPT", "H2O Driverless AI", "H2O Hydrogen Torch", "H2O Document AI" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "zilliz-cloud", "name": "Zilliz Cloud", "company": "Zilliz", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 69, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://zilliz.com", "apiLink": null, "aiCoreShare": 81, "contentReview": "needs-review" }, { "id": "cline-rules", "name": "Cline Rules", "company": "Cline", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 68, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://cline.bot", "apiLink": null, "aiCoreShare": 85, "contentReview": "needs-review" }, { "id": "datarobot-ai-platform", "name": "DataRobot", "company": "DataRobot", "region": "global", "tagline": "Enterprise AI platform for AutoML, apps and observability", "description": "DataRobot automates model building, deployment, monitoring and generative application development for enterprises, with governance and observability across the lifecycle. It was an AutoML pioneer now repositioned around generative AI.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "Developer workflow automation", "Model operations", "Observability", "Secure deployment", "Suite modules: DataRobot GenAI Apps, DataRobot AI Observability, DataRobot AutoML" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://datarobot.com", "apiLink": null, "aiCoreShare": 85, "sourceUrls": [], "versions": [ { "name": "DataRobot", "variants": [ "DataRobot GenAI Apps", "DataRobot AI Observability", "DataRobot AutoML" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "siliconflow-serverless-api", "name": "SiliconFlow API Docs", "company": "SiliconFlow", "region": "china", "tagline": "Developer documentation for SiliconFlow model APIs", "description": "SiliconFlow API Docs is the developer entry point for integrating SiliconFlow's model APIs and inference services.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 68, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://api-docs.siliconflow.cn/", "apiLink": "https://api-docs.siliconflow.cn/", "aiCoreShare": 85, "sourceUrls": [ "https://api-docs.siliconflow.cn/" ] }, { "id": "tabnine", "name": "Tabnine", "company": "Tabnine", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 68, "competitors": [ "GitHub Copilot", "Cursor", "Windsurf Editor", "Sourcegraph Cody" ], "features": [ "Suite modules: Tabnine Chat" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://tabnine.com", "apiLink": null, "aiCoreShare": 90, "sourceUrls": [], "versions": [ { "name": "Tabnine", "variants": [ "Tabnine Chat" ] } ], "versionsUpdated": "2026-07-29", "contentReview": "needs-review" }, { "id": "v0", "name": "v0", "company": "Vercel", "region": "global", "tagline": "AI-powered UI generation by Vercel", "description": "v0 is Vercel's AI product for generating and iterating web interfaces from prompts. Duplicate Vercel v0 rows resolve to this canonical slug.", "users": "1M+", "userCount": 1000000, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks", "replit" ], "features": [ "Prompt-to-UI generation", "React and frontend iteration", "Vercel workflow integration", "Design-to-code prototyping", "Suite modules: v0 by Vercel", "Text-to-UI generation", "React/Next.js code", "Tailwind CSS", "shadcn/ui components" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://v0.dev", "apiLink": "https://v0.dev", "aiCoreShare": 85, "sourceUrls": [], "adoptionSignal": "Vercel reports 1M+ users for v0. Distribution runs through Vercel's existing developer base, and paid conversion is not disclosed.", "scoringRationale": "Raised from 73 to 78. The row had 1M disclosed users and no rationale. That is real adoption in a crowded prompt-to-UI category — above Lovable (76) and Bolt (73), both of which disclose nothing — and v0 converts directly into Vercel hosting, which is durable workflow. Held at 78 because 1M is modest against the category leaders and much of the reach is Vercel's existing developer base rather than earned independently." }, { "id": "replicate", "name": "Replicate", "company": "Replicate", "region": "global", "tagline": "Run and fine-tune open models via API", "description": "Replicate hosts open-source models behind a uniform API, with the open Cog packaging format defining how models are containerised, plus fine-tuning and dedicated deployments. Cog's role as the packaging standard is the ecosystem hook.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "Developer workflow automation", "Model operations", "Observability", "Secure deployment", "Suite modules: Cog, Replicate Deployments, Replicate Training" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://replicate.com", "apiLink": null, "aiCoreShare": 86, "sourceUrls": [ "https://github.com/replicate/cog" ], "versions": [ { "name": "Replicate", "variants": [ "Cog", "Replicate Deployments", "Replicate Training" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "builder-design-to-code", "name": "Builder Design-to-Code", "company": "Builder.io", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 66, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://builder.io", "apiLink": null, "aiCoreShare": 79, "contentReview": "needs-review" }, { "id": "charlotte-ai", "name": "Charlotte AI", "company": "CrowdStrike", "region": "global", "tagline": "Cybersecurity AI analyst and platform", "description": "Charlotte AI is CrowdStrike's cybersecurity AI analyst and platform.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 66, "competitors": [ "Splunk", "Palo Alto Networks", "SentinelOne", "Darktrace" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://www.crowdstrike.com/en-us/platform/charlotte-ai/", "apiLink": null, "aiCoreShare": 78, "sourceUrls": [ "https://www.crowdstrike.com/en-us/platform/charlotte-ai/" ], "contentReview": "needs-review" }, { "id": "portkey-guardrails", "name": "Portkey Guardrails", "company": "Portkey", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 66, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://portkey.ai", "apiLink": null, "aiCoreShare": 79, "contentReview": "needs-review" }, { "id": "w-and-b-models", "name": "Weights & Biases", "company": "Weights & Biases", "region": "global", "tagline": "Experiment tracking and LLM observability for ML teams", "description": "Weights & Biases tracks training runs, datasets and model versions, and extends into LLM tracing and evaluation through Weave, job orchestration through Launch, and hosted inference. These are modules of one platform under a single account. W&B was acquired by CoreWeave in 2025.