The easiest way to run open models locally — the developer default for on-device AI
HQ: San Francisco, CA, United States | Founded: 2023 | Employees: ~40 (2026, est.) | Stage: Early Stage VC (Series B) | Website: https://ollama.com
Ollama was founded in 2023 by Jeff Morgan (previously a principal engineer at LinkedIn) with a simple but powerful idea: running an open-weights language model should be as easy as running any other developer tool. The result is a single, polished CLI and app that downloads, runs, and manages local models — with a model library covering Llama, Mistral, Qwen, DeepSeek, Gemma, and thousands of community models, plus an OpenAI-compatible API so local models drop into existing applications.\n\nThe product spread entirely through developer word-of-mouth: it became the fastest way to try open models privately, without cloud costs or data leaving the machine. By 2025-2026 Ollama's reach was extraordinary for a tiny team — 8.9 million monthly developers, ~85% penetration among Fortune 500 companies (who use it for private, on-premise AI experimentation), and deep integration into the broader AI toolchain (VS Code extensions, Docker, Homebrew, and countless tutorials).\n\nFinancially, Ollama raised a $22M Series A (2024, reportedly led by Benchmark) and a $65M Series B led by Theory Ventures in July 2026, bringing total funding to $88M — with Benchmark, 8VC, Y Combinator, and Docker founder Solomon Hykes among backers. The company's strategy is to stay the neutral, open default for local and on-premise model execution while building commercial products (Ollama Cloud, enterprise offerings) on top of the massive distribution. Its founding from Y Combinator (S24 batch) and its position as the top-of-mind local-AI brand make it one of the most influential small companies in the AI stack.
Ollama is the 'Docker for AI models': it took the friction out of running open-weights models locally and became the developer default for on-device AI — 8.9M monthly developers and ~85% Fortune 500 penetration by mid-2026, a staggering reach for a ~40-person company. It rides the open-models wave (Llama, Mistral, Qwen, DeepSeek and thousands of community models run locally via Ollama), and investors bet it becomes a durable platform layer: $88M total funding including a $65M Series B led by Theory Ventures (July 2026) with Benchmark, 8VC, and YC. The bull case is local-AI gravity — privacy, cost, and offline capability — plus a brand that owns 'run models locally'; the bear case is that the core product is free open source, monetization is still early, and local compute keeps getting commoditized by hardware and cloud defaults.
Browse: Home | Blog | About | All Startups A-Z | View full profile on StartupWiki