xAI Co-Founder's $1.1B Anti-Replacement AI
The man who helped build xAI walked away to rebuild AI from scratch, not to replace workers, but to give everyone a personal AI "guardian angel." And investors just threw $1.1 billion at the idea before the company was three months old.
Igor Babuschkin, the co-founder of xAI who previously held AI roles at DeepMind and OpenAI, has raised $1.1 billion in a seed/Series A round for his new startup, River AI. The round was led by General Catalyst and AMP PBC, with participation from Nvidia, AMD Ventures, Y Combinator, and Temasek. AMP PBC was founded in 2026 by former Andreessen Horowitz general partner Anjney Midha, who backed Black Forest Labs, Mistral AI, LMArena, and OpenRouter during his time at a16z.
The Anti-Replacement AI
The most striking thing about River AI is not the size of the round, though $1.1 billion for a company that only emerged from stealth in June is certainly eye-popping. It is the explicit rejection of the direction most AI labs are heading.
"To get there, we believe the stack has to be rebuilt end to end: training, models, the product layer, and new hardware that lets personal AI live close to you," Babuschkin wrote in his launch blog. "Capable agents will be a normal part of everyday life. Less like the assistants you call on today when you need a task done, more like guardian angels: quietly present, on your side, helping with what actually matters to you. They will know you well, and they will be yours, not someone else's."
This "guardian angel" framing is a direct counterpoint to the dominant narrative in AI today. While OpenAI, Anthropic, and xAI chase ever-larger frontier models designed to automate knowledge work, Babuschkin is betting that the real market is in personal AI that individuals own and control, not AI that replaces them.
What River AI Actually Builds
River AI already has a product. The company offers an API that lets developers use reinforcement learning and LoRA (low-rank adaptation) fine-tuning on open models. Pricing is per 1 million tokens, with rates varying by model.
"Prompting steers a model you don't own and can't improve," the company's product literature says. "River lets you train open models into ones that are truly yours, and serve them like any other endpoint."
The company claims any enterprise can complete a complex reinforcement learning run in 15 to 20 minutes with no infrastructure team required, at two to four times the cost savings relative to closed-source alternatives like GPT-4o or Claude. This is a bold claim in a market where fine-tuning has historically required dedicated ML engineering teams and expensive GPU clusters.
The $1.1 Billion Question
The size of the round raises an obvious question: does a 2-month-old startup with a beta API really need $1.1 billion? In the overheated AI funding environment, the answer seems to be that investors are betting on the founder, not the product.
Babuschkin's pedigree is exceptional. He was a co-founder of xAI, deeply involved in the development of Grok. Before that, he worked on AI research at DeepMind under Demis Hassabis and at OpenAI under Ilya Sutskever. He has seen the frontier from three different angles, and his thesis is that all three are wrong about the endgame.
The broader context is on his side. Enterprises are increasingly waking up to the risks of locking themselves into a single AI vendor. The rise of open-weight models and the growing interest in local, personal AI agents, visible in the explosion of projects like OpenClaw and its derivatives, suggests that Babuschkin is betting on a real trend, not a fictional one.
Nvidia, one of the investors in this round, is simultaneously partnering with PC makers like Dell, Microsoft, and HP to build AI-capable hardware for exactly this kind of local AI future. The pieces are falling into place, even if River AI's exact role in that future remains unproven.
The Sceptical Take
For all the vision, River AI is a bet on a future that does not yet exist. Personal AI agents that individuals actually own and control would require a fundamentally different computing infrastructure than what exists today. Babuschkin acknowledges this, the "new hardware that lets personal AI live close to you" is part of the stack he wants to rebuild, but he is essentially promising to rebuild the entire foundation of AI while flying a plane he is still assembling.
The $1.1 billion gives him the runway to try. Whether the market actually wants personalized AI that is "yours, not someone else's", or whether most users are perfectly happy with the centralised, corporate-owned models they already use for free, is a question that $1.1 billion cannot answer by itself.
Sources
- TechCrunch: General Catalyst leads $1.1B round into 2-month-old River AI
- Reuters: XAI co-founder's startup River AI raises $1.1 billion
- The New York Times: His Start-Up's Goal: A.I. That Is Trainable and Not Controlled by a Big Company
- River AI (official site)
- Forbes: xAI Cofounder Igor Babuschkin In Talks To Raise Up To $1 Billion For A New AI Startup