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "Developer workflow automation", "Model operations", "Observability", "Secure deployment", "Suite modules: W&B Weave, W&B Launch, W&B Prompts, W&B Inference" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://wandb.ai", "apiLink": null, "aiCoreShare": 88, "sourceUrls": [], "versions": [ { "name": "Weights & Biases", "variants": [ "W&B Weave", "W&B Launch", "W&B Prompts", "W&B Inference" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "windsurf-editor", "name": "Windsurf Editor", "company": "Windsurf", "region": "global", "tagline": "", "description": "Windsurf Editor is the agentic AI IDE originally built by Codeium (renamed Windsurf in 2025), now owned and developed by Cognition after the July 2025 carve-up in which Google licensed the technology and hired key founders and Cognition acquired the product, brand, and team. Cascade agentic flows, live previews, and autocomplete are modules of the editor. Reported ~$82M ARR at acquisition; current usage under Cognition is not separately disclosed.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 66, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://windsurf.com", "apiLink": null, "aiCoreShare": 78, "adoptionSignal": "Acquired by Cognition Jul 2025 (~$82M ARR reported at the time); Google licensed the tech (~$2.4B) and hired founders", "sourceUrls": [ "https://techcrunch.com/2026/05/27/ai-coding-startup-cognition-raises-1b-at-25b-pre-money-valuation/" ], "contentReview": "needs-review" }, { "id": "milvus", "name": "Milvus", "company": "Zilliz", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 66, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://github.com/milvus-io/milvus", "apiLink": null, "aiCoreShare": 81, "sourceUrls": [ "https://github.com/milvus-io/milvus" ], "contentReview": "needs-review" }, { "id": "evidently", "name": "Evidently", "company": "Evidently AI", "region": "global", "tagline": "Evidently's open-source ML and LLM evaluation and monitoring", "description": "Evidently is an evaluation and monitoring tool for ML and LLM systems, available as the open-source library and as Evidently Cloud. Reports and test suites are features of the library, and cloud versus self-hosted is packaging rather than a different product.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 73, "competitors": [ "LangSmith", "LangFuse", "Arize", "Helicone" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance", "Suite modules: Evidently Cloud, Evidently Reports, Evidently Test Suites" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://evidentlyai.com", "apiLink": null, "aiCoreShare": 90, "sourceUrls": [], "versions": [ { "name": "Evidently", "variants": [ "Evidently Cloud", "Evidently Reports", "Evidently Test Suites" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "langfuse", "name": "Langfuse", "company": "Langfuse", "region": "global", "tagline": "Open-source LLM engineering and observability platform", "description": "Langfuse traces LLM application calls, manages prompt versions and runs evaluations, self-hosted or cloud. Open-source and self-hostable is the wedge against closed observability vendors for teams with data constraints.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "Developer workflow automation", "Model operations", "Observability", "Secure deployment", "Suite modules: Langfuse Prompt Management, Langfuse Evaluations, Langfuse Tracing" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://langfuse.com", "apiLink": null, "aiCoreShare": 82, "sourceUrls": [], "versions": [ { "name": "Langfuse", "variants": [ "Langfuse Prompt Management", "Langfuse Evaluations", "Langfuse Tracing" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "pydantic-validation", "name": "Pydantic Validation", "company": "Pydantic", "region": "global", "tagline": "Typed agent and validation tooling", "description": "Pydantic Validation is Pydantic's typed agent and validation tooling.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 63, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://pydantic.dev", "apiLink": null, "aiCoreShare": 70, "contentReview": "needs-review" }, { "id": "dataiku-dss", "name": "Dataiku", "company": "Dataiku", "region": "global", "tagline": "End-to-end data science and AI platform for enterprises", "description": "Dataiku covers data preparation, AutoML, deployment and governance in one workspace, with an LLM Mesh layer for routing enterprise applications across model providers. It sells to large enterprises standardising analytics and AI on a single platform.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 76, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks", "GPT-5", "claude", "Gemini 2.5", "Grok 4" ], "features": [ "Developer workflow automation", "Model operations", "Observability", "Secure deployment", "Suite modules: Dataiku LLM Mesh, Dataiku AutoML, Dataiku Govern" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://dataiku.com", "apiLink": null, "aiCoreShare": 84, "sourceUrls": [], "versions": [ { "name": "Dataiku", "variants": [ "Dataiku LLM Mesh", "Dataiku AutoML", "Dataiku Govern" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "instructor", "name": "Instructor", "company": "Instructor", "region": "global", "tagline": "Structured outputs and extraction library", "description": "Instructor is Instructor's structured outputs and extraction library.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 62, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://python.useinstructor.com", "apiLink": null, "aiCoreShare": 72, "contentReview": "needs-review" }, { "id": "errors-inbox-ai", "name": "Errors Inbox AI", "company": "New Relic", "region": "global", "tagline": "Observability assistant and analytics AI", "description": "Errors Inbox AI is New Relic's observability assistant and analytics AI.