The popular open source AI tool Olama has raised a $65 million series round, led by Theory Ventures, founder and CEO Jeff Morgan tells TechCrunch.
This round follows an earlier $15 million Series A led by Benchmark’s Peter Fenton. In total, the company has now raised $88 million.
Launched in 2023, Ollama helps developers run open-source AI models on their computers, getting them up and running in minutes. It has been praised by developers to countless education locations, video, histology and social media posts. It has collected 176,000 stars and almost 17,000 forks on GitHub.
Developers can also use Ollama to find models and access larger, more complex ones that it hosts on its neocloud through several subscription tiers, from free to $100/month. It also tracks usage based on GPU time, not token limits.
If the mission to help developers build more easily on their PCs sounds vaguely familiar, it should. Morgan and co-founder Michael Chiang previously helped build Docker Desktop. They landed at Docker after it bought their previous startup, Kitematic. Docker makes containers that help cloud applications move easily from cloud to cloud or desktop to cloud, removing all the pesky hardware configuration issues.
So Ollama essentially did for AI what Docker and Docker Desktop did for the cloud.
“The open models started coming out in 2023, but they were really hard to use,” Morgan said. They were geared towards researchers at the time, not developers. “As a result, it was very difficult to get them up and running.” Three years after its launch, Ollama “is now used by more than 8.9 million developers every month, in 85% of the Fortune 500, and growing like crazy,” he said. All with only 14 employees.
It is this professional experience that prompted Benchmark’s Peter Fenton to lead the previous round and join the board.
“What Jeff and Michael built with Docker is used by more than 10 million developers every day. The creative forces to build a product that is ubiquitous for developers are extremely rare,” Fenton told TechCrunch.
Morgan and Fenton declined to discuss the startup’s revenue and new valuation. But Morgan says the proof point for Ollama as a business happened around January, when OpenClaw became hot. That’s when the larger open models “suddenly were able to do these representative tasks, like coding. Obviously, we saw the explosion of helpers like OpenClaw and the idea that open models can do real work.”
Since then, the industry has been abuzz with the idea that paying users (especially deep-pocketed enterprises and fast-growing AI startups) will increasingly turn to more affordable open models, reserving the use of closed models like Anthropic for more base as needed.
“I still think that’s where most of the debate is wrong. It’s not either/or,” Fenton says of open versus closed AI models. There will be plenty of work for both, he argues. But any company with high inference costs — the cost of using the models — has a “vital existential task” that pushes them to move “to open weight models,” he says.
There is plenty of evidence that such startups and enterprises are already turning to open models for their day-to-day needs. This obviously bodes well for Ollama’s cloud business.
But even more interestingly, Ollama is another example of how AI is spawning a great new crop of open source projects that are turning into VC-pursued companies. There are open source inference providers like Inferact, vLLM maker, and RadixArk, maker of SGLang. There is OpenClaw and its alternatives like NanoClaw. There are even tiny startups building their own open models from scratch, like Arcee.
To be sure, not all Ollama fans were happy that the company was trying to make a living. About a year ago, a bunch of blog and social media posts she complained that her cloud business was taking attention away from her beloved free project and cited Ollama as an example the so-called “Enshittification” of dev toolsas the voltage is called.
But Morgan sees its cloud service as an evolution of its open source mission to help developers find and use models easily. These state-of-the-art, large, open models are often “too big to run on your computer. So we said, ‘Hey, let’s help figure out the computation for this,'” he explained.
Board member Fenton adds, “Nothing has changed about the core product that’s free on PC. There’s no change in the premise that this is the place to discover and run local models.”
When you purchase through links in our articles, we may earn a small commission. This does not affect our editorial independence.
