The AI startup River AI has raised $1.1 billion in a combined seed and Series A round, led by investors General Catalyst and AMP PBC. Chip manufacturers Nvidia and AMD Ventures are also strategically participating in the company, which was founded in June 2026 by xAI co-founder Igor Babuschkin. The fresh capital is meant to accelerate the build-out of an open AI infrastructure owned and controlled by users.
River API trains open models in minutes
The startup’s first product is the River API, as River AI announced on its own blog. Through the interface, companies adapt open AI models using reinforcement learning and LoRA fine-tuning – a complete training run takes 15 to 20 minutes according to the company and requires no dedicated infrastructure team. River AI puts the cost savings compared with closed providers like OpenAI or Anthropic at two to four times; this figure has not been independently verified. Billing is usage-based per processed million tokens, tiered by model size. Alongside lead investors General Catalyst and AMP PBC, early-stage investor Y Combinator and Singapore’s sovereign wealth fund Temasek are also participating in the round, according to River AI.
As TechCrunch reports, River AI is only two months old: Babuschkin took the company out of stealth mode in June 2026. Beyond the training infrastructure, the startup also plans hardware for the local operation of personal AI systems, according to its own statements – a rollout date has not yet been set. River AI does not disclose a company valuation for the funding round.
Babuschkin wants to hand AI ownership back to users
Babuschkin previously worked at Google DeepMind and OpenAI on generative models and reinforcement learning before becoming a co-founder of xAI. As Bloomberg reported at the June launch, Babuschkin left xAI to build a personal AI. Users are meant to train and modify it themselves, rather than follow the rules set by large providers. This positioning deliberately sets River AI apart from closed systems like ChatGPT or Claude.
The startup joins a growing number of well-funded bets on open AI infrastructure. Just in July, developer service Ollama raised $65 million for open AI models. Inference providers like Fireworks AI are now valued in the double-digit billions. That chip manufacturers Nvidia and AMD Ventures are now backing River AI too underscores their interest in open training infrastructure outside the established cloud providers.
Whether a two-month-old company with no published valuation and no broad user base can justify its investors’ billion-dollar expectations remains to be seen. What will matter is whether River AI actually delivers the hardware it has announced for running personal AI models locally – a date for that has not yet been named.