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 62, "competitors": [ "Superhuman", "Shortwave", "Lavender", "Spike" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://newrelic.com/platform/errors-inbox", "apiLink": null, "aiCoreShare": 72, "sourceUrls": [ "https://newrelic.com/platform/errors-inbox" ], "contentReview": "needs-review" }, { "id": "openhands-resolver", "name": "OpenHands Resolver", "company": "OpenHands", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 62, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://github.com/All-Hands-AI/OpenHands", "apiLink": null, "aiCoreShare": 75, "contentReview": "needs-review" }, { "id": "openhands-runtime", "name": "OpenHands Runtime", "company": "OpenHands", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 62, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://github.com/All-Hands-AI/OpenHands", "apiLink": null, "aiCoreShare": 82, "contentReview": "needs-review" }, { "id": "promptfoo", "name": "Promptfoo", "company": "Promptfoo", "region": "global", "tagline": "Open-source prompt testing and LLM red-teaming", "description": "Promptfoo runs declarative test suites over prompts and models, with assertions, side-by-side comparison and automated red-teaming for injection and jailbreak coverage, wired into CI. Config-file simplicity is the adoption path.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 68, "competitors": [ "LangSmith", "LangFuse", "Arize", "Helicone" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance", "Suite modules: Promptfoo Red Teaming, Promptfoo Assertions, Promptfoo Evaluations" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://promptfoo.dev", "apiLink": null, "aiCoreShare": 86, "sourceUrls": [], "versions": [ { "name": "Promptfoo", "variants": [ "Promptfoo Red Teaming", "Promptfoo Assertions", "Promptfoo Evaluations" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "truera-observability", "name": "TruEra Observability", "company": "TruEra", "region": "global", "tagline": "LLM observability and evaluation tooling", "description": "TruEra Observability is TruEra's lLM observability and evaluation tooling.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 62, "competitors": [ "LangSmith", "LangFuse", "Arize", "Helicone" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://truera.com", "apiLink": null, "aiCoreShare": 84, "contentReview": "needs-review" }, { "id": "pydantic-settings", "name": "Pydantic Settings", "company": "Pydantic", "region": "global", "tagline": "Typed agent and validation tooling", "description": "Pydantic Settings is Pydantic's typed agent and validation tooling.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 61, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://pydantic.dev", "apiLink": null, "aiCoreShare": 75, "contentReview": "needs-review" }, { "id": "zilliz-cloud-pipelines", "name": "Zilliz Cloud Pipelines", "company": "Zilliz", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 61, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://zilliz.com", "apiLink": null, "aiCoreShare": 77, "contentReview": "needs-review" }, { "id": "splunk-itsi-predictive-analytics", "name": "Splunk AI", "company": "Splunk", "region": "global", "tagline": "Predictive analytics and anomaly detection over machine data", "description": "Splunk's AI applies anomaly detection, incident prediction and natural-language querying to the log and telemetry data enterprises already store in Splunk. Cisco acquired Splunk in 2024. The value is the existing data estate, not the models.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 60, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance", "Suite modules: Splunk AI Assistant, Splunk App for Anomaly Detection, Splunk Observability AI" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://splunk.com", "apiLink": null, "aiCoreShare": 76, "sourceUrls": [], "versions": [ { "name": "Splunk AI", "variants": [ "Splunk AI Assistant", "Splunk App for Anomaly Detection", "Splunk Observability AI" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "trulens-eval", "name": "TruLens Eval", "company": "TruEra", "region": "global", "tagline": "LLM observability and evaluation tooling", "description": "TruLens Eval is TruEra's lLM observability and evaluation tooling.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 60, "competitors": [ "LangSmith", "LangFuse", "Arize", "Helicone" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://www.trulens.org/", "apiLink": null, "aiCoreShare": 86, "sourceUrls": [ "https://www.trulens.org/" ], "contentReview": "needs-review" }, { "id": "dspy", "name": "DSPy", "company": "DSPy", "region": "global", "tagline": "Framework for programming and optimising LLM pipelines", "description": "DSPy treats prompting as a compilation problem: you declare the pipeline structure and DSPy optimises the prompts and few-shot examples against a metric, instead of hand-tuning strings. It is a research-originated framework with strong academic uptake.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 79, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance", "Suite modules: DSPy Assertions, DSPy Optimizers, DSPy Retrieval Pipelines" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://dspy.ai", "apiLink": null, "aiCoreShare": 77, "sourceUrls": [], "versions": [ { "name": "DSPy", "variants": [ "DSPy Assertions", "DSPy Optimizers", "DSPy Retrieval Pipelines" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "instructor-validation", "name": "Instructor Validation", "company": "Instructor", "region": "global", "tagline": "Structured outputs and extraction library", "description": "Instructor Validation is Instructor's structured outputs and extraction library.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 59, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://python.useinstructor.com", "apiLink": null, "aiCoreShare": 75, "contentReview": "needs-review" }, { "id": "bolt-for-teams", "name": "Bolt for Teams", "company": "StackBlitz", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 59, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://bolt.new", "apiLink": null, "aiCoreShare": 87, "contentReview": "needs-review" }, { "id": "builder-generate", "name": "Builder Generate", "company": "Builder.io", "region": "global", "tagline": "", "description": "", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 58, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://builder.io", "apiLink": null, "aiCoreShare": 84, "contentReview": "needs-review" }, { "id": "monte-carlo-incident-iq", "name": "Monte Carlo Incident IQ", "company": "Monte Carlo", "region": "global", "tagline": "Data observability with AI incident workflows", "description": "Monte Carlo Incident IQ is Monte Carlo's data observability with AI incident workflows.