A single command can bring up an open-source large model locally. This tool, which has freed countless developers from API keys and exorbitant bills, has just secured the fuel for its next phase. On July 9, Ollama announced in its official blog that it has completed an $88 million funding round. The investor list spans top-tier venture capital and veterans of the open-source world: Benchmark's Peter Fenton, Theory Ventures' Tomasz Tunguz, and 8VC's Alex Kolicich led the round, with Solomon Hykes, founder of Docker, Aaron Katz, CEO of ClickHouse, Spencer Kimball, co-founder of GIMP and co-founder of Cockroach Labs, Quinn Slack, CEO of Amp, Marianna Tessel, a Cisco board member, Michael Montano, former engineering head at Twitter, as well as Y Combinator, Garage Capital, Pace Capital, 49 Palms, and GTMFund among others as angel investors.
The company's founders, Jeff and Michael, have been doing the same thing for ten years. The two met in college and founded Kitematic, which made running Docker extremely simple. In 2015, Docker acquired it, and their work later became Docker Desktop, now used by over ten million developers worldwide. Ten years later, they applied the same philosophy of making complex technology accessible to all to AI—Ollama's goal is to allow developers to start and run open-source models with the least effort. Beyond their wildest imagination, this platform has already served 8.9 million developers and is adopted by 85% of Fortune 500 companies.
In Ollama's narrative, open-source models are ushering in a PC-like moment for AI: just as PCs brought computing power from mainframe rooms to every desktop, allowing people to own, customize, and build upon them, open-source models are doing the same today. A few years ago, powerful and free open-source models already existed, but getting them to run smoothly was fraught with obstacles. The capability was there, yet it felt locked behind a door, making it difficult for developers to use them as easily as calling a proprietary model's API. Ollama's solution is an application you can download to your computer, starting the latest open-source model with a single command and then building on it through a simple API, making running open-source models as easy as running any ordinary software—without needing permissions, API keys, or expensive server hardware. Your model, your machine, your data.
Its product is built around three principles: ownership, affordability, and privacy. Ownership means open-source models belong to the user, who can keep, customize, and optimize them anytime without being locked into a specific model that their agent or application depends on. Affordability comes from running models on your own hardware, eliminating runaway billing based on tokens, so experiments, iterations, and releases no longer need to worry about each prompt adding to the cost. Privacy ensures data never leaves your local machine, and even if it does go to the cloud for expansion, the same level of trust can be carried along.
Open-source models are no longer just toys in laboratories. Ollama Cloud has become the most convenient one-stop access to the most powerful open-source models, including GLM, Nemotron, DeepSeek, Kimi, MiniMax, and more. Its token usage grows by more than 100% monthly on average. What began as the joy of running a model for the first time on a personal computer has now expanded to tackle the tough challenges once reserved for closed-source models.
This $88 million has been positioned by Ollama as fuel to push the ecosystem forward. It is currently at the core of the open-source model ecosystem, and the funds will be invested in three areas: seamless hybrid reasoning, support for newly released open-source models on the day they are launched, and a cloud service that allows any developer and their team to use the most powerful models without sacrificing ownership or privacy. Jeff and Michael wrote in the conclusion that they had stood at the beginning of similar changes before, betting that openness and ease of use would ultimately prevail. Now, they are putting everything on the line again. With 85% of Fortune 500 companies quietly moving open-source models into their internal systems, the next carriage on this open-source train clearly still has plenty of seats available.
