Foundation Model Labs

Anthropic

San Francisco, United States

Anthropic is one of the worlds leading AI safety and research companies, founded by former OpenAI researchers Dario and Daniela Amodei and headquartered in San Francisco, California. The company develops the Claude family of large language models including Claude Opus, Claude Sonnet, and Claude Haiku, focused on building reliable, interpretable, and steerable AI systems. Anthropic provides Claude through claude.ai consumer chat interface, the Claude API for developers, Claude Code for terminal-based agentic coding, and enterprise integrations. The company is reportedly preparing for an IPO with major investments from Google and Amazon.

Cohere

Toronto, Canada

Cohere is a leading enterprise-focused foundation model company, founded and headquartered in Toronto, Ontario, Canada with offices in San Francisco and London. Founded by former Google Brain researchers Aidan Gomez (co-author of the original Transformer paper), Ivan Zhang, and Nick Frosst, Cohere develops the Command family of large language models, the Embed family of embedding models, and the Rerank model for search and retrieval. Cohere focuses on enterprise deployments through Oracle Cloud, Amazon SageMaker, Microsoft Azure, and on-premise installations, plus North (the secure enterprise AI platform) for Fortune 500, financial services, and government clients requiring private model deployment.

Google DeepMind

London, United Kingdom

Google DeepMind is the unified AI research division of Alphabet, formed in April by merging Google Brain and DeepMind Technologies. Headquartered in London, United Kingdom with major research operations in Mountain View, California, the company is led by Sir Demis Hassabis (Nobel Prize in Chemistry) and develops the Gemini family of multimodal foundation models including Gemini Ultra, Pro, and Flash. DeepMind also created AlphaFold for protein structure prediction (winning the Nobel Prize), AlphaGo, AlphaZero, and continues research on artificial general intelligence, reinforcement learning, and scientific applications of AI through Google Cloud, Google Workspace, and Bard now Gemini.

Meta AI

Menlo Park, United States

Meta AI is the artificial intelligence research and product division of Meta Platforms (NASDAQ: META), headquartered in Menlo Park, California. The division includes Fundamental AI Research (FAIR) led by Yann LeCun and the GenAI product organization developing the Llama family of open-weight large language models including Llama 3 and Llama 4. Meta releases Llama models with permissive licensing for commercial and research use, powering the Meta AI consumer assistant integrated across WhatsApp, Instagram, Messenger, Facebook, and Ray-Ban Meta smart glasses. Meta also develops Code Llama, the Segment Anything computer vision model, and major contributions to PyTorch and open-source AI tooling.

Mistral AI

Paris, France

Mistral AI is Europes leading foundation model laboratory, founded and headquartered in Paris, France by former Meta and Google DeepMind researchers Arthur Mensch, Guillaume Lample, and Timothee Lacroix. The company develops the Mistral family of open-weight and proprietary large language models including Mistral Large, Mistral Small, Codestral for coding, Mistral Embed for embeddings, and the Pixtral multimodal model. Mistral provides API access through La Plateforme, the Le Chat consumer assistant, and enterprise solutions including on-premise model deployment. The company has emerged as Europes flagship AI champion with major partnerships with Microsoft Azure, IBM, and the French government.

OpenAI

San Francisco, United States

OpenAI is one of the worlds leading artificial intelligence research and deployment companies, founded and headquartered in San Francisco, California. The company develops the GPT family of large language models including GPT-4 and GPT-5, the ChatGPT consumer product with over 800 million weekly active users, the Sora video generation model, the DALL-E image generation model, and the Whisper speech recognition system. OpenAI provides API access for developers, ChatGPT subscriptions for consumers and enterprises, and is preparing for a US public listing reportedly valuing the company at up to one trillion dollars.

xAI

Palo Alto, United States

xAI is an artificial intelligence company founded by Elon Musk and headquartered in Palo Alto, California with major data center operations in Memphis, Tennessee. The company develops the Grok family of large language models including Grok 4 with real-time access to X (formerly Twitter) data and built-in web search. xAI operates the Colossus supercomputer in Memphis featuring over 200,000 Nvidia H100 GPUs making it one of the largest AI training clusters in the world. Products include Grok consumer chat through X Premium, the xAI API for developers, and integration into Tesla and SpaceX product ecosystems.