River AI secured $1.1 billion to enable companies to develop and maintain their own models.

      River AI has secured $1.1 billion in funding, roughly four months after its establishment. The funding round was led by General Catalyst and AMP PBC, with NVIDIA and AMD Ventures also making strategic investments, alongside contributions from Y Combinator and Temasek. Founded by Igor Babuschkin, previously a co-founder of xAI, the company is based in Palo Alto.

      The core proposition is that businesses should train and own their models instead of relying on general-purpose alternatives. River provides an API for LoRA fine-tuning and reinforcement learning on open weight models, handling the underlying infrastructure such as weight transfers, sampling-training consistency, and elastic computing.

      The claims about performance are self-reported and have not been validated by external testing. River asserts that a complex reinforcement learning process can be completed in 15 to 20 minutes through its API without needing an infrastructure team, delivering savings of two to four times compared to closed-source options. Billing is based on tokens used for training and inference, which the firm argues eliminates costs associated with idle GPU capacity.

      “The current method of AI development is not indicative of its future,” stated Babuschkin, who serves as River's CEO. “AI should be open, accessible, and affordable. It should act in the best interest of its user, not the lab that trained it.”

      River's ambitions extend beyond software. The company is also developing hardware designed to place personal AI closer to users, along with the necessary training infrastructure and products aimed at personalization and continuous learning.

      General Catalyst highlighted the national implications of the investment. “For the United States to maintain its leadership in AI, it is crucial to lead in open weight models while also keeping a strong position in closed frontier models,” said CEO Hemant Taneja, positioning River’s mission as vital for American resilience. This perspective has garnered additional support, with Nvidia among those endorsing an open letter advocating for American leadership in open weight models, a letter not signed by OpenAI or Anthropic.

      Babuschkin previously worked on generative modeling and reinforcement learning at Google DeepMind, later leading large-scale training efforts at OpenAI, before co-founding xAI. He departed Elon Musk's company in August 2025 to initially pursue setting up a venture firm. All 11 co-founders of xAI have since left.

      The fundraising exceeded expectations. In May, Forbes reported that Babuschkin was targeting up to $1 billion, valuing the company at as much as $5 billion, and planned to invest up to $100 million of his personal funds.

      River AI was officially incorporated in Nevada on April 20. No valuation details were shared on Tuesday.