David Silver, the former lead of Google DeepMind’s reinforcement learning team, has raised $1.1 billion in seed funding for his new AI venture, Ineffable Intelligence. The British startup, founded just months ago, is valued at $5.1 billion and aims to build a so-called superlearner that discovers knowledge entirely through trial and error — without relying on human-generated data.
A bet on pure reinforcement learning
Silver is best known for leading the development of AlphaZero, the DeepMind program that mastered chess and the board game Go by playing against itself, learning purely from experience. Ineffable Intelligence intends to apply the same principle at a much larger scale, aiming to create an AI system that can learn any skill or knowledge domain without human examples or labeled datasets.
Also read: Medicare’s quiet bet on AI: A new payment model that most of tech hasn’t noticed
The company’s website describes the ambition in stark terms: if successful, this would represent a scientific breakthrough comparable to Darwin’s theory of evolution — a law that explains and builds all intelligence. The venture is Silver’s self-described life’s work, and he has stated publicly that any personal profits will go to high-impact charities.
The investors and the London AI ecosystem
The funding round was led by Sequoia Capital and Lightspeed Venture Partners, with participation from Index Ventures, Google, Nvidia, the British Business Bank, and Sovereign AI — the UK government’s recently launched venture fund for artificial intelligence. The size of the round, unusual for a seed-stage company, reflects the confidence investors have in Silver’s track record and the potential of reinforcement learning to break through the limitations of current large language models.
Also read: Altman testifies Musk once proposed handing OpenAI to his children during safety dispute
Ineffable Intelligence joins a growing list of AI startups founded by star researchers that have raised massive early rounds — sometimes called coconut rounds. Last month, AMI Labs, co-founded by Turing Award winner Yann LeCun, raised $1.03 billion at a $3.5 billion valuation. Recursive Superintelligence, founded by former DeepMind principal scientist Tim Rocktäschel, reportedly raised $500 million with demand that could push it to $1 billion.
These developments signal mounting momentum around London as a global AI hub. DeepMind’s continued presence after its 2014 acquisition by Google has created a powerful alumni network. Several former DeepMind researchers are expected to join Ineffable’s executive team. Additionally, Jeff Bezos’ AI lab, Project Prometheus, is reportedly in talks to secure office space near Google’s London AI hub.
What this means for the AI industry
The emergence of Ineffable Intelligence and similar ventures represents a significant shift in AI research funding. Investors are placing large bets on approaches that diverge from the dominant large language model standard, betting that reinforcement learning — which has proven effective in games and robotics — can scale to general intelligence. If successful, this could lead to AI systems that are less dependent on the vast amounts of human-generated text and data that power today’s models, potentially reducing bias and expanding capabilities in domains where human data is scarce.
However, the path from AlphaZero’s success in board games to a general-purpose superlearner is not straightforward. Reinforcement learning remains computationally expensive and difficult to apply to open-ended problems. Ineffable Intelligence has not disclosed a timeline for its research or any specific product plans, and it remains unclear how or when the venture will generate revenue.
Conclusion
David Silver’s $1.1 billion raise for Ineffable Intelligence is a major vote of confidence in reinforcement learning as a path to more capable and autonomous AI systems. The funding also underscores London’s growing importance as a center for leading AI research, fueled by DeepMind’s alumni network and government-backed sovereign investment funds. While the technical and commercial challenges are substantial, the scale of investment suggests that investors see real potential in an approach that could fundamentally change how AI systems are built.
FAQs
What is Ineffable Intelligence building?
Ineffable Intelligence is developing a superlearner AI system that uses reinforcement learning to discover knowledge and skills through trial and error, without relying on human-generated data or examples.
Who is David Silver?
David Silver is a professor at University College London and former lead of the reinforcement learning team at Google DeepMind. He was instrumental in developing AlphaZero, which mastered chess and Go through self-play.
How does this differ from large language models like GPT?
Large language models learn from vast datasets of human-written text. Ineffable’s approach uses reinforcement learning, where the AI learns from its own actions and outcomes, potentially enabling it to discover knowledge that humans haven’t explicitly documented.
Be the first to comment