River AI, the startup founded by xAI co-founder Igor Babuschkin, has raised $1.1 billion in a seed and Series A round led by General Catalyst and AMP PBC, with support from Nvidia, AMD Ventures, Y Combinator and Temasek.
Launched from stealth in June, River is pursuing an ambitious idea: rebuilding AI infrastructure from the ground up so models can be trained and adapted more directly for individual use. Babuschkin, who previously worked at DeepMind and OpenAI, says the goal is to move beyond generic assistants and toward personal agents that can be shaped by each user.
A new approach to AI training
The company already offers an API that lets developers fine-tune open models using reinforcement learning and LoRA methods. River positions this as an alternative to prompt engineering, giving teams more control over how models evolve and perform in production.
According to the company, its platform is designed to simplify complex training workflows and reduce infrastructure overhead, while supporting open-model strategies that many enterprises are now exploring. River also says its system can help complete advanced reinforcement learning runs in minutes without requiring a dedicated infrastructure team.
The broader vision is clear: AI agents that feel more personal, more adaptable, and more closely aligned with the people who use them. As hardware, model training, and software layers continue to converge, River is betting that the next phase of AI will be defined by ownership, customization, and everyday utility. That shift could help shape a more personal and flexible AI future.
