PALO ALTO, Calif.: Ollama closes a $65 million Series B led by Theory Ventures. The number worth staring at, though, is fourteen. That’s the entire headcount.
Those 14 people run a platform that 8.9 million developers open every month. Usage has doubled since January. Installs are running close to a million a week, and the tool now sits inside 85% of the Fortune 500, including banks, hospitals, and government agencies.
The July 9 round brings total funding to $88 million. Benchmark, 8VC, Y Combinator, Pace Capital, 49 Palms, and GTMFund all join. The angel list reads like an infrastructure hall of fame: Docker founder Solomon Hykes, ClickHouse CEO Aaron Katz, Cockroach Labs’ Spencer Kimball. Valuation stays undisclosed.
Ollama’s product is almost aggressively simple. One terminal command pulls an open-weight model such as Llama, DeepSeek, Gemma, or Qwen and runs it on your own laptop or server. No cloud API, no per-token bill, no prompt leaving the building. More than 67,000 community integrations plug it into coding assistants and agents.
Founders Jeff Morgan and Michael Chiang built Docker Desktop before this, which explains a lot. Benchmark’s Peter Fenton, who led the Series A and sits on the board, states the thesis plainly: “Open-weight models will generate the supermajority of tokens” within two years.
The contrast with rivals is stark. Hugging Face has raised over $395 million. Together AI just took $800 million at $8.3 billion. Ollama’s $88 million total is a rounding error next to either, yet it claims the largest developer network in the open-model world.
Not everyone cheers. Parts of the community accused Ollama of enshittification last year over its paid cloud tier. The Fortune 500 apparently disagrees.
