Two months. That's all it took for Igor Babuschkin to go from idea to a $5 billion company with $1.1 billion in the bank.
On August 11, 2026, River AI announced a $1.1 billion seed/Series A led by General Catalyst and AMP PBC, with participation from NVIDIA, AMD Ventures, Y Combinator, and Temasek. The company came out of stealth just two months ago. The round is one of the largest ever raised by a company this young.
Babuschkin is not a newcomer. Before founding River, he was a co-founder of xAI — Elon Musk's AI lab — and before that spent years at DeepMind and OpenAI. His thesis with River is pointed: the current AI industry is structurally broken in a way that most people haven't named yet.
The Problem River Is Solving
Right now, if you want a powerful AI model for your company, you rent it from one of a handful of hyperscalers. You don't own the model. You don't control how it's trained, what data it's seen, or how it will behave next quarter when the provider decides to change it. You pay per token, and every token you spend makes someone else's model better — not yours.
River AI's pitch is simple: that should change. According to TechCrunch, River's platform lets any developer train, fine-tune, and serve their own custom models using LoRA fine-tuning and reinforcement learning on top of frontier open-weight models — without needing a dedicated infrastructure team or expensive hardware. Token-metered billing, instant deployment to production, and full model ownership are the core value props.
The company's slogan is "Own Your Intelligence," and it's less a marketing line than a philosophy.
Why This Round Is Different
$1.1 billion into a two-month-old AI company is extraordinary even by 2026 standards. The investors who signed on — NVIDIA and AMD Ventures in particular — signal that this is about more than a product: it's about infrastructure. River is positioning itself as a layer in the open AI stack, not a consumer app or a model vendor.
AMP PBC's involvement is also notable. AMP is one of the larger mission-aligned funds that explicitly backs open-source or public-benefit technology companies, suggesting the "open" part of River's pitch is structural, not just a marketing adjective.
Dealroom's coverage frames River's round as part of a broader capital push into what investors now call "sovereign AI" — the idea that companies, countries, and individuals should be able to run and own their own models rather than depending on a centralized provider. River's emphasis on personal AI and open-weight foundations puts it squarely in that camp.
What Comes Next
River is still in early access. The public-facing API lets developers train custom models on their own data, fine-tune existing open-weight models, and deploy to production instantly. The company hasn't shipped a consumer product, and it hasn't publicly named which open-weight models it supports — though the combination of NVIDIA and AMD backing strongly implies multi-architecture support from day one.
Babuschkin has said he wants to "reinvent AI from scratch," which is a big claim. But so far the company is executing precisely on the part of that vision that matters most to developers right now: making it possible to own what you build, instead of renting someone else's intelligence by the token.
$1.1 billion in two months suggests the market agrees.