Popular AI developer tool Ollama raises $65M, grows to nearly 9M users

Popular open source AI tool Ollama has raised a $65 million Series B round, led by Theory Ventures, founder and CEO Jeff Morgan tells TechCrunch.

This round follows a previous Series A of $15 million led by Peter Fenton of Benchmark. In total, the company has now raised $88 million.

Ollama, launched in 2023, helps devs run open-source AI models on their PCs, get them up and running in minutes. It has been recommended by developers across countless training sites, videos, blogs and social media posts. It has accumulated 176,000 stars and nearly 17,000 forks on GitHub.

Developers can also use Ollama to discover models and access the larger, more complex ones it hosts in its neocloud through several subscription tiers, from free to $100/month. It also tracks usage based on GPU time, not token limits.

If the goal of helping developers easily build on their PCs sounds familiar, it is. Morgan and his co-founder Michael Chiang previously helped create Docker Desktop. They came to Docker after buying their previous store, Kitematic. Docker creates containers that enable cloud applications to be easily moved from cloud to cloud, or from desktop to cloud, removing all complex hardware configuration issues.

So Ollama essentially does for AI what Docker and Docker Desktop do for the cloud.

“The open models started coming out in 2023 but they were difficult to use,” Morgan said. They were intended for researchers at the time, not programmers. “Because of this, it was really hard to get them to work.” Three years after launch, Ollama is now “used by over 8.9 million developers every month, sits in 85% of the Fortune 500 and is growing like crazy,” he said. All have only 14 employees.

That work experience is what drew Benchmark’s Peter Fenton to lead its previous round and join the board.

“What Jeff and Michael built with Docker is used by 10 million developers every day. The creative ability to create a product that goes universal for developers is very rare,” Fenton told TechCrunch.

Morgan and Fenton declined to discuss startup revenues and new valuations. However, Morgan says the proving ground for Ollama as a business happened in January, when OpenClaw was on fire. That’s when the big open models “suddenly were able to do these agent tasks, like writing code. Obviously, we saw an explosion of assistants like OpenClaw, and this idea that open models could get the real work done.”

Since then, the industry has been nervous about the idea that paying users (especially deep-pocketed enterprises and fast-growing AI startups) will turn to affordable open models, keeping their use of closed models like Anthropic to get the necessary foundation.

“I still think this is the most wrong part of the debate. It’s not either/or,” Fenton said of open versus closed AI models. There will be a lot of business for both of them, he argues. However, every company with a high cost of thinking — the cost of running models — has a “critical project at hand” that forces them to move to “open-weight models,” he said.

There is a lot of evidence that such startups and businesses are already opening models for their daily needs. That, obviously, bodes well for Ollama’s cloud business.

But more interestingly, Ollama is another example of how AI is creating a huge new crop of open source projects that turn into companies followed by VCs. There are open source idea providers such as Inferact, maker of vLLM, and RadixArk, maker of SGlang. There is OpenClaw and its alternatives like NanoClaw. There are also small startups that build their open models from scratch, like Arcee.

To be sure, not all of Ollama’s fans are happy that the company was pursuing a living. About a year ago, a number of blog and social media posts complained that its cloud business was drawing attention away from its popular free project and cited Ollama as an example of the so-called “Enshittification” of dev tools, as the trend is called.

But Morgan sees its cloud service as an evolution of its open source mission to help programmers find and use models more easily. Those high-end, big, open models are often “too big to run on your computer. So we said, ‘Hey, let’s help get a computer for that,'” he explained.

Board member Fenton adds, “Nothing has changed in the main free product on the desktop. There is no change in the principle that this is a place where you can find and use local models.”