Sequoia and Nvidia support David Silver's Ineffable Intelligence with a valuation of $5.1 billion.

      David Silver departed Google DeepMind in late 2025 after over a decade of developing AlphaGo, AlphaZero, AlphaStar, and contributing to Gemini. He founded Ineffable Intelligence in November 2025, a company that currently has no product, revenue, or public roadmap. What it does possess is a fundamental thesis and a founder whose proven track record holds a billion-dollar value to investors based solely on their confidence.

      Ineffable Intelligence, the AI startup based in London established by Silver, who led the creation of AlphaGo, AlphaZero, and AlphaStar at Google DeepMind, has garnered backing from Sequoia Capital and Nvidia, receiving a valuation of $5.1 billion. This funding round, one of the largest raised by a startup at such a nascent stage, was primarily led by Sequoia, with managing partner Alfred Lin and partner Sonya Huang personally traveling to London to meet Silver and finalize the deal. Nvidia’s venture arm invested at least $250 million, valuing a company incorporated in November 2025 with no product, revenue, or public roadmap at over five billion dollars.

      During his time at Google DeepMind, Silver made significant contributions that capture the modern narrative of AI. AlphaGo, launched in 2016, was the first AI program to defeat a professional Go player without a handicap, later triumphing 4-1 against 18-time world champion Lee Sedol in a widely viewed match across Asia. The importance of this moment has now become a part of AI culture: it inspired Marc Andreessen’s "Sputnik moment," initiated a surge in deep learning investments, and set the stage for Demis Hassabis to eventually earn a Nobel Prize in Chemistry for AlphaFold.

      Following AlphaGo, Silver developed AlphaZero, which mastered Go, Chess, and Shogi purely through self-play, without utilizing any human data, demonstrating that a single reinforcement learning system could achieve superhuman performance across multiple complex games simultaneously. AlphaStar was the next step, showcasing grandmaster-level prowess in StarCraft II against professional players.

      The thesis of Ineffable Intelligence directly contests the current paradigm of AI development. Silver, in a 2025 paper co-authored with Richard Sutton, a leading researcher in reinforcement learning, argues that large language models have inherent limitations as they are trained exclusively on human-generated data. This means they can reorganize, enhance, and remix existing human knowledge, but they lack the ability to uncover genuinely novel insights. In contrast, reinforcement learning enables an AI to learn through interaction with its environment, engaging in trial and error and self-play, thus generating strategies and insights that have never been conceived by humans. For instance, AlphaGo's notable Move 37 in Game 2 against Lee Sedol was not part of any human game record; it was a discovery made by the AI functioning beyond human intuition. Silver believes that scaling this approach could lead to the development of superintelligence.

      The context of valuation is noteworthy. Ilya Sutskever, former Chief Scientist at OpenAI who founded Safe Superintelligence after leaving in 2024, raised $3 billion with a valuation that escalated to $32 billion by April 2025, all for a company without any product. Mira Murati, the former Chief Technology Officer at OpenAI, established Thinking Machines Lab and secured a multibillion-dollar cloud infrastructure deal with Google. Meanwhile, Meta’s former Chief AI Scientist Yann LeCun is raising around €500 million for AMI Labs. This trend suggests that the AI investment landscape in 2025-26 values not current capabilities but rather the credibility of the researcher, the feasibility of the thesis, and the team’s track record as indicators of the likelihood of future breakthroughs. By this measure, Silver, who has developed three of the most renowned AI systems in history, commands a premium.

      Ineffable Intelligence was incorporated in November 2025, with Silver appointed as director in January 2026. The startup is located in London, a strategic choice by Silver. The UK hosts Google DeepMind’s headquarters and has a robust academic pipeline from institutions like UCL and Oxford, along with a growing cluster of frontier AI researchers who left major laboratories. Silver himself remains a professor at University College London. The $5.1 billion valuation positions Ineffable Intelligence as one of the most valuable pre-product AI startups in Europe. The funding round was reported in February 2026 to be around $4 billion pre-money; and the Bloomberg valuation of $5.1 billion likely reflects the post-money value after the capital influx, or an updated final closing figure. February reports indicated that Nvidia, Google, and Microsoft were considering participation, while the confirmed deal consists of Sequoia and Nvidia as committed investors.

      Critics of Silver’s thesis are vocal. Reinforcement learning has shown remarkable results in constrained domains with clear victory conditions, such as Go, Chess, and StarCraft, but it has historically faced challenges in open