Analysis
Ollama, the open-source tool that lets developers run AI models on their own hardware, closed a $65 million Series B led by Theory Ventures, with Benchmark, 8VC, Y Combinator, Pace Capital, 49 Palms and GTMFund also participating. The round brings Ollama's total funding to $88 million since the company launched in 2023, according to founder and CEO Jeff Morgan.
The growth numbers are the real story: Ollama says usage has doubled since January to nearly 8.9 million monthly developers, with the platform now used inside 85% of the Fortune 500 and adding close to a million new installs a week. That's an unusually steep adoption curve for developer infrastructure, and it reflects a broader shift among engineering teams toward running open-weight models locally rather than depending entirely on hosted frontier-lab APIs for every workload.
Ollama's core pitch is simplicity: developers can get an open-weight model running locally with a single command, and the company layers on the option to seamlessly scale workloads to Ollama's own cloud infrastructure without changing APIs or rewriting code -- a hybrid model that lets teams start cheap and local, then scale to cloud compute only when a workload actually needs it. That flexibility is increasingly valuable as enterprises balance the cost unpredictability of hosted frontier APIs against the latency, privacy and control benefits of running models on infrastructure they own.
The company sits at the center of the open-weight model ecosystem alongside Hugging Face, which focuses more on model hosting and dataset distribution, and cloud-native inference providers like Together AI and Fireworks AI, which compete more directly on hosted-inference pricing and scale. Ollama's differentiation is specifically the local-first developer experience -- the tool developers reach for first when experimenting with a new open model before deciding whether a workload needs to move to production-scale cloud infrastructure.
For enterprise buyers, Ollama's 85% Fortune 500 penetration is a meaningful data point: even companies with large frontier-API budgets are running meaningful workloads locally, whether for cost control, data sensitivity, or simply developer experimentation that never needed cloud API costs in the first place. For founders building developer tools, Ollama's growth is proof that the open-weight model ecosystem has real developer demand independent of which specific open models happen to be trending at any given time.
The bear case is that Ollama's growth is partly a function of the broader open-weight model boom -- Llama, Qwen, DeepSeek and others releasing increasingly capable models developers want to try locally -- and a slowdown in open-weight model releases or capability could soften Ollama's own growth curve even if its product execution stays strong. What to watch next: how Ollama's cloud-compute revenue mix evolves relative to its free local-tool usage, and whether the company faces more direct competition from Hugging Face or a frontier lab building its own local-run tooling.