River AI, a startup established just two months ago, has secured $1.1 billion in its first funding round. General Catalyst and AMP PBC led the investment, with participation from NVIDIA, AMD Ventures, Y Combinator, and Temasek. The company's stated goal is to develop personal artificial intelligence agents that users can fully own and continuously retrain.
The startup was co-founded by Igor Babuschkin, who previously co-founded xAI and also worked at DeepMind and OpenAI. Babuschkin is reportedly contributing up to $100 million of his own capital to the venture. River AI's mission is to challenge the prevailing model of centralized AI development, advocating for open, accessible, and affordable artificial intelligence that serves individual users.
River AI's platform offers an API for LoRA fine-tuning and reinforcement learning on open-weight models. The company aims to simplify the process of training, tuning, and deploying custom AI models, removing the need for dedicated infrastructure teams or specialized hardware for enterprises. River claims that its API can complete complex reinforcement learning training runs in 15 to 20 minutes, offering cost savings of two to four times compared to closed-source alternatives. Billing is based on tokens used, which River states eliminates the cost associated with idle GPU capacity.
The company's long-term vision includes developing hardware that allows personal AI to operate in close proximity to the user, alongside a vertically integrated stack encompassing training infrastructure and personalized products. This approach positions River AI as an alternative to the large, closed-model labs that currently dominate the AI landscape.
The substantial early-stage funding for River AI underscores a broader trend of investment in AI infrastructure and the development of models that enterprises can own and customize. Companies like Together AI, Fireworks AI, and Baseten have also recently secured significant funding for platforms focused on specialized AI models and training infrastructure.
While River AI has not disclosed its valuation or the specific allocation of funds between its seed and Series A components, the company was incorporated in Nevada in April 2026. The startup currently has approximately 20 employees and is several months away from releasing its open-source technologies. Investors are betting on Babuschkin's track record in building large AI systems and his argument that businesses will increasingly prefer owning customized models over renting general-purpose ones.
General Catalyst, a lead investor, has been actively investing in AI, including a strategy focused on acquiring labor-intensive service businesses where agentic AI can automate workflows. This approach aims to re-rate the economic profiles of these companies towards software-like margins. NVIDIA and AMD's strategic investments suggest a focus on the hardware and compute infrastructure necessary for advanced AI development.
The success of River AI's model hinges on convincing customers to shift from the convenience of closed, general-purpose AI models to owning and continuously training their own intelligence. The company has yet to release a product, generate revenue, or provide public demonstrations.