The founders behind Docker Desktop have raised $88 million to scale their open AI model platform beyond its current 8.9 million developer users.

Jeff Hwang and Michael Grinich, who previously built Kitematic before its 2015 acquisition by Docker, announced the Series A round for Ollama on Tuesday. Benchmark's Peter Fenton led the funding, with participation from Theory Ventures' Tomasz Tunguz and 8VC's Alex Kolicich.

Ollama lets developers download and run open-source AI models locally through a single command-line interface. The platform has become the leading way to access models from labs like DeepSeek, MiniMax, and others without requiring API keys or cloud infrastructure.

Why local AI matters

The company built Ollama around three core principles: ownership, affordability, and privacy. Users can customize models without vendor lock-in, avoid per-token pricing that scales with usage, and keep sensitive data on their own hardware.

"Open models are now enabling that moment for AI," the founders wrote in their announcement. They compared the shift to personal computers moving computing power from mainframes to individual desks.

Ollama now serves 85% of Fortune 500 companies, according to the company. Its cloud service has more than doubled in token volume each month as enterprises adopt open models for production workloads.

The funding will expand Ollama's hybrid inference capabilities and same-day support for newly released open models. The company plans to launch a cloud platform that maintains the privacy and ownership benefits of local deployment while offering enterprise-scale infrastructure.

Notable angels in the round include Docker founder Solomon Hykes, ClickHouse CEO Aaron Katz, and former Twitter engineering head Michael Montano. Y Combinator, Garage Capital, and Pace Capital also participated.

Ollama's growth reflects broader adoption of open-source AI models as alternatives to proprietary services from OpenAI and Anthropic. The platform competes with Hugging Face for developer mindshare in the open model ecosystem.

The company plans to maintain its focus on making powerful AI models "effortless to run for everyone" while keeping the technology open-source.