River AI, a Palo Alto start-up founded by xAI co-founder Igor Babuschkin, said on 11 August 2026 that it had raised $1.1bn across a seed round and a Series A. General Catalyst and AMP PBC led the funding.
NVIDIA and AMD Ventures came in as strategic investors, and Y Combinator and Temasek also took part. River did not say how the $1.1bn was divided between the two rounds, and it did not disclose a valuation.
TechCrunch described River as a two-month-old company; it came out of stealth in June 2026. According to The Next Web, River was incorporated in Nevada on 20 April, and the funding came in above its target.
“The way AI is built today is not how it will be built in the future. AI should be open, freely available, and affordable. It should feel like it is working for the person using it, not the lab that trained it. We started River to allow people and companies to own their intelligence.”Igor Babuschkin, Co-Founder & CEO, River AI
What River AI does
River's first product is the River API, which lets developers and companies run LoRA fine-tuning and reinforcement learning on open-weight models without building their own training infrastructure. The platform takes care of weight transfers, consistency between sampling and training, and elastic compute. Models trained on it can be deployed to production straight away.
Customers pay for the tokens used in training and inference rather than for reserved GPU capacity. TechCrunch reported that prices are set per million tokens and vary with the open model used. River said a company can finish a complex reinforcement learning run in 15 to 20 minutes and spend two to four times less than with closed-source alternatives; The Next Web noted that these claims had not been independently tested.
The company has said its longer-term aim is personal AI that learns from and acts for an individual user. It plans products built around personalisation and continual learning, as well as hardware that would let such models run close to the user instead of in a remote data centre.
Who is behind it
Babuschkin is River's co-founder and chief executive. Before co-founding xAI he worked on generative modelling and reinforcement learning at Google DeepMind and led large-scale training work at OpenAI. The Next Web reported that he left xAI in August 2025, at first to set up a venture firm. River said its founding team also includes people with experience at xAI and Tesla.
The Next Web also reported that Forbes said in May that Babuschkin was seeking up to $1bn at a valuation of as much as $5bn, and that he planned to put up to $100m of his own money into the company. River has not confirmed a valuation.
Who invested
General Catalyst framed the deal around open-weight models. Its chief executive, Hemant Taneja, said US leadership in AI required leadership in open-weight models as well as closed frontier models, and called River's plans a priority for American resilience. Marc Bhargava, a managing director at the firm, said companies had lacked an affordable way to train and own custom models.
AMP PBC is an AI-focused investment firm set up in 2026 by former Andreessen Horowitz general partner Anjney Midha, TechCrunch reported. Its portfolio includes Black Forest Labs, Mistral AI, LMArena and OpenRouter.
What the money is for
River said the funding would go into every layer of the stack it is building: training infrastructure that makes fine-tuning accessible to developers, products for personalisation and continual learning, and new hardware for personal AI.
TechCrunch called the size of the round unusual for such a young company and suggested it could be a sign of an overheated AI market. It linked the bet to rising demand from enterprises that want more control over their AI models, including by mixing in open-weight ones.
Company website: river.ai