AlphaGo’s Creator Just Made AI’s Boldest Bet Yet

The man who taught a computer to beat the world at Go just convinced Sequoia, Nvidia and the British government to wager $1.1 billion that the next leap in AI will not come from scraping more human text. Ineffable Intelligence — David Silver’s stealth lab, public for the first time on April 27 — wants to skip the internet entirely and let machines learn the way AlphaZero did: by playing against themselves.

The Bet: $1.1B for a Lab With No Product

Europe’s largest seed round, ever

According to TechCrunch and CNBC, the months-old London startup closed a $1.1 billion seed at a $5.1 billion post-money valuation. That alone would be a record-setting Series B in most years; as a seed, it is unprecedented in Europe and rivals the largest first checks ever written into a frontier-AI lab. Sequoia Capital and Lightspeed Venture Partners co-led the round, joined by Index Ventures, Google, Nvidia, the British Business Bank and Sovereign AI — the UK’s newly-launched sovereign venture vehicle.

No demo, no product, just a thesis

Ineffable has not shipped a model. It has not published benchmarks. What it has is a thesis — and the only researcher in the world with a credible track record of building agents that taught themselves superhuman behaviour from a blank slate. Silver, a UCL professor and DeepMind veteran of more than a decade, was the technical lead on AlphaGo, AlphaZero, AlphaStar and AlphaProof, projects that collectively re-defined what reinforcement learning can do.

The “Era of Experience” Thesis

The internet has been read

Silver’s public writing — most notably the 2024 essay “Welcome to the Era of Experience,” co-authored with Richard Sutton — argues that LLM scaling is bumping into a wall: the supply of high-quality human text is finite, and the marginal token now teaches the model very little. Ineffable’s pitch deck, per Lightspeed’s announcement post, calls this “the end of the imitation era.” The next decade, the firm bets, belongs to agents that generate their own training data through interaction with simulators, tools and the physical world.

“Superlearner,” not chatbot

The product the company describes is not a model that answers questions. It is a system that discovers — in Silver’s framing, an agent capable of “Move 37”-style insights across science, mathematics and engineering, where AlphaGo’s famously unhuman 37th move in its 2016 match against Lee Sedol becomes the template, not the exception. Lightspeed’s partner essay is literally titled “Move 37, Ten Years On.” The audacity of the bet is in the framing: Ineffable is not trying to out-OpenAI OpenAI. It is trying to make the LLM paradigm look like the steam-engine era of AI.

Why It Matters Beyond Silicon Valley

A sovereign-AI signal from the UK

The UK government’s involvement — through both the British Business Bank and the new Sovereign AI fund — is the first marquee deployment of Britain’s sovereign-AI strategy, announced earlier this year. A government press release framed Ineffable as a company “building breakthrough AI that can discover new knowledge,” explicitly positioning London as a frontier-research hub independent of US hyperscalers. That is a sharper geopolitical statement than the cheque size suggests.

A philanthropy commitment without precedent

Silver has pledged 100% of his personal equity gains to high-impact charities through Founders Pledge — the largest individual commitment in the organisation’s history, per its own announcement. At a $5.1B valuation, that is theoretically a multi-billion-dollar pledge. It also reframes the company’s narrative: Ineffable is being marketed not as a unicorn-in-waiting, but as a quasi-public research project with a profit motive bolted on.

Related

• Google’s $40B Anthropic Bet and the TPU Endgame
• Google Said Yes to the Pentagon When Anthropic Said No
• YC’s 2026 RFS: The Harness and the Dual-Agent Workflow
• Google Gemma 4 Goes Open-Source — and Why Apache Matters

Sources

1. TechCrunch — DeepMind’s David Silver just raised $1.1B to build an AI that learns without human data
2. CNBC — Ex-DeepMind David Silver raises $1.1 billion for AI startup Ineffable
3. Lightspeed Venture Partners — Move 37, Ten Years On: Partnering with Ineffable Intelligence
4. GOV.UK — UK backs company building breakthrough AI that can discover new knowledge