David Silver’s Ineffable Intelligence raises a $1.1B seed round at a $5.1B valuation to build a reinforcement-learning ‘superlearner’ without human data

Fortune reported on January 30, 2026 that David Silver had left Google DeepMind (he had been on sabbatical) to found Ineffable Intelligence in London. A DeepMind spokesperson said: “Dave’s contributions have been invaluable.” In February the Financial Times reported talks on a seed round of about $1B led by Sequoia. On April 27 the company announced $1.1B at a $5.1B valuation, led by Sequoia and Lightspeed. Index Ventures, which also invested, wrote that the company will build a “superlearner” that finds knowledge and skills from its own experience and aims to rediscover and go beyond human inventions such as language, science and mathematics. The UK government invested through the British Business Bank ($20M) and its Sovereign AI fund; Science Secretary Liz Kendall said the money “will support a company at the very frontier of AI”.

Silver told Wired that the plan is to put agents in simulations where they pursue goals, fail and improve, and that he wants “an elite AI lab that actually focuses a hundred percent on this approach”, “not just a corner of another place dedicated to LLMs”. The company’s site says that success “will represent a scientific breakthrough of comparable magnitude to Darwin”. No product, model or benchmark result has been published.

On October 12, 2026 the FT ran an interview with Silver by Madhumita Murgia. It is paywalled and was not read for this entry; only Techmeme’s one-line summary is used. CNBC and Wired were also not readable directly: their facts are taken from search summaries and from Decrypt and City AM, which quote them.

It is the largest bet so far on a route to advanced AI that does not rest on large language models, led by one of the main authors of AlphaGo and AlphaZero. Silver argues that systems which learn from their own experience can keep learning “without limit”, while human data is a shortcut that runs out.