General Catalyst leads $1.1B round into 2-month-old River AI
TechCrunch Julie Bort ● Covered by 3 sources
River AI just raised $1.1B in seed/Series A led by General Catalyst. It’s a 2-month-old startup trying to make AI agents personal, trainable, and yours.
Based on reporting by TechCrunch, Julie Bort — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
River AI, the startup from xAI co-founder Igor Babuschkin, has landed a huge $1.1 billion seed/Series A round. General Catalyst and AMP PBC led it, with Nvidia, AMD Ventures, Y Combinator, and Temasek also joining in. That is a lot of money for a company that only came out of stealth in June.
Babuschkin’s pitch is blunt: rebuild AI from the ground up. He wants to start with training, not just bolt on better prompts or a shinier chat window. The aim is to produce agents that people can train for themselves, instead of models that drift toward replacing human workers.
River is already selling an API, priced per 1 million tokens and tied to whatever open model a customer uses. It also supports reinforcement learning and LoRA fine-tuning, which is the company’s way of making “prompt engineering” feel old and limited. The pitch is that users should be able to train open models into something they actually own, then run them like any other endpoint.
The timing is doing River a favor. Enterprises are increasingly interested in controlling their model stack and mixing different models, including open weight ones. River is trying to own the messy middle after the model is chosen: the post-training work, the tuning, the practical stuff that turns a generic model into something specific enough to matter.
In its funding announcement, River said an enterprise can finish a complex reinforcement learning run in 15 to 20 minutes without an infrastructure team, and at two to four times the cost savings versus closed-source alternatives. That is the kind of claim that gets investors leaning forward. It also sets up a very large test: whether personal agents become a real product category, or just another grand AI promise with expensive hardware attached.
My take — AI-written commentary, not fact-checked reporting
This is the familiar Silicon Valley move: turn “your own AI” into a platform pitch and then raise a billion dollars to prove it. Open models keep getting treated like raw material, while the real prize is who controls the training layer and the hardware around it. That is less romantic than the agent utopia, but a lot more honest.
Read more about this at: TechCrunch