Ollama: Local AI's Future
Ollama, the open-source AI platform, secures $65 million Series B funding.
Key Points
Ollama offers neocloud hosting for larger open models, enabling developers to run more complex models remotely with the same local user experience; pricing tiers run from free to $100/month and are based on GPU time rather than tokens.
Founders Jeff Morgan and Michael Chiang previously built Docker Desktop and led the Kitematic startup before Docker acquired it, informing Ollama’s container-centric approach to running open models.
Ollama maintains a very lean team, with only about 14 employees, despite rapid growth and broad adoption.
In addition to Fortune 500-scale usage, Ollama notes customers in regulated industries such as government, healthcare, and finance, highlighting its appeal to regulated sectors.
The company’s Series A was a $15 million round led by Benchmark, with Peter Fenton joining the board, underscoring early investor interest in open-model tooling.
Ollama, the open-source tool that lets developers run AI models directly on their own computers, has raised $65 million in a Series B round led by Theory Ventures, according to finsmes. The deal brings Ollama's total funding to $88 million and cements its place as a leading platform in the fast-growing local AI space.
The company says roughly 8.9 million developers use Ollama every month, and its GitHub repository has collected 176,000 stars, per cryptobriefing. Ollama also claims its tools run inside 85% of Fortune 500 companies — a remarkable footprint for a team of just 14 employees.
Founders Who Built Docker Now Simplify AI
Ollama was co-founded by Jeff Morgan and Michael Chiang, who previously built Docker Desktop and led Kitematic before Docker acquired it. That background shaped Ollama's core idea: make running complex AI models as simple as running a software container.
The company's pitch is straightforward. Developers download a model and run it locally on their own machine — no cloud account needed, no usage fees per question asked. For bigger tasks, Ollama also offers cloud hosting that works the same way, so developers don't have to rewrite their code to scale up, according to zamin.
GPU-Based Pricing Targets Enterprises Needing Cost Certainty
Ollama runs on a freemium model. Its paid cloud tiers go up to $100 per month and charge based on GPU time — not per token or per API call. That matters for enterprises, because it makes costs predictable, according to whalesbook.
The company has found particular traction in regulated industries. Ollama counts customers in government, healthcare, and finance — sectors where sending data to a third-party cloud can raise legal and privacy red flags. Running models locally sidesteps many of those concerns entirely.
Theory Ventures Leads, Benchmark Returns for Series B
Theory Ventures led this $65 million Series B. Benchmark, 8VC, and other investors also participated, per finsmes. Benchmark previously led Ollama's $15 million Series A, with partner Peter Fenton joining the board at that time, showing strong early conviction in open-model tooling.
The back-to-back rounds reflect a broader investor bet on AI that runs outside the big cloud providers. As more companies seek to avoid locking into OpenAI or Google, platforms like Ollama — which work with open-weight models like Meta's Llama — are drawing serious attention and capital.
Tiny Team, Massive Reach — What Comes Next
With only 14 employees and $88 million in total funding now in hand, Ollama is in a rare position. The company has proven it can grow fast and lean. The new capital gives it room to expand its cloud infrastructure and support more and larger open-weight models, according to zamin.
The broader trend is clear. Developers increasingly want AI tools they control — tools that run on their own hardware, keep data private, and don't send every query to a remote server. Ollama has built exactly that, and investors are betting the demand will keep growing, per citybiz.
Sentiment
Mostly PositivePublishers
11
Articles
9
Reach
20
