River AI raises $1.1bn for custom model development platform

River AI enters a market where the gap between AI capability and practical enterprise adoption remains wide. General Catalyst's Marc Bhargava articulated the core problem the company addresses: companies lack cost-efficient pathways to train, tune, and own custom models. Babuschkin, who previously led large-scale training efforts at both OpenAI and Google DeepMind before co-founding xAI, brings deep expertise in the infrastructure and ML systems required to solve this. His vision, stated in the funding announcement, is explicit: "AI should be open, freely available, and affordable" and should "feel like it is working for the person using it, not the lab that trained it." This philosophy underpins River AI's stated aim to place ownership of AI with the people and organizations using it, not with the vendors who built the models.

The funding coalition—anchored by General Catalyst and AMP PBC but including hardware makers Nvidia and AMD Ventures—signals confidence in the infrastructure play. Nvidia and AMD's participation is notable, as it suggests the companies see River AI as a workload driver for their chips rather than a threat; the platform's token-metered billing and elastic compute model may create predictable, high-volume training demand. The presence of Y Combinator and Temasek reflects both the startup ecosystem's backing and international capital's interest in infrastructure-layer AI startups.