Top AI Product

So, David Silver quietly left Google DeepMind late last year, and now we know why. He’s building [Ineffable Intelligence](https://www.ineffable.limited/), a London-based AI lab with a thesis that’s going to make a lot of people uncomfortable: large language models alone will never get us to superintelligence. And he’s got the receipts to back it up.

If the name doesn’t ring a bell, Silver is the brain behind AlphaGo — the system that beat world champion Lee Sedol at Go back in 2016 and basically proved reinforcement learning could tackle problems everyone assumed were decades away. He went on to lead AlphaZero, which taught itself chess, Go, and shogi from scratch, no human data needed. The guy was one of DeepMind’s very first employees when it started in 2010, and he spent over a decade there pushing the boundaries of what RL can do.

Now here’s where it gets wild. As [reported by SiliconANGLE](https://siliconangle.com/2026/02/18/new-ai-startup-ineffable-intelligence-reportedly-raising-1b-funding-round/), [Sifted](https://sifted.eu/articles/david-silver-1bn-ineffable-intelligence), and [Fortune](https://fortune.com/2026/01/30/google-deepmind-ai-researcher-david-silver-leaves-to-found-ai-startup-ineffable-intelligence/), Ineffable Intelligence is raising a **$1 billion seed round** led by Sequoia Capital, valuing the company at roughly $4 billion. If it closes, that would be the largest seed round in European history. And the investor list reads like a who’s who — Nvidia, Google, and Microsoft are all reportedly in talks to participate. Sequoia’s Alfred Lin and Sonya Huang apparently flew to London personally to meet Silver. That’s the kind of founder gravity that doesn’t come around often.

The company’s stated mission is to build “an endlessly learning superintelligence that self-discovers the foundations of all knowledge.” Big words, sure. But Silver’s approach is fundamentally different from what OpenAI, Anthropic, or Google are doing with their flagship models. Instead of training on mountains of internet text, Ineffable Intelligence wants AI that learns by interacting with environments — the same reinforcement learning philosophy that made AlphaGo and AlphaZero work so remarkably well.

There’s no product yet. No revenue. No public roadmap. This is pure conviction investing based on one person’s track record and a compelling bet against the current LLM paradigm. Whether Silver can pull it off is anyone’s guess, but if there’s one person on the planet who’s earned the right to make that bet, it’s probably the guy who already proved RL can achieve what everyone said was impossible.

Keep an eye on this one. The company was [incorporated in the UK in November 2025](https://find-and-update.company-information.service.gov.uk/company/16865241), and things are clearly moving fast.