Stay informed with free updates<\/p>\n
Simply sign up to the Artificial intelligence myFT Digest — delivered directly to your inbox.<\/p>\n
David Silver, one of Britain\u2019s top AI researchers, is raising $1bn for a new company aiming to build \u201csuperhuman intelligence\u201d in a deal led by Sequoia Capital that would be the largest seed round ever in Europe.<\/p>\n
The round is expected to value the nascent start-up at about $4bn, not including the new investment, according to people with knowledge of the matter. It comes amid a rush to back prominent researchers leaving big AI<\/a> labs to work on experimental technologies.<\/p>\n Silver left Google DeepMind late last year to launch the London-based start-up Ineffable Intelligence. <\/p>\n His departure from DeepMind kicked off a highly competitive process to back his new venture, with Sequoia<\/a> partners Alfred Lin and Sonya Huang flying to London to visit Silver, who is also a professor at University College London. <\/p>\n Big Tech groups including Nvidia, Google and Microsoft are also in talks to invest, the people said, adding that negotiations are live and the final terms could change. If concluded at $1bn, it would be the largest first round ever raised by a European start-up<\/a>, according to PitchBook.<\/p>\n Top researchers have raised unprecedented sums for untested start-ups promising to build AI that will outperform large language models from OpenAI<\/a>, Google or Anthropic.<\/p>\n Such is the flurry of activity that last month two separate start-ups, Ricursive Intelligence and Recursive AI, both undertook funding talks with a price tag of $4bn, according to people familiar with those deals.<\/p>\n Ineffable Intelligence will build on Silver\u2019s work training AI systems through experience or interactions with their environment \u2014 a process known as reinforcement learning. Most leading AI models are trained on large volumes of text culled from the internet, before being fine-tuned via feedback. <\/p>\n In a paper on the \u201cEra of Experience\u201d, published last year, Silver and Canadian computer scientist Richard Sutton predicted that a \u201cnew generation of agents will acquire superhuman capabilities by learning predominantly from experience\u201d.<\/p>\n AI agents capable of independently completing complex tasks based on human instructions are increasingly regarded as the key to advancing and commercialising AI by large labs and start-ups alike.<\/p>\n Experience will \u201cbecome the dominant medium of improvement and ultimately dwarf the scale of human data used in today\u2019s systems\u201d, Silver and Sutton predicted. <\/p>\n Silver joined DeepMind soon after it was founded in 2010 and has been behind some of the AI group\u2019s landmark breakthroughs. He worked closely on AlphaGo and AlphaStar, AI programmes capable of beating the world\u2019s best players of the strategy game Go and the video game StarCraft.<\/p>\n He has also been instrumental in developing Google\u2019s Gemini family of LLMs, after the Big Tech company bought DeepMind in 2014. <\/p>\n The 49-year-old is one of a number of prominent researchers in the field who have formed their own companies in recent months. <\/p>\n OpenAI executives Ilya Sutskever and Mira Murati have raised billions since leaving the ChatGPT maker in