Popular open source AI tool Ollama has raised $65 million in Series B led by Theory Ventures, founder and CEO Jeff Morgan tells TechCrunch.
This round follows a previous $15 million Series A led by Benchmark’s Peter Fenton. In total, the company has now raised $88 million.
Released in 2023, Ollama helps developers run open-weight AI models on their PC and get up and running in minutes. It has been praised by developers through countless training sites, videos, blogs, and social media posts. GitHub has accumulated 176,000 stars and approximately 17,000 forks.
Developers can also use Ollama to search for models and access large and complex models hosted on Neocloud through several subscription tiers ranging from free to $100 per month. It also tracks usage based on GPU time rather than token limits.
Our mission to help developers build more easily on the PC may sound vaguely familiar, and it should be. Morgan and his co-founder Michael Chan previously worked together to build Docker Desktop. They came to Docker after acquiring their previous startup, Kitematic. Docker creates containers that make it easy to move cloud apps from cloud to cloud or from desktop to cloud, abstracting away all the troublesome hardware configuration issues.
So Ollama essentially did for AI what Docker and Docker Desktop did for the cloud.
“Open models started appearing in 2023, but they were very difficult to use,” Morgan said. At the time, they were aimed at researchers, not programmers. “As a result, it was very difficult to get them up and running.” Now, three years after its launch, Ollama is “used by more than 8.9 million developers every month, ranks 85% of the Fortune 500, and is growing at breakneck speed,” he said. We have just 14 employees.
That career experience led Benchmark’s Peter Fenton to lead an early round and join the company’s board.
“What Jeff and Michael built with Docker is used by more than 10 million developers every day. The creativity to create a product that is ubiquitous to developers is extremely rare,” Fenton told TechCrunch.
Morgan and Fenton declined to discuss the startup’s revenue or new valuation. But Morgan said Ollama’s proving point as a business came around January, when OpenClaw gained traction. That’s when larger open models “suddenly became able to perform agent tasks, such as coding. Clearly, assistants like OpenClaw exploded, and we started to see the idea that open models could do real work.”
Since then, the industry has been buzzing with the idea that paying users (particularly deep-pocketed companies and fast-growing AI application layer startups) will increasingly move to more affordable open models, securing the use of closed models like Anthropic’s as needed.
“I still think this is where most discussions get it wrong: It’s not an either/or,” Fenton says of open versus closed AI models. There will be plenty of business for both, he argues. But any company with high inference costs, or the cost of using a model, has “significant survival projects” driving the move to “open-weight models,” he said.
There is plenty of evidence that these startups and companies are already turning to open models to meet their everyday needs. This clearly bodes well for Ollama’s cloud business.
But more interestingly, Ollama is another example of how AI is creating new large open source projects that are turning into companies pursued by VCs. There are open source inference providers such as Inferact, creator of vLLM, and RadixArk, creator of SGLang. There is OpenClaw and its alternatives (such as NanoClaw). There are even small startups like Arcee that are building their own open models from scratch.
Admittedly, not all Orama fans are happy with the company’s pursuit of making a living. About a year ago, a number of blogs and social media posts complained that the company’s cloud business was drawing attention away from its beloved free projects, citing Ollama as an example of the so-called “Enshittification” of development tools (as the trend is called).
But Morgan sees its cloud service as an evolution of open source’s mission of making it easier for programmers to discover and use models. These cutting-edge, large-scale open models “are often too big to run on your own computer, so we said, ‘Let’s help you find the computing for that,'” he explained.
Board member Fenton added, “Nothing has changed about our core product, which is free on the desktop. The premise remains that this is where you can find and run local models.”
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