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 58, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://montecarlodata.com", "apiLink": null, "aiCoreShare": 69, "contentReview": "needs-review" }, { "id": "vectra-ai-platform", "name": "Vectra AI Platform", "company": "Vectra AI", "region": "global", "tagline": "AI-driven network detection and response", "description": "Vectra AI Platform is Vectra AI's AI-driven network detection and response.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 58, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://vectra.ai", "apiLink": null, "aiCoreShare": 75, "contentReview": "needs-review" }, { "id": "new-relic-vulnerability-management-ai", "name": "New Relic Vulnerability Management AI", "company": "New Relic", "region": "global", "tagline": "Observability assistant and analytics AI", "description": "New Relic Vulnerability Management AI is New Relic's observability assistant and analytics AI.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 57, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://newrelic.com", "apiLink": null, "aiCoreShare": 68, "contentReview": "needs-review" }, { "id": "guardrails", "name": "Guardrails", "company": "Guardrails AI", "region": "global", "tagline": "Guardrails AI's validation framework for LLM output", "description": "Guardrails is an open-source framework for validating and correcting LLM output. The server deployment, the validator hub and individual validators are parts of one framework rather than separate products.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 72, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance", "Suite modules: Guardrails Server, Guardrails Hub, Guardrails Validators" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://guardrailsai.com", "apiLink": null, "aiCoreShare": 80, "sourceUrls": [], "versions": [ { "name": "Guardrails", "variants": [ "Guardrails Server", "Guardrails Hub", "Guardrails Validators" ] } ], "versionsUpdated": "2026-07-29" }, { "id": "instructor-structured-extraction", "name": "Instructor Structured Extraction", "company": "Instructor", "region": "global", "tagline": "Structured outputs and extraction library", "description": "Instructor Structured Extraction is Instructor's structured outputs and extraction library.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 56, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://python.useinstructor.com", "apiLink": null, "aiCoreShare": 73, "contentReview": "needs-review" }, { "id": "pydantic-logfire", "name": "Pydantic Logfire", "company": "Pydantic", "region": "global", "tagline": "Typed agent and validation tooling", "description": "Pydantic Logfire is Pydantic's typed agent and validation tooling.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 56, "competitors": [ "AWS Bedrock", "gemini-enterprise-agent-platform", "Azure AI Foundry", "Databricks" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://pydantic.dev", "apiLink": null, "aiCoreShare": 77, "contentReview": "needs-review" }, { "id": "trulens-feedback-functions", "name": "TruLens Feedback Functions", "company": "TruEra", "region": "global", "tagline": "LLM observability and evaluation tooling", "description": "TruLens Feedback Functions is TruEra's lLM observability and evaluation tooling.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 55, "competitors": [ "LangSmith", "LangFuse", "Arize", "Helicone" ], "features": [ "AI engineering workflow", "Observability", "Evaluation", "Security and governance" ], "pricing": { "api": "Usage-based or enterprise pricing", "enterprise": "Contact sales" }, "link": "https://www.trulens.org/", "apiLink": null, "aiCoreShare": 80, "sourceUrls": [ "https://www.trulens.org/" ], "contentReview": "needs-review" }, { "id": "panmnesia-cxl-switch", "name": "Panmnesia CXL Switch", "company": "Panmnesia", "region": "korea", "tagline": "PCIe 6.4 / CXL 3.2 fabric switch for AI servers.", "description": "Panmnesia's CXL switch enables disaggregated memory pools and high-bandwidth AI server fabrics.", "users": "Not disclosed", "userCount": 0, "adoptionSignal": "Early hyperscaler engagements", "technicalImpact": 76, "competitors": [ "Astera Labs", "Marvell", "Microchip", "Synopsys" ], "features": [ "PCIe 6.4 / CXL 3.2", "Disaggregated memory", "AI server fabric" ], "pricing": { "free": "N/A", "pro": "Per unit", "enterprise": "Contact sales" }, "link": "https://panmnesia.com", "apiLink": null, "aiCoreShare": 70 }, { "id": "panmnesia-cxl-ip", "name": "Panmnesia CXL IP", "company": "Panmnesia", "region": "korea", "tagline": "CXL controller and switch IP for AI infrastructure.", "description": "Panmnesia licenses its CXL controller and switch IP to silicon vendors.", "users": "Not disclosed", "userCount": 0, "adoptionSignal": "Licensed by silicon partners", "technicalImpact": 72, "competitors": [ "Synopsys", "Cadence", "Rambus", "Astera Labs" ], "features": [ "CXL controller IP", "Switch IP", "Licensable" ], "pricing": { "free": "N/A", "pro": "License", "enterprise": "Contact sales" }, "link": "https://panmnesia.com", "apiLink": null, "aiCoreShare": 64 }, { "id": "dnotitia-mnemos", "name": "Mnemos", "company": "Dnotitia", "region": "korea", "tagline": "On-device personal AI with vector long-term memory.", "description": "Mnemos is Dnotitia's on-device personal AI pairing small language models with vector databases.", "users": "Not disclosed", "userCount": 0, "adoptionSignal": "Early consumer-AI partnerships", "technicalImpact": 65, "competitors": [ "Apple Intelligence", "Microsoft Copilot+ PC", "Mem", "Reflect" ], "features": [ "On-device LLM", "Vector DB", "Personal memory", "Privacy-first" ], "pricing": { "free": "Limited tier", "pro": "Subscription", "enterprise": "Contact sales" }, "link": "https://dnotitia.com", "apiLink": null, "aiCoreShare": 88 }, { "id": "enerzai-optimizer", "name": "ENERZAi Optimizer", "company": "ENERZAi", "region": "korea", "tagline": "1.58-bit quantization for AI models on Arm SoCs.", "description": "ENERZAi's optimizer compresses and quantizes deep-learning models for Arm-based mobile and edge SoCs.", "users": "Not disclosed", "userCount": 0, "adoptionSignal": "Korean mobile / device manufacturer pilots", "technicalImpact": 70, "competitors": [ "Nota AI", "Qualcomm AI Engine", "Deci AI", "Neural Magic" ], "features": [ "1.58-bit quantization", "Arm SoC runtime", "LLM compression" ], "pricing": { "free": "Free trial", "pro": "Subscription", "enterprise": "Contact sales" }, "link": "https://enerzai.com", "apiLink": null, "aiCoreShare": 82 }, { "id": "squeezebits-yeter", "name": "Yeter Inference Engine", "company": "SqueezeBits", "region": "korea", "tagline": "High-throughput