What Ollama raised, and what it is
On 9 July 2026, Ollama said it had raised 65 million dollars in a Series B round led by Theory Ventures, with Benchmark, 8VC, Y Combinator and others taking part, bringing its total funding to 88 million dollars. If the name is unfamiliar, the tool is not: Ollama lets a developer run an open AI model on their own hardware with a single command, then scale up to larger open models in Ollama's own cloud when they need more power. It is the plumbing many teams use to run AI locally rather than calling a cloud provider for every request.
The numbers behind the round are unusual. Ollama says 8.9 million developers use it, across more than 67,000 integrations, and that it runs inside 85 percent of the Fortune 500 - all built by a team of just 14 people, two of them former Docker Desktop engineers. The founders would not discuss revenue or a valuation. So the amount raised is public, but what the company is worth, and how it plans to earn, are not.
Why this matters if you self-host AI
For many operators, running models locally is not a preference but a strategy. Self-hosting keeps sensitive data off third-party APIs, avoids a bill that scales with every token, and reduces dependence on a US cloud - the practical backbone of a lot of European data-sovereignty plans. Ollama is, for a large share of those teams, how that actually gets done. Its becoming a funded company is a signal that local, open AI is now a real market rather than a hobbyist corner.
The tension is in the business model. Ollama is free where it runs on your hardware, but the growth path the round is meant to fund points toward its paid cloud - the familiar open-core pattern where the local tool stays free and the money gathers around the hosted tier. That is not a betrayal; it is how these companies survive. But it does mean the thing you may be building your independence on now has investors who will, in time, expect a return.
What to check before you build on it
Treat the funding as a reason to read the fine print, not to relax. The valuation and revenue are undisclosed, which tells you the monetization strategy is still being set - and that is exactly the moment to understand what stays free. Look at which capabilities live in the open tool versus the paid cloud, what the licence permits, and whether anything you depend on today could move behind that cloud tomorrow. An open-core tool is only a sovereignty asset for as long as the part you rely on stays open.
The discipline is portability. Keep your model weights and your data in formats you control, so that if Ollama's roadmap moves value toward its cloud, you can run the same open models another way without a rebuild. The base rate is clear enough: free developer tools that raise this kind of money almost always shift some value behind a paywall eventually. Plan for that now, and the investment works in your favour; ignore it, and you find out the terms when they change.
Read next: The EU Product Passport Drives Its First Takeover | Europe's Shift Software Enters Its Roll-Up Phase
