xAI Co-Founder Raises $1.1 Billion for River AI to Build Personal Models That Actually Belong to Users

River AI appeared out of nowhere. In just two months the startup secured $1.1 billion. General Catalyst led the round. AMP PBC joined as co-lead. Nvidia and AMD Ventures added strategic bets. Y Combinator and Temasek came along too. The sum stunned even hardened AI watchers. A company incorporated in April had suddenly amassed one of the largest early-stage hauls ever.

Igor Babuschkin founded River. He co-founded xAI with Elon Musk. He left that venture in August 2025. Before xAI he shaped large-scale training at OpenAI. His earlier work at Google DeepMind focused on generative modeling and reinforcement learning. Those credentials lent instant credibility. Yet the speed and size still raised eyebrows. Forbes had reported in May that Babuschkin sought up to $1 billion at a potential $5 billion valuation. He reportedly planned to commit up to $100 million himself. The final round exceeded those targets. No valuation was disclosed Tuesday.

The pitch cuts against the dominant narrative. Most AI labs chase ever-larger general models meant to replace human workers. River wants the opposite. It aims to create agents that feel like they work for the individual user. Not for the lab that trained them. Babuschkin put it plainly. “The way AI is built today is not how it will be built in the future,” he told The Next Web. “AI should be open, freely available, and affordable. It should feel like it is working for the person using it, not the lab that trained it.”

So the company rebuilds the stack. Training. Models. Product layer. Even new hardware. The goal is personal AI that lives close to the user. Agents that know their owner intimately. That learn and adapt continuously. Less like today’s task-oriented assistants. More like guardian angels. Quiet. Present. On your side. Babuschkin described the vision in his launch post. “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.”

Enterprises already sense the shift. They want control over their models. They tire of renting opaque services. They prefer to train on proprietary data and own the result. River offers an immediate answer. Its API lets developers fine-tune open-weight frontier models. The methods include low-rank adaptation and reinforcement learning. No massive infrastructure team required. The company handles weight transfers, sampling consistency, elastic compute. All of it.

Performance claims sound aggressive. A complex reinforcement learning run finishes in 15 to 20 minutes. Costs run two to four times lower than closed-source alternatives. Billing runs on tokens used for training and inference. No charges for idle GPUs. “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,” the company stated in its announcement, as reported by TechCrunch. Prompt engineering becomes obsolete. “Prompting steers a model you don’t own and can’t improve. River lets you train open models into ones that are truly yours — and serve them like any other endpoint.”

But. The claims remain unverified by independent tests. River itself acknowledges that. Skeptics question whether such speed and savings hold at scale. Still, the timing feels right. Open-weight models gain ground. Nvidia signed an open letter supporting American leadership in open models. OpenAI and Anthropic notably stayed away. Hemant Taneja, General Catalyst’s chief executive, framed the bet in stark terms. American leadership in AI demands strength in open-weight systems alongside closed frontier models. He called River’s agenda “a priority for American resilience.”

The investment firm AMP PBC also led. Former Andreessen Horowitz partner Anjney Midha launched it this year. AMP had backed Black Forest Labs, Mistral AI, LMSYS Arena and OpenRouter. Its participation signals alignment with the open-source side of the house. Strategic checks from Nvidia and AMD Ventures add hardware credibility. Those chip giants clearly see value in tools that drive demand for their silicon in enterprise and personal settings.

River came out of stealth in June. It had incorporated in Nevada on April 20. The entire journey from legal formation to billion-dollar raise took roughly four months. That pace reflects the current frenzy. Capital floods toward any credible team that promises a different path. Babuschkin is not alone in seeing limits to today’s proprietary model builders. As he noted in one discussion, proprietary model makers “are kind of starting to get squeezed.”

Yet questions linger. Can one company truly rebuild every layer? Training infrastructure is expensive. Hardware for personal AI remains nascent. Local agents already emerge in projects like OpenClaw. Nvidia partners with Dell, Microsoft and HP on AI-capable PCs. River must differentiate its full-stack approach. The war chest gives it runway. Whether it translates into defensible technology is the test ahead.

Recent coverage reinforces the momentum. The Economic Times highlighted the focus on custom tools for enterprises building personalized models on their own data. General Catalyst published its own perspective on owning the AI experience, quoting Babuschkin’s physics roots and desire to build tools that elevate rather than replace people. X posts lit up Tuesday with reactions ranging from astonishment at the valuation implied to cautious optimism about shifting power back to users and companies.

The broader industry watches closely. If River delivers, enterprises gain genuine ownership. Individuals might one day train their own guardian agents on private data. The alternative is continued dependence on a handful of labs. Babuschkin believes the future tilts toward openness and personalization. His backers just placed an enormous wager on that conviction. The coming months will reveal whether the vision matches the hype. For now the cash is real. The ambition is larger.