LLM inference engine.", "description": "Yeter is SqueezeBits' LLM inference engine integrated with Modular MAX and vLLM.", "users": "Not disclosed", "userCount": 0, "adoptionSignal": "Demonstrated at NVIDIA GTC 2026; Modular and Rebellions partnerships", "technicalImpact": 72, "competitors": [ "vLLM", "TensorRT-LLM", "Together AI", "Modal" ], "features": [ "LLM serving", "Diffusion pipelines", "Modular MAX integration" ], "pricing": { "free": "Open source path", "pro": "Subscription", "enterprise": "Contact sales" }, "link": "https://www.squeezebits.com", "apiLink": null, "aiCoreShare": 84 }, { "id": "flitto-datalab", "name": "Flitto DataLab", "company": "Flitto", "region": "korea", "tagline": "AI training-data construction service for LLMs.", "description": "Flitto DataLab builds Korean and multilingual instruction, evaluation, and pretraining datasets.", "users": "Not disclosed", "userCount": 0, "adoptionSignal": "Selected by Upstage's sovereign-AI consortium", "technicalImpact": 74, "competitors": [ "Scale AI", "Surge AI", "Labelbox", "SuperAnnotate" ], "features": [ "LLM training datasets", "Korean / multilingual", "Instruction & eval data" ], "pricing": { "free": "N/A", "pro": "Project-based", "enterprise": "Contact sales" }, "link": "https://datalab.flitto.com", "apiLink": null, "aiCoreShare": 70 }, { "id": "energy-x-platform", "name": "Energy X Platform", "company": "Energy X", "region": "korea", "tagline": "AI for renewable-energy site selection and PPA structuring.", "description": "Energy X combines geospatial data, financial modeling, and AI-driven site analysis for solar and renewable-energy asset development.", "users": "Not disclosed", "userCount": 0, "adoptionSignal": "Korean renewables developer customers", "technicalImpact": 66, "competitors": [ "Aurora Solar", "Watershed", "Pachama", "Persefoni" ], "features": [ "AI site selection", "Renewable asset analytics", "PPA structuring" ], "pricing": { "free": "N/A", "pro": "Subscription", "enterprise": "Contact sales" }, "link": "https://energyx.ai", "apiLink": null, "aiCoreShare": 70 }, { "id": "tunibridge", "name": "TUNiBridge", "company": "TUNiB", "region": "korea", "tagline": "LLM safety and de-identification API.", "description": "TUNiBridge gives developers Korean-language LLM safety, persona, and de-identification APIs.", "users": "Not disclosed", "userCount": 0, "adoptionSignal": "Used by Korean LLM-app developers", "technicalImpact": 66, "competitors": [ "Lakera", "Guardrails AI", "Protect AI", "WhyLabs" ], "features": [ "LLM safety", "De-identification", "Persona chatbots" ], "pricing": { "free": "Free tier", "pro": "Subscription", "enterprise": "Contact sales" }, "link": "https://tunib.ai", "apiLink": null, "aiCoreShare": 88 }, { "id": "ai-network-gpu", "name": "AI Network GPU Cloud", "company": "Common Computer", "region": "korea", "tagline": "Decentralized GPU cloud for AI training and inference.", "description": "Common Computer's decentralized GPU compute network with A100/H100 access from ~$0.99/hr.", "users": "Not disclosed", "userCount": 0, "adoptionSignal": "Korean AI dev community deployments", "technicalImpact": 64, "competitors": [ "RunPod", "Vast.ai", "CoreWeave", "Lambda Labs" ], "features": [ "Decentralized GPU", "A100 / H100", "Blockchain settlement" ], "pricing": { "free": "Free credits", "pro": "Pay-per-hour", "enterprise": "Contact sales" }, "link": "https://comcom.ai", "apiLink": null, "aiCoreShare": 70 }, { "id": "alchera-firescout", "name": "FireScout", "company": "Alchera", "region": "korea", "tagline": "AI wildfire detection on utility camera networks", "description": "FireScout detects wildfire smoke and flame on existing camera networks (including ALERTWildfire and utility feeds), alerting operators in as little as 60 seconds — often before 911 reports. Operated by US subsidiary Alchera X, it monitors 300+ cameras across the US and Australia (company-claimed) with named deployments at PG&E, Sonoma County Department of Emergency Management, NV Energy, and San Diego Gas & Electric. Contract values are not disclosed.", "users": "Not disclosed", "userCount": 0, "adoptionSignal": "Named US deployments: PG&E, Sonoma County DEM, NV Energy, SDG&E; 300+ cameras monitored (company-claimed); ABC7-covered pilots", "technicalImpact": 76, "aiCoreShare": 92, "competitors": [ "Pano AI", "AlertCalifornia AI", "Dryad Networks", "OroraTech" ], "features": [ "Real-time smoke/flame detection on existing cameras", "~60-second ignition alerts", "300+ cameras across the US and Australia", "Runs on existing CCTV/ALERTWildfire infrastructure" ], "pricing": { "free": "N/A", "pro": "Per deployment", "enterprise": "Contact sales" }, "link": "https://firescout.ai/", "apiLink": null, "sourceUrls": [ "https://firescout.ai/", "https://abc7news.com/pge-fire-camera-wildfire-al-cameras-alchera/11240447/", "https://www.globenewswire.com/news-release/2025/01/16/3011061/0/en/California-Wildfire-Crisis-Alchera-X-s-FireScout-AI-Technology-Could-Have-Mitigated-and-Prevented-Devastation.html" ], "scoringRationale": "76 reflects rare verified overseas deployment for Korean vision AI — named US utility and county customers in a safety-critical monitoring workflow; held below 80 because camera counts are company-claimed and contract scale is undisclosed." }, { "id": "alchera-face-trust", "name": "Face Trust", "company": "Alchera", "region": "korea", "tagline": "Enterprise face-verification engine", "description": "Face Trust is Alchera's facial-recognition engine — a top Korean performer in NIST FRVT categories (company claim) and the first Korean RGB-camera solution to pass the iBeta anti-spoofing test — powering Incheon International Airport's Smart Pass since July 2023 and formerly Shinhan Card's Face Pay (being wound down in 2026). It ships as a cross-platform SDK with distributed biometric storage. Enrollment volumes are not disclosed.", "users": "Not disclosed", "userCount": 0, "adoptionSignal": "Incheon Airport Smart Pass deployment (2023-); iBeta PAD anti-spoofing certification; Shinhan Face Pay historical deployment", "technicalImpact": 72, "aiCoreShare": 92, "competitors": [ "Suprema", "CUBOX", "NEC NeoFace", "Idemia" ], "features": [ "High-accuracy face matching robust to aging and masks", "iBeta PAD anti-spoofing on RGB cameras", "Distributed biometric data management", "Airport smart-pass and access-control deployments" ], "pricing": { "free": "N/A", "pro": "Per deployment", "enterprise": "Contact