Ollama, an open-source AI tool that helps developers run open-weight models locally and via a hosted cloud, has closed a $65 million Series B round led by Theory Ventures, founder and CEO Jeff Morgan told TechCrunch. The financing follows a $15 million Series A led by Benchmark partner Peter Fenton; total funding to date is $88 million.
What Ollama does
Launched in 2023, Ollama enables developers to get open-weight models running on their PCs within minutes. The project has gained significant traction among developers: it has roughly 176,000 stars and nearly 17,000 forks on GitHub. In addition to the desktop client, Ollama offers model discovery and hosted access to larger models through its so-called neocloud across several subscription tiers, ranging from a free tier to $100 per month. Usage is billed based on GPU time rather than token limits.
Founders and background
Jeff Morgan and co-founder Michael Chiang previously helped build Docker Desktop after selling an earlier startup, Kitematic, to Docker. The team draws a parallel between what Docker did for cloud portability and what Ollama aims to do for AI: reduce friction so developers can run and experiment with models without wrestling with hardware configuration.
Scale and team size
According to Morgan, Ollama is now used by over 8.9 million developers each month, is present at 85% of the Fortune 500, and operates with just 14 employees. The company declined to disclose current revenue figures or a new valuation.
Market timing and strategy
Morgan says a proving moment came around January, when OpenClaw and other large open models became capable of "agentic" tasks like coding. That shift reinforced the idea that paying customers — especially well-funded enterprises and fast-growing AI application startups — may increasingly choose cost-effective open models for routine workloads while reserving closed models such as Anthropic for specific needs.
Peter Fenton, who led Ollama's earlier round and joined the board, argues that the market for open and closed models is not mutually exclusive: both will have significant business. Still, he believes many companies facing high inference costs have a compelling reason to migrate toward open-weight models.
Ecosystem and competitors
Ollama's rise is part of a larger movement in which open-source AI projects are becoming viable startups pursued by venture capital. Other open-source inference providers and projects mentioned in the discussion include Inferact (maker of vLLM), RadixArk (maker of SGLang), OpenClaw and alternatives like NanoClaw, as well as very small teams building models from scratch, such as Arcee.
Community concerns and company response
Not all community reaction has been positive: about a year ago a number of blog and social media posts criticized Ollama's cloud business, arguing it diverted attention from the free project and citing it as an example of the so-called "enshittification" of developer tools.
Morgan frames the cloud offering as an extension of Ollama's open-source mission: many state-of-the-art open models are "too big to run on your own computer," so the company aims to help developers find the necessary compute. Board member Peter Fenton added that nothing has changed for the core free desktop product; it remains the place to discover and run local models.
