The popular open source AI utility Ollama has secured a $65 million Series B round led by Theory Ventures, according to founder and CEO Jeff Morgan.

This funding follows an earlier $15 million Series A led by Benchmark's Peter Fenton. In total the company has now raised $88 million.

Ollama, which debuted in 2023, assists developers in running open-weight AI models on personal computers and getting them operational within minutes. It has earned praise from developers through numerous tutorials, videos, articles and online discussions. The project has collected 176,000 stars and nearly 17,000 forks on GitHub.

Developers can also leverage Ollama to discover models and reach larger, more intricate ones hosted on its cloud platform through various subscription levels, ranging from free to $100 per month. Usage is measured by GPU time rather than token counts.

If the goal of helping developers build more easily on their PCs seems familiar, that is understandable. Morgan and co-founder Michael Chiang previously contributed to Docker Desktop after Docker acquired their earlier venture, Kitematic. Docker provides containers that simplify moving cloud applications between environments while hiding hardware setup complexities.

Thus Ollama has essentially replicated for AI what Docker achieved for cloud infrastructure.

"Open models began appearing in 2023 yet they proved difficult to operate," Morgan noted. At the time they targeted researchers rather than programmers, making setup challenging. Three years after launch, Ollama serves over 8.9 million developers monthly, appears in 85 percent of the Fortune 500 and continues rapid growth, all with a team of just 14 employees.

That background is what attracted Benchmark's Peter Fenton to lead the prior round and join the board.

"What Jeff and Michael created at Docker reaches more than 10 million developers daily. The ability to build a product that achieves widespread developer adoption is uncommon," Fenton said.

Morgan and Fenton chose not to share revenue figures or the latest valuation. Morgan indicated the business validation arrived around January when open models gained traction for agentic tasks such as coding, coinciding with the rise of tools like OpenClaw and demonstrating that open models can handle meaningful work.

Since then the sector has focused on how paying customers, especially enterprises and growing AI startups, may shift toward cost-effective open models while using closed options like those from Anthropic only when necessary.

"I still believe the discussion often misses the mark. It is not an either-or situation," Fenton observed regarding open versus closed models. Both approaches will find demand, yet firms facing high inference costs face a critical need to adopt open-weight models.

Evidence shows startups and enterprises already rely on open models for routine operations, which supports Ollama's cloud offerings.

Even more notable, Ollama illustrates how AI is spawning new open source initiatives that attract venture funding. Examples include inference providers such as the maker of vLLM and the creator of SGLang, along with projects like OpenClaw and its peers, plus emerging efforts to build open models from the ground up.

Not every Ollama supporter welcomed the move toward commercialization. Roughly a year ago some posts criticized the cloud service for diverting focus from the free project and labeled it an instance of dev-tool commercialization trends.

Morgan views the cloud service as an extension of the open source goal to help programmers locate and use models. Many advanced open models are too large for local hardware, so the company offers compute access for those cases.

Board member Fenton noted that the core free desktop product remains unchanged, preserving the original promise of a place to find and run local models.