sales" }, "link": "https://alchera.ai/solutions/face-verification", "apiLink": null, "sourceUrls": [ "https://alchera.ai/solutions/face-verification", "https://www.aitimes.kr/news/articleView.html?idxno=28492" ], "scoringRationale": "72 reflects a nationally visible airport deployment and certified anti-spoofing; kept modest because the Shinhan Face Pay reference is ending and enrollment/usage volumes are undisclosed." }, { "id": "gauss-panoptes-vm", "name": "Panoptes VM", "company": "Gauss Labs", "region": "korea", "tagline": "AI virtual metrology for semiconductor fabs", "description": "Panoptes VM predicts process outcomes for every wafer in real time from equipment sensor data, replacing sampled physical measurement, and plugs into advanced process control to cut variability and improve yield. It has run in SK hynix mass production since December 2022 (thin-film deposition) across five sites in Korea and China, with SK hynix reporting a 21.5% average reduction in process variability. VM 2.0 shipped in 2024; customers beyond SK hynix are not disclosed.", "users": "Not disclosed", "userCount": 0, "adoptionSignal": "SK hynix mass-production deployment since Dec 2022 across 5 sites in Korea/China; SK hynix-reported 21.5% process-variability reduction", "technicalImpact": 78, "aiCoreShare": 90, "competitors": [ "PDF Solutions Exensio", "Applied Materials SmartFactory", "Onto Innovation", "Tignis" ], "features": [ "Real-time process-outcome prediction for 100% of wafers", "Automated management of hundreds of thousands of ML models", "APC integration for variability reduction", "Sampling optimization and maintenance prediction" ], "pricing": { "free": "N/A", "pro": "Per deployment", "enterprise": "Contact sales" }, "link": "https://www.gausslabs.ai/vm", "apiLink": null, "sourceUrls": [ "https://www.gausslabs.ai/vm", "https://news.skhynix.com/gauss-labss-ai-based-virtual-metrology-solution/", "https://www.prnewswire.com/news-releases/gauss-labs-releases-ai-based-virtual-metrology-solution-panoptes-vm-2-0--302219771.html" ], "scoringRationale": "2026-07-29 stress-test follow-up: reduced 80 -> 78. The 5-fab production deployment with a quantified outcome is real and sourced — but the named customer is SK hynix, Gauss Labs' founding investor. A captive-parent deployment is weaker evidence than an arms-length customer (compare ATOM-Max, 82, whose KT Cloud deployment is a CSAP-certified public commercial service). Returns to 80 with a named customer beyond the SK family." }, { "id": "rtzr-stt-api", "name": "RTZR STT API", "company": "Return Zero", "region": "korea", "tagline": "Korean/Japanese speech-to-text API", "description": "RTZR STT is Return Zero's enterprise speech-to-text API for Korean and Japanese, trained on 15M+ hours of Korean audio, with claims of 35% lower error rates and 2.5x faster conversion versus rivals. Referenced deployments concentrate in finance, telecom, public-sector, and fire-department call handling, and it underpins the company's Proactive Voice Agent stack now deployed inside parent Ubase's contact-center operations. Product-specific customer counts are not disclosed.", "users": "Not disclosed", "userCount": 0, "adoptionSignal": "Finance/telecom/public-sector references in funding and acquisition coverage; deployed in Ubase contact-center operations post-acquisition", "technicalImpact": 72, "aiCoreShare": 92, "competitors": [ "Google Cloud Speech-to-Text", "selvy-stt", "Naver Clova Speech", "Deepgram" ], "features": [ "Korean/Japanese STT with speaker diarization", "Real-time streaming at large concurrency", "Developer portal with API keys and docs", "Powers Proactive Voice Agent stack" ], "pricing": { "free": "Free trial credits", "pro": "Usage-based", "enterprise": "Contact sales" }, "link": "https://www.rtzr.ai/stt", "apiLink": "https://developers.rtzr.ai", "sourceUrls": [ "https://platum.kr/archives/274227", "https://www.aitimes.kr/news/articleView.html?idxno=40213" ], "scoringRationale": "72 reflects a differentiated Korean-language STT API with a proprietary 15M-hour corpus and sector references; held below 80 because customer counts are undisclosed and hyperscaler STT is an ever-present substitute." }, { "id": "nvidia-nim", "name": "NVIDIA NIM", "company": "NVIDIA", "region": "global", "tagline": "Containerized inference microservices for deploying models on NVIDIA GPUs", "description": "NIM packages optimized model inference as prebuilt containers with standard APIs, so a team can deploy a frontier or open-weight model on NVIDIA GPUs in their own cloud or datacenter without building a serving stack. It is distributed through NVIDIA AI Enterprise and the NVIDIA API catalog, and is built on TensorRT-LLM and Triton underneath. NVIDIA publishes no customer count for it.", "users": "Enterprise AI deployment infrastructure; customer counts not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AMD Instinct", "Cloud TPU v5", "AWS Trainium 2" ], "features": [ "Prebuilt inference containers with OpenAI-compatible APIs", "Self-hosted deployment in customer cloud or datacenter", "Built on TensorRT-LLM and Triton", "Distributed through NVIDIA AI Enterprise and the NVIDIA API catalog", "Frontier and open-weight model coverage" ], "pricing": { "free": "Free for development via the NVIDIA API catalog", "enterprise": "Licensed through NVIDIA AI Enterprise (per-GPU)" }, "link": "https://www.nvidia.com/en-us/ai-data-science/products/nim-microservices/", "apiLink": null, "aiCoreShare": 80, "adoptionSignal": "Distributed through NVIDIA AI Enterprise and the public NVIDIA API catalog. NVIDIA discloses no customer count, seat count or named production deployment for NIM specifically.", "sourceUrls": [ "https://www.nvidia.com/en-us/ai-data-science/products/nim-microservices/" ], "scoringRationale": "Reduced from 85 to 78. The previous score argued deployment-cost reduction and ecosystem standardization without any adoption figure, named customer or revenue attribution. 