Ollama lands $65M Series B as open-source AI tool passes 9M monthly developers

Ollama, an open-source tool that lets developers run machine-learning models on personal computers, has secured $65 million in Series B funding led by Theory Venture, bringing the startup's total capital raised to $88 million. Founder and CEO Jeff Morgan disclosed the funding round to TechCrunch.

The platform has grown dramatically since its 2023 launch, now serving more than 8.9 million monthly users across 85% of Fortune 500 companies—all with a lean team of just 14 people.

A faster path to AI development#

Ollama streamlines the process of deploying open-weight AI models by packaging them for quick setup on standard computers. Rather than requiring specialized infrastructure or extensive configuration, the tool gets developers working with state-of-the-art models within minutes. The project has accumulated 176,000 GitHub stars and nearly 17,000 forks, signaling robust community adoption.

The tool's popularity stems from solving a real friction point. When open models began proliferating in 2023, they were largely tailored for academic researchers rather than working developers. Ollama's approach brought accessibility to that ecosystem.

Building on proven success#

Morgan's track record informed investor confidence in the round. Together with co-founder Michael Chiang, he previously helped create Docker Desktop—the breakthrough containerization tool that simplified cloud development. Benchmark's Peter Fenton, who led Ollama's $15 million Series A and now sits on the board, attributed his conviction to that history: "The creative powers to create a product that goes to ubiquity for developers is extremely rare."

Fenton sees parallels between what Docker did for cloud infrastructure and what Ollama is accomplishing for AI: making powerful technology accessible to everyday developers without requiring deep infrastructure knowledge.

Expanding beyond the desktop#

While the free desktop tool remains at the core, Ollama has begun offering cloud-hosted services, allowing developers to access larger models that exceed typical local hardware capacities. Subscription tiers range from free to $100 monthly, with billing tied to GPU usage rather than token consumption.

The business case crystallized around January, when open-source models demonstrated real-world capability in tasks like code generation and autonomous operations. This shift convinced Morgan and Fenton that enterprises with significant inference costs—the computational expense of running models—would increasingly adopt affordable open alternatives alongside proprietary services.