River AI, founded by xAI co-founder Igor Babuschkin, announced $1.1 billion in combined seed and Series A funding led by General Catalyst and AMP PBC, with strategic investment from NVIDIA and AMD Ventures, plus Y Combinator and Temasek. The funding accelerates development of River's full-stack platform for training and deploying custom AI models on open-weight architectures, enabling enterprises to train, fine-tune, and own models tailored to their specific workflows without building dedicated infrastructure teams.
The company's API delivers reinforcement learning (RL) and LoRA fine-tuning for frontier open-weight models (Qwen, Kimi, GLM families), allowing any enterprise to complete a complex RL training run in 15–20 minutes at 2–4x lower cost than closed-source alternatives. Trained models deploy instantly to production with token-metered billing (no idle GPU costs), and customers retain full ownership of model checkpoints. This directly addresses the post-training and personalization gap between frontier model capabilities and what enterprises can realistically deploy.
River frames the shift as inevitable: enterprises will migrate from single off-the-shelf frontier models to a mix of private, open-weight models tuned to their data and workflows. The company is building vertically—training infrastructure (live), personalization/continual-learning products (forthcoming), and hardware to run personal AI closer to users (in development). Babuschkin's team brings xAI and Tesla execution; the strategic chip backer support signals confidence in the demand for open-model training infrastructure.