we_are_coded.by CODE · The world, decoded
БГ

Ollama · investment

Ollama raised 88 million dollars to stay free and local

88 million dollars for a tool whose whole value is being free and local - Ollama raised funding and now serves 8.9 million developers. The number alone shows local AI work stopped being a niche whim a long time ago.

Source
Ollama Blog
Published
The event
Checked on
Responsible editor
Cvetelin Ivanov
How we work
MethodCorrections

You download an open model, it runs on your machine. No cloud, no stranger's server, no token passing through a third party. That's Ollama. For a team that works locally and doesn't trust anything by default, this isn't exotic. It's daily life.

Fact

The facts: Ollama announced 88 million dollars in funding on 9 July 2026, in a post titled All aboard open models. Among the investors are Peter Fenton of Benchmark, Tomasz Tunguz of Theory Ventures, Alex Kolicich of 8VC, the Y Combinator fund (a well-known startup investment fund), as well as founders like Solomon Hykes (Docker), Aaron Katz (ClickHouse) and Spencer Kimball (Cockroach Labs). The tool serves 8.9 million developers.

8.9 million people on a tool that, by design, sends nothing out - a serious number, precisely because it's chosen by people who care about control. But 88 million in venture capital (money from investment funds) in something whose whole value is being free and local raises the question of return. An investor doesn't put in money for a good cause. Sooner or later they want an exit. And the exit usually runs through a cloud, an enterprise tier, or a paid layer on top of the free one.

Free and local is a product decision, until it becomes an obstacle to returning the capital.

We keep using Ollama as before, but we're watching three things. Whether the local mode stays fully capable. Whether an enterprise version appears that pushes the local one to second class. And whether the license on the tool itself changes. Until then, it's a tool. Not a promise.

Official primary sources

Ollama Blog - All aboard open models

The visual is generated code-art, no third-party images.