Ollama raises $65M in Series B led by Theory Ventures
Ollama, the open source AI tool that lets developers run models on their own machines, has closed a $65 million Series B. Theory Ventures led the round, founder and CEO Jeff Morgan confirmed to TechCrunch. The company has now raised $88 million total, following a $15 million Series A led by Benchmark’s Peter Fenton.
That’s a serious vote of confidence for a startup with just 14 employees. Ollama launched in 2023 with a simple pitch: make open-weight AI models actually usable on a developer’s PC, with setup measured in minutes, not days.
The numbers back up the hype. Ollama now counts over 8.9 million developers using the tool every month. Morgan says it’s sitting inside 85% of the Fortune 500. On GitHub, the project has racked up 176,000 stars and nearly 17,000 forks.
From Docker to AI: the founders’ playbook
If the story feels familiar, it should. Morgan and co-founder Michael Chiang previously built Docker Desktop. They landed at Docker after the company acquired their earlier startup, Kitematic. Docker’s containers made cloud apps portable, abstracting away hardware headaches.
Ollama essentially did for AI what Docker did for cloud. The timing mattered. “Open models started coming out in 2023 but they were really hard to use,” Morgan said. They were built for researchers, not programmers. “As a result, it was really hard to get them up and running.”
That experience is exactly what drew Benchmark’s Peter Fenton to lead the earlier round and join the board. “What Jeff and Michael built with Docker is being used by 10 million-plus developers every day,” Fenton told TechCrunch. “The creative powers to create a product that goes to ubiquity for developers is extremely rare.”
What developers get from Ollama
Ollama isn’t just a local runtime. The company also hosts larger, more complex models on its neocloud, with subscription tiers ranging from free to $100 per month. Notably, it tracks usage by GPU time rather than token limits — a deliberate choice that sets it apart from API-based competitors.
Developers can use Ollama to:
- Discover and pull open-weight models directly from their terminal
- Run models locally on their own hardware
- Access bigger models via Ollama’s hosted neocloud
- Switch between models without reconfiguring their environment
The open model moment arrived in January
Morgan says the real inflection point for Ollama as a business came around January, when OpenClaw went hot. Suddenly, larger open models “became able to do these agentic tasks, like coding,” he said. “Obviously, we saw the explosion of the assistants like OpenClaw, and this idea that open models can get real work done.”
That shift has fueled industry-wide speculation that enterprises and AI application startups will move away from closed models like Anthropic’s toward cheaper open alternatives. Fenton pushes back on the framing. “I still think that this is the part that most of the debate gets wrong. It’s not an either/or,” he says. Both ecosystems will thrive, he contends. But every company with heavy inference expenses faces a “vital existential project” pushing them toward open-weight models.
That dynamic bodes well for Ollama’s cloud business, which monetizes exactly those enterprise workloads.
Ollama is part of a bigger open source wave
Ollama isn’t alone. AI is spawning a new crop of open source projects turning into VC-backed companies. There are inference providers like Inferact (maker of vLLM) and RadixArk (maker of SGLang). There’s OpenClaw and its alternatives like NanoClaw. Even tiny startups like Arcee are building open models from scratch.
But not every Ollama fan has cheered the company’s commercial push. About a year ago, a wave of blog posts and social media complaints accused the cloud business of distracting from the beloved free project. Critics cited Ollama as an example of the so-called “enshittification” of dev tools.
Morgan sees it differently. The cloud service is an evolution of the open source mission, he argues. Those state-of-the-art open models are often “too big to run on your own computer. So we said, ‘Hey, let’s help find the compute for that.'”
Fenton echoes that. “Nothing has changed for the core product that’s free on the desktop. There’s zero change to the premise that this is the place you can discover and run local models.”
For developers watching the open source AI space, Ollama’s trajectory is worth tracking. The tool that started as a convenience for running models locally is now a serious business — and one of the clearest signals yet that open-weight AI has moved from hobbyist territory to enterprise infrastructure.
