River AI, founded by xAI co-founder Igor Babuschkin, raised $1.1 billion to expand enterprise AI tools that let companies train and fine-tune custom models on their own data. General Catalyst and AMP PBC co-led the round, with NVIDIA, AMD Ventures, Y Combinator, and Temasek also participating; no valuation was disclosed.
How much did River AI raise, and what will the money fund?
River AI raised $1.1 billion to expand its enterprise AI toolsetCITE:E1. The startup announced the round on August 11, 2026, saying the capital will go toward helping enterprise customers train and fine-tune their own proprietary models using their own dataCITE:E1.
What is founder Igor Babuschkin's background?
Igor Babuschkin, River AI's founder, previously co-founded xAICITE:E2. Before that, he worked at Google DeepMind researching generative models and reinforcement learning, then led large-scale training efforts at OpenAICITE:E5. River AI describes its mission as moving enterprises away from the general-purpose models offered by large AI labs and toward customizable, open-weight-based models that companies can own outrightCITE:E2.
Which investors led the funding round?
General Catalyst and AMP PBC co-led River AI's $1.1 billion roundCITE:E3. NVIDIA and AMD Ventures joined as strategic investors, alongside Y Combinator and TemasekCITE:E3. River AI did not disclose a valuation for the roundCITE:E3.
What does River AI's API offer, and how does its cost compare to rivals?
River AI says its API completes complex reinforcement-learning training runs in 15 to 20 minutesCITE:E4. The company states this requires no dedicated infrastructure team on the customer's side, and claims its costs run two to four times lower than closed-source competitorsCITE:E4.
| Metric | Figure |
|---|
| Funding raised | $1.1 billionCITE:E1 |
| Reinforcement-learning training time | 15–20 minutesCITE:E4 |
| Claimed cost advantage vs. closed-source rivals | 2–4x lowerCITE:E4 |
What kind of AI stack is River AI building?
River AI is working to build an open AI technology stack alongside its $1.1 billion raiseCITE:E6.
What does this mean?
The round pairs two chipmakers, NVIDIA and AMD Ventures, as strategic investors in a startup whose pitch is training efficiency and lower cost rather than proprietary scaleCITE:E3CITE:E4. That positioning tracks with Babuschkin's own path — through Google DeepMind, OpenAI, and xAI, three labs built around large proprietary models — before founding a company explicitly aimed at giving enterprises open-weight alternatives to that approachCITE:E5CITE:E2.