Ollama, the company behind the popular tool for running open models locally, has announced a $65 million (about 10.5 billion yen) Series B funding round. The round brings Ollama's total funding to $88 million (about 14.2 billion yen). With its developer base growing to 8.9 million and a reported 85% of the Fortune 500 using it, Ollama is cementing its presence as a platform for working with open models.
A $65M Series B, Bringing Total Funding to $88M
Announced on July 9 (local time), the round was led by Tomasz Tunguz of Theory Ventures[1][2]. It also included Peter Fenton of Benchmark, who led the Series A and joined the board, along with 8VC, Y Combinator, Pace Capital, 49 Palms, and GTMFund[1][2].
On its official blog, Ollama says it has "raised $88M," but that figure refers to total funding to date, including this $65 million Series B[1][2]. Angel investors include Docker founder Solomon Hykes, ClickHouse CEO Aaron Katz, Cockroach Labs co-founder and GIMP co-creator Spencer Kimball, and Amp CEO Quinn Slack[1].
※1 USD = about 161 JPY (as of July 10, 2026)[3]
What Is Ollama?
Ollama is open-source software built by co-founders Jeffrey Morgan (CEO) and Michael Chiang[1]. The pair previously created Kitematic, a tool that made Docker easy to use, which Docker acquired in 2015. That work went on to become Docker Desktop, now used by more than 10 million developers[1].
Ollama's appeal is that a single command downloads and runs the latest open models, and a simple API lets developers build on them in their own apps[1]. There is no need to obtain an API key or set up expensive servers; models run on your own machine. Morgan and Chiang compare this to the way the personal computer once brought computing to individual desks, describing it as "the personal computer moment for AI"[1].
Three Principles and Ollama Cloud
Ollama is built around three principles[1]. The first is ownership: open models are yours to keep, customize, and optimize. The second is affordability: running models on your own hardware avoids runaway per-token bills. The third is privacy: running locally means your data never has to leave your machine.
The scale keeps growing. The developer base has reached 8.9 million, with close to one million new installs added each week[2]. Integrations number more than 67,000, and 85% of the Fortune 500 reportedly use Ollama, including in heavily regulated sectors such as government, healthcare, and finance[1][2]. All of this is supported by a team of just 14 employees[2].
Ollama also offers Ollama Cloud, a service for larger models that will not fit on a local machine. It provides access to powerful open models such as GLM, Nemotron, DeepSeek, Kimi, and MiniMax, and the company says its token volume has been more than doubling every month on average[1].
A Bet on Open AI, and the Challenges Ahead
The new funding is expected to go toward supporting new models the day they are released, seamless hybrid inference across local and cloud, and expanding the cloud platform[1]. Morgan and Chiang say their bet that open and easy would win proved right once before, and that they are betting again "with everything we've got"[1].
At the same time, how Ollama balances its "own it locally" philosophy with the expansion of its cloud service is worth watching. Whether it can deliver the same trust to privacy- and ownership-focused developers even through the cloud may shape its future growth. How far a 14-person team can sustain and scale support for 8.9 million users is another point to watch.
Summary
Ollama has closed a $65 million (about 10.5 billion yen) Series B led by Theory Ventures, lifting its total funding to $88 million (about 14.2 billion yen). Backed by 8.9 million developers and adoption across 85% of the Fortune 500, it is solidifying its position as the entry point for running open models on your own hardware. Including how it balances a local-first ethos with cloud expansion, it is a move worth following as the contest for leadership in open AI unfolds.
