AlphaGo Creator David Silver Raises $1.1B Seed to Build AI That Learns Without Human Data

On Monday, a new contender entered the crowded field of frontier AI labs — and it arrived with a bang. Ineffable Intelligence, a London-based startup founded by David Silver, the former head of reinforcement learning at DeepMind and the architect of AlphaGo and AlphaZero, has closed a $1.1 billion seed round at a $5.1 billion valuation. The round, co-led by Sequoia Capital and Lightspeed Venture Partners, is the largest seed funding round in European startup history by a wide margin.

For a company that launched only a few months ago, the figure is staggering. It also signals something broader: investors believe the next frontier in AI may not be built by scaling up human-labeled datasets, but by teaching machines to learn the way humans do — through direct experience with the world.

The Man Behind AlphaGo

David Silver’s credentials are almost without parallel in applied AI research. As a faculty member at University College London and a longtime lead researcher at DeepMind, he was the principal architect behind some of the most celebrated AI milestones of the past decade.

AlphaGo, which in 2016 defeated world Go champion Lee Sedol — a feat widely considered a decade ahead of schedule — used a combination of supervised learning on human games and reinforcement learning to master the ancient board game. But its successor, AlphaZero, was more radical: it played only against itself from scratch, with no human game data at all. Starting with only the rules of chess, shogi, and Go, AlphaZero surpassed all previous AI systems and all human champions within hours of self-play.

AlphaZero became a touchstone moment in AI — proof that a system could achieve superhuman capability in complex domains purely through self-generated experience. Silver has spent years since asking a deeper question: what would it look like to generalize that approach across all of human knowledge, not just board games?

Ineffable Intelligence is his answer.

What Is a “Superlearner”?

The company’s central thesis is that the current paradigm of AI training — massive datasets of human-generated text, code, and images, curated, labeled, and filtered at enormous cost — is a bottleneck that will eventually constrain the field.

Ineffable’s approach, which Silver refers to as building a “superlearner,” aims to create an AI that can acquire broad knowledge and skills primarily through experience and self-directed exploration, rather than passive absorption of human data. The technical approach draws heavily on reinforcement learning, the branch of AI where agents learn by taking actions, receiving feedback, and adjusting their behavior accordingly.

The ambition is not to replace large language models, but to give AI systems a fundamentally more active learning substrate. Where today’s models are trained once on fixed datasets and then deployed, a superlearner would continuously refine its understanding by interacting with its environment — whether that environment is code execution, scientific simulation, or structured problem-solving.

This connects directly to one of the most pressing scaling challenges in the industry: leading AI labs have begun to report diminishing returns from simply adding more human-generated training data. If the next wave of capability gains requires moving beyond human data, Silver’s timing — and backers — suggest many in the industry share that hypothesis.

The $5.1 Billion Bet

The fundraise is remarkable not just for its size but for the diversity and prestige of its participants. Sequoia Capital and Lightspeed Venture Partners co-led the round, bringing two of Silicon Valley’s most powerful VC firms into a European AI startup at an early stage. Additional investors include Nvidia — whose GPU dominance makes it a strategic backer interested in the next wave of compute demand — Google, DST Global, Index Ventures, and the UK government’s Sovereign AI Fund, a relatively new vehicle created to ensure Britain maintains strategic AI capability.

The UK government’s involvement is especially notable. In an era of intensifying AI nationalism, having a sovereign fund participate in the largest seed round in European history is a clear statement of national intent. The UK has been aggressively courting AI talent and capital since its own domestic AI strategy was relaunched following DeepMind’s continued commercial success.

Silver himself has made an unusual commitment: he has pledged, via Founders Pledge, to donate 100% of any personal financial gains from his Ineffable equity to charity. The pledge — a relatively rare commitment at this scale — underscores Silver’s stated belief that the work is primarily a scientific mission rather than a wealth-creation exercise.

Europe’s AI Moment

The funding round lands at a significant moment for European AI. For years, the narrative has been one of brain drain: elite European researchers, many trained at institutions like DeepMind, Oxford, Cambridge, and ETH Zurich, have departed for better-funded roles at OpenAI, Anthropic, Google DeepMind, and Meta in the United States.

At $1.1 billion, Ineffable’s seed round — the largest in European startup history — is a counterargument. It demonstrates that world-class AI talent can raise world-class capital while staying in London, and that the UK ecosystem, augmented by the Sovereign AI Fund, can compete for frontier AI investment.

Sequoia’s participation is significant in this regard. The firm has become increasingly active in European AI after years of primarily backing US startups, and its co-lead here suggests it is prepared to invest at scale in European founders when the opportunity warrants it.

The round also arrives as the European Union’s AI Act begins to shape the regulatory landscape. Whether Ineffable’s research will be classified under the Act’s highest-risk categories — and what that means for deployment — will be a key question as the company moves toward building and releasing models.

The Quiet Competition With OpenAI and Anthropic

Ineffable Intelligence enters a crowded field. OpenAI, Anthropic, Google DeepMind, xAI, and Meta’s AI division are all pursuing frontier AI with resources that dwarf even a $1.1 billion seed. But Silver’s framing is deliberately orthogonal to the current race.

Most of today’s leading labs are competing on the same axis: better reasoning, longer context windows, faster inference, cheaper API access. Ineffable is positioning itself on a different axis entirely — one defined not by how much human knowledge a model has absorbed, but by how effectively it can learn new things without human input.

Whether that bet pays off depends on unresolved scientific questions that Silver has arguably spent more of his career on than almost anyone alive. That, ultimately, is what justified the $5.1 billion pre-money valuation on day one.

The company says it will publish its research openly, a nod to the academic traditions Silver comes from. What it builds — and when — will be closely watched by an industry increasingly asking whether the next major AI breakthrough will look more like AlphaZero than GPT.