Ollama, the open source platform enabling developers to run large language models locally on consumer hardware, raised $65 million in fresh funding led by Benchmark. The round values the company at a substantial valuation, though specifics remain undisclosed.

The funding milestone reflects explosive growth. Ollama now serves nearly 9 million users and has accumulated 176,000 GitHub stars alongside 17,000 forks. These numbers position it as one of the most-adopted developer tools in the AI infrastructure space.

Ollama simplifies local AI deployment. Rather than relying on cloud APIs or expensive GPU clusters, developers download models like Llama 2 or Mistral and run them directly on their machines. This approach appeals to engineers building AI applications without incurring ongoing API costs, managing data privacy concerns, or dealing with latency issues tied to remote inference.

The competitive landscape includes several players. Hugging Face dominates model hosting and discovery. Replicate offers simpler API-based inference. vLLM provides optimization for serving models at scale. But Ollama occupies a distinct niche: making local inference accessible to everyday developers without deep ML infrastructure expertise.

Benchmark's backing carries weight. The firm backs numerous infrastructure startups including Datasette creator Simon Willison's Data layer ventures. The partnership suggests confidence in Ollama's path to becoming critical developer infrastructure.

The funding fuels ambitions in a hot market. Open source AI tooling has drawn significant capital as enterprises and startups seek alternatives to proprietary AI platforms. Ollama's growth reflects broader developer appetite for running models locally, driven by cost concerns, privacy requirements, and the desire for offline-capable AI applications.

Earlier funding rounds and the Benchmark lead indicate strong investor confidence in Ollama's staying power. The tool's accessibility and the developer community's embrace position it as foundational infrastructure in