Ollama raised $65M. Running models locally is now the point.

Ollama — the tool that lets you run open-weight AI models on your own machine in about two minutes — raised a $65 million Series B led by Theory Ventures, with Benchmark, 8VC, and Y Combinator along for the ride. The headline is the money. The number that actually matters is 8.9 million monthly developers and a footprint inside 85% of the Fortune 500. Running a capable model locally, instead of renting one through a metered API, has quietly gone mainstream — and that changes who owns the model layer of your software.

What actually happened

Per TechCrunch, the round brings Ollama's total funding to $88 million. Usage doubled since January to 8.9 million monthly developers, with roughly a million new installs a week and 176,000 GitHub stars. The company runs on 14 employees; founders Jeff Morgan and Michael Chiang previously built Docker Desktop.

The pricing detail is the tell. Ollama runs open-weight models locally for free, and its cloud tier ($0–$100/month) is priced on GPU time, not tokens. No per-token meter. No usage bill that scales with your success. You pull a model — Llama, Qwen, GLM, whatever — and run it. When a better open model ships, you pull that one instead.

Why it matters for your business

Every AI feature you ship sits on top of a model. The question that decides your leverage is: do you rent that model or own the seam? When your feature runs on open weights you can pull and run yourself, the model becomes a swappable, ownable part — not an API a vendor can reprice, deprecate, or gate overnight. We've written before about treating the model as a dial and keeping your stack portable; Ollama is the plumbing that makes it concrete.

You don't have to self-host everything. Most of our builds route heavy reasoning to a hosted frontier model and keep the cheaper, high-volume calls — classification, extraction, drafting — on an open model that can run local or on a rented GPU. That split caps your bill and kills lock-in. Build the seam so the model is a config value, and no vendor's pricing email is ever an emergency.

Sources: TechCrunch — Open-source AI developer tool Ollama raises $65M, grows to nearly 9M users, SiliconANGLE — Open-source AI developer tool Ollama raises $65M to grow its platform.