Ollama Raises $65M Series B As Open Model Runner Hits 9M Developers
The developer platform that packages open-weight models into a one-line install closed a $65 million Series B led by Theory Ventures, taking total funding to $88M.
Key Takeaways
Ollama closed a $65 million Series B led by Theory Ventures on July 9, 2026, bringing total funding to $88 million three years after founding.
Benchmark, 8VC, Y Combinator, Pace Capital, 49 Palms and GTMFund joined the round; Benchmark's Peter Fenton remains on the board.
Usage hit 8.9 million monthly developers, up from 4.5 million in January, with nearly 1 million new installs per week and presence in 85% of the Fortune 500 - all with a team of 14.
Ollama pairs one-command local model runs with its own GPU cloud, billed by GPU time rather than per token, hosting open models like DeepSeek, Nemotron, GLM, Kimi and MiniMax via deals with labs and NVIDIA, AMD, Intel and Qualcomm.
Co-founders Jeff Morgan and Michael Chiang previously built Kitematic, which became Docker Desktop, and are running the same platform playbook for open-source AI as open models close the gap with closed frontier models.
Kaan Tınmaz
Ollama, the developer tool that turns any open-weight large language model into a one-line install-and-run command, has closed a $65 million Series B led by Theory Ventures, taking the three-year-old company's total funding to $88 million as it doubles down on open-source AI.
Theory Ventures leads with Benchmark, 8VC and Y Combinator
The round was announced on July 9, 2026 in a blog post titled “All aboard open models” and reported first by TechCrunch. Benchmark, 8VC, Y Combinator, Pace Capital, 49 Palms and GTMFund joined the round along with a group of angel investors. Peter Fenton, the Benchmark partner who led Ollama's earlier round, remains on the board.
An open-model bet, with a Docker-era playbook
Ollama's pitch is straightforward: a developer downloads an open model and runs it on a laptop with a single command; if the model is too big for local hardware, Ollama's own GPU cloud takes over, billed by GPU time rather than per token. Co-founders Jeff Morgan and Michael Chiang previously built Kitematic, which became Docker Desktop after Docker's 2015 acquisition. Ollama runs the same play for AI.
The company says 8.9 million developers now use it every month, up from 4.5 million in January, with almost one million new installs a week. It sits inside 85% of the Fortune 500, including regulated verticals such as government, healthcare and finance, all delivered by a team of 14.
Cloud hosts Nemotron, GLM, DeepSeek, Kimi and MiniMax
Ollama's cloud runs heavyweight open models including DeepSeek, Nemotron, GLM, Kimi and MiniMax on day one via distribution deals with the labs behind them and with silicon vendors NVIDIA, AMD, Intel and Qualcomm. Theory Ventures partner Tomasz Tunguz frames Ollama as the platform layer everything else plugs into.
Why the timing matters
The raise lands as open-weight models close in on closed frontier models on quality and cost. Fenton told TechCrunch that open versus closed is “not an either/or,” but companies with heavy inference bills have a strong incentive to migrate. Morgan pegs the turning point around January, when open models became good enough for agentic coding work — the same shift that's fueling raises across the open-source AI stack, including Generalist AI and Prime Intellect.
Reporting based on coverage from The Next Web and TechCrunch.
Ollama raised a $65 million Series B led by Theory Ventures, with participation from Benchmark, 8VC, Y Combinator, Pace Capital, 49 Palms, GTMFund and angel investors, taking total funding to $88 million.
What does Ollama do?
Ollama lets developers download and run any open-weight large language model on a laptop with a single command; if a model is too big for local hardware, Ollama's GPU cloud takes over, billed by GPU time instead of per token.
How many developers use Ollama?
About 8.9 million developers use Ollama every month, up from 4.5 million in January 2026, with almost one million new installs a week and adoption inside 85% of the Fortune 500, including government, healthcare and finance.
Why is the funding timing significant?
The raise comes as open-weight models approach closed frontier models on quality and cost; co-founder Jeff Morgan says the turning point was around January, when open models became good enough for agentic coding work, driving raises across the open-source AI stack.