78 is the ceiling for a real product with no recorded adoption evidence. Clears upward when NVIDIA discloses NIM customers, licensed-GPU volume, or named production deployments." }, { "id": "nvidia-triton-inference-server", "name": "NVIDIA Triton Inference Server", "company": "NVIDIA", "region": "global", "tagline": "Open-source model serving for multi-framework inference", "description": "Triton Inference Server is NVIDIA's open-source serving layer for deploying models from any framework — TensorRT, PyTorch, TensorFlow, ONNX, Python — across GPU and CPU targets, with dynamic batching, concurrent model execution and model-repository management. It is production serving infrastructure rather than a hardware product. NVIDIA publishes no adoption figures for it.", "users": "Open inference serving infrastructure; user counts not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "vLLM", "TensorRT-LLM", "Together AI", "Anyscale" ], "features": [ "Multi-framework model serving (TensorRT, PyTorch, TensorFlow, ONNX)", "Dynamic batching and concurrent model execution", "GPU and CPU deployment targets", "Model repository and version management", "Apache 2.0 open source" ], "pricing": { "free": "Open source (Apache 2.0)", "enterprise": "Included with NVIDIA AI Enterprise" }, "link": "https://developer.nvidia.com/triton-inference-server", "apiLink": null, "aiCoreShare": 82, "adoptionSignal": "Open-source serving project published on GitHub and packaged into NVIDIA AI Enterprise. NVIDIA discloses no user, download or named-deployment count.", "sourceUrls": [ "https://developer.nvidia.com/triton-inference-server" ], "scoringRationale": "Reduced from 86 to 78 alongside the rest of NVIDIA's software rows. The score rested on ecosystem argument rather than evidence: no adoption figure, no named deployment, no revenue attribution. 78 is the ceiling for a real product with no recorded adoption evidence. Clears upward with quantified adoption or named production deployments." }, { "id": "nvidia-dgx-cloud", "name": "NVIDIA DGX Cloud", "company": "NVIDIA", "region": "global", "tagline": "Managed AI training capacity on NVIDIA GPU clusters, rented through major clouds", "description": "DGX Cloud is NVIDIA's managed AI supercomputing service: dedicated NVIDIA GPU cluster capacity with the NVIDIA software stack, sold through AWS, Azure, Google Cloud and Oracle Cloud rather than as hardware a buyer racks themselves. It competes with the clouds' own GPU instances and with neoclouds such as CoreWeave and Lambda. NVIDIA does not disclose DGX Cloud customer counts or revenue separately.", "users": "Cloud AI supercomputing service; customer counts not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AMD Instinct", "Cloud TPU v5", "AWS Trainium 2" ], "features": [ "Dedicated managed NVIDIA GPU cluster capacity", "Available through AWS, Azure, Google Cloud and Oracle Cloud", "Bundled NVIDIA AI Enterprise software stack", "Large-model training and fine-tuning workloads", "Monthly-commitment rental rather than hardware purchase" ], "pricing": { "enterprise": "Monthly per-instance commitment; contact NVIDIA or the host cloud" }, "link": "https://www.nvidia.com/en-us/data-center/dgx-cloud/", "apiLink": null, "aiCoreShare": 83, "adoptionSignal": "Sold through AWS, Azure, Google Cloud and Oracle Cloud marketplaces. NVIDIA does not break out DGX Cloud customer counts, committed capacity or revenue, and the catalog records no named customer.", "sourceUrls": [ "https://www.nvidia.com/en-us/data-center/dgx-cloud/" ], "scoringRationale": "Reduced from 87 to 78. Availability across four hyperscaler marketplaces is a distribution fact, not adoption evidence — the catalog rule is that platform distribution cannot carry a score. No customer count, committed capacity, named customer or revenue figure is disclosed. Clears upward when NVIDIA discloses DGX Cloud customers, capacity or revenue." }, { "id": "nvidia-tensorrt-llm", "name": "NVIDIA TensorRT-LLM", "company": "NVIDIA", "region": "global", "tagline": "Open-source LLM inference optimization library for NVIDIA GPUs", "description": "TensorRT-LLM is NVIDIA's open-source library for compiling and optimizing large language models for inference on NVIDIA GPUs, covering kernel fusion, quantization, in-flight batching and multi-GPU/multi-node serving. It is the layer between a trained model and production throughput on NVIDIA hardware, and it is what NIM microservices and much of the NVIDIA serving stack are built on. It competes directly with vLLM, SGLang and the serving layers of Together, Anyscale and Modal. NVIDIA publishes no adoption figures for it.", "users": "Developer inference optimization stack; user counts not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "vLLM", "Together AI", "Anyscale", "Modal" ], "features": [ "LLM inference compilation and kernel optimization", "Quantization (FP8, INT4/INT8) and in-flight batching", "Multi-GPU and multi-node serving", "Apache 2.0 open source on GitHub", "Backend for Triton Inference Server and NIM" ], "pricing": { "free": "Open source (Apache 2.0)" }, "link": "https://github.com/NVIDIA/TensorRT-LLM", "apiLink": null, "aiCoreShare": 86, "adoptionSignal": "Apache-2.0 library published on GitHub and used as the inference backend for NVIDIA's own serving stack. NVIDIA discloses no user, download or deployment count, and the catalog records no named production deployment.", "sourceUrls": [ "https://github.com/NVIDIA/TensorRT-LLM" ], "scoringRationale": "Reduced from 87 to 78. The previous score was an assertion — 'a key developer/inference optimization layer' — with no adoption figure, no named deployment, no revenue and no source beyond the project's own GitHub repository. It also escaped the AI-hardware evidence cap only because its features were generic templated tags rather than chip terms. 78 is the catalog's ceiling for a real product with no recorded adoption evidence; the row's own listed competitor Modal scores 76. Clears upward with quantified adoption (downloads, GitHub scale, named production deployments) or disclosed revenue attribution." }, { "id": "nvidia-nemo", "name": "NVIDIA NeMo", "company": "NVIDIA", "region": "global", "tagline": "Framework for customizing and evaluating models on NVIDIA infrastructure", "description": "NeMo is NVIDIA's framework for training, fine-tuning, aligning and evaluating language, speech and multimodal models on NVIDIA infrastructure, with guardrails and retrieval components for enterprise deployment. It is a model-customization toolchain rather than a hardware or managed-service product. NVIDIA publishes no adoption figures for it.", "users": "Enterprise model customization framework; user counts not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 78, "competitors": [ "AMD Instinct", "Cloud TPU v5", "AWS Trainium 2" ], "features": [ "Model training, fine-tuning and alignment", "Speech, language and multimodal model support", "NeMo Guardrails for output control", "Evaluation