Popular open source AI developer tool Ollama raises $65M, grows to nearly 9M users

Ollama raised a $65M Series B led by Theory Venture (following a $15M Series A), bringing total funding to $88M. The open-source AI tool is used by over 8.9M developers monthly and runs open-weight models locally, while its neocloud subscriptions charge from free up to $100/month. Investors view the shift toward lower-cost open models as accelerating enterprise adoption, supporting Ollama’s cloud business despite limited revenue/valuation disclosure.

Analysis

The market implication is less about this company and more about who captures the next dollar of AI spend. Open-weight tooling reaching mainstream enterprise workflows shifts bargaining power away from frontier model APIs and toward the infrastructure layers that make model routing, serving, and GPU allocation frictionless. That is structurally positive for accelerator and server demand, but it can be negative for the gross-margin narrative of closed-model vendors if customers increasingly reserve premium APIs for edge cases.

Second-order, this looks like a mix shift inside AI rather than a simple demand expansion. Routine inference moving to cheaper open models can compress per-task monetization for hyperscalers and model vendors, while still leaving total compute demand intact or even higher if usage broadens. That favors names tied to throughput and utilization over pricing power: NVDA, SMCI, and select neoclouds like CRWV/NBIS are better positioned than software businesses whose AI attach rates depend on expensive proprietary calls.

The contrarian miss is that “open vs closed” is probably not a winner-take-all trade; it is a routing layer problem. If developer defaults get standardized early, the value accrues to whoever sits in the workflow, not to the model itself. Over the next 1-3 months, the key falsifier is not user growth but whether hyperscaler commentary shows inference ARPU and AI gross margin holding up despite open-model adoption; over 6-18 months, a durable decline in closed-model pricing would be the real bearish signal for AI software multiples.

Key Decisions for Investors

• No standalone trade in the private company; treat this as a sector signal, not a catalyst with immediate P&L relevance.
• 1-3 month relative-value idea: long SMH / short XSW or IGV on the thesis that open-weight adoption supports GPU throughput while pressuring AI software pricing power; target 1.5-2.0x downside capture if enterprise routing to open models accelerates.
• Tactical long on weakness in NVDA or CRWV/NBIS if the market interprets open models as bearish compute; the cleaner read is that cheaper models expand total inference volume. Falsify if cloud AI capex or utilization commentary rolls over in the next two earnings cycles.
• Do not short MSFT or GOOGL on this headline alone; wait for evidence in cloud AI margin disclosure or assistant monetization that low-cost model routing is actually displacing premium API usage.
• Set a watch item on enterprise AI spend surveys and hyperscaler commentary over the next 1-2 quarters; if closed-model pricing re-accelerates or compliance requirements force hosted inference, unwind any bearish software-multiple view.