and data-curation tooling", "Open source, packaged into NVIDIA AI Enterprise" ], "pricing": { "free": "Open source", "enterprise": "Licensed through NVIDIA AI Enterprise" }, "link": "https://www.nvidia.com/en-us/ai-data-science/products/nemo/", "apiLink": null, "aiCoreShare": 88, "adoptionSignal": "Open-source framework packaged into NVIDIA AI Enterprise. NVIDIA discloses no user count, named customer or production-deployment figure for NeMo.", "sourceUrls": [ "https://www.nvidia.com/en-us/ai-data-science/products/nemo/" ], "scoringRationale": "Reduced from 84 to 78. The row already conceded that public adoption evidence was absent; 78 is the catalog's ceiling for that position rather than 84. Clears upward with quantified adoption or named enterprise deployments." }, { "id": "paddlepaddle", "name": "PaddlePaddle", "nativeName": "飞桨", "company": "Baidu", "region": "china", "tagline": "China's largest open-source deep-learning framework and model platform", "description": "PaddlePaddle is Baidu's open-source deep-learning framework, China's first independently developed one, spanning the core training framework, pre-trained model libraries, deployment toolkits and industrial toolchains. Baidu reports 21.85 million developers and 670,000 enterprises building on it — the largest disclosed developer base of any AI platform in the catalogue — with heavy use in Chinese industrial, manufacturing and public-sector deployments where foreign frameworks face procurement constraints.", "features": [ "Open-source deep-learning training framework", "Pre-trained model libraries and industrial toolkits", "Inference deployment across server, mobile and edge", "Domestic hardware support including Kunlun and Ascend", "Industrial and public-sector toolchains" ], "pricing": { "free": "Open source (Apache 2.0)", "enterprise": "Baidu AI Cloud services" }, "link": "https://www.paddlepaddle.org.cn/en", "apiLink": "https://www.paddlepaddle.org.cn/documentation/docs/en/guides/index_en.html", "users": "21.85M+ developers and 670,000+ enterprises (Baidu-reported)", "userCount": 21850000, "releaseDate": "2016", "technicalImpact": 86, "aiCoreShare": 82, "competitors": [ "huggingface", "pytorch", "tensorflow", "mindspore", "modelscope" ], "adoptionSignal": "Baidu reports 21.85M developers and 670,000 enterprises on PaddlePaddle. The figure is company-reported and counts registered developers rather than active ones, but it is an order of magnitude above any other disclosed Chinese developer platform and is corroborated by PaddlePaddle's role as the default framework in Chinese industrial and public-sector AI procurement.", "scoringRationale": "86 reflects the largest disclosed developer base in the catalogue (21.85M) plus genuine framework-level lock-in: models, toolchains and domestic-hardware support are built against PaddlePaddle in Chinese industrial deployments, which is durable switching cost rather than distribution. Below Hugging Face (90) because Hugging Face's ecosystem role is global — model IDs are hardcoded in production code worldwide — while PaddlePaddle's is concentrated in China, and its figure is registered rather than active developers.", "sourceUrls": [ "https://www.paddlepaddle.org.cn/en", "https://www.yicaiglobal.com/news/baidu-upgrades-china-biggest-ai-platform-paddlepaddle-for-nearly-41-million-developers" ] }, { "id": "modelscope", "name": "ModelScope", "nativeName": "魔搭", "company": "Alibaba", "region": "china", "tagline": "Alibaba's open model hub and Model-as-a-Service community", "description": "ModelScope is Alibaba's open model hub, hosting downloadable model weights, datasets and demo spaces with a unified Python SDK for loading and fine-tuning — the primary distribution channel for Qwen and much of China's open-weight ecosystem, and the domestic counterpart to Hugging Face. It matters to a Korea/China comparison because it is where Chinese open-weight models are actually consumed.", "features": [ "Open model hub with downloadable weights", "Unified SDK for loading and fine-tuning", "Datasets and demo spaces", "Primary distribution channel for Qwen", "Alibaba Cloud deployment integration" ], "pricing": { "free": "Free model and dataset hosting", "api": "Alibaba Cloud inference pricing" }, "link": "https://www.modelscope.cn/models", "apiLink": "https://www.modelscope.cn/docs", "users": "Not disclosed", "userCount": 0, "releaseDate": "2022-11", "technicalImpact": 80, "aiCoreShare": 78, "competitors": [ "huggingface", "paddlepaddle", "qwen", "siliconflow-inference" ], "adoptionSignal": "ModelScope is the primary distribution channel for Qwen and a large share of Chinese open-weight releases, with a public model catalogue and SDK. Alibaba discloses no registered-developer, download or active-user figure for the hub itself.", "scoringRationale": "80 reflects a real ecosystem position — the default hub for Chinese open-weight model distribution, including Qwen — without any disclosed adoption figure of its own. It sits well below Hugging Face (90, 5M disclosed users) and PaddlePaddle (86, 21.85M disclosed developers) precisely because those rows disclose numbers and this one does not. Clears upward when Alibaba publishes developer or download figures.", "sourceUrls": [ "https://www.modelscope.cn/models" ] }, { "id": "audiocraft", "name": "AudioCraft", "company": "Meta", "region": "global", "tagline": "Research toolkit for audio generation and compression", "description": "AudioCraft is Meta's downloadable audio research library, including MusicGen, AudioGen and EnCodec, not a hosted commercial assistant or enterprise API. Code and model weights have different licences, including non-commercial weight terms. Current production adoption and maintenance commitments are not established.", "users": "Not disclosed", "userCount": 0, "releaseDate": "", "technicalImpact": 68, "competitors": [ "GPT-5", "claude", "Gemini 2.5", "Grok 4" ], "features": [ "MusicGen and AudioGen research models", "EnCodec audio compression", "Research code and pretrained weights" ], "pricing": { "research": "Code and weights have separate licences; review terms before use" }, "link": "https://github.com/facebookresearch/audiocraft", "apiLink": null, "aiCoreShare": 100, "sourceUrls": [ "https://github.com/facebookresearch/audiocraft" ], "scoringRationale": "Reviewed 2026-09-29: 76 → 68. Available research toolkit, not a commercial foundation-model service; no current production-scale adoption evidence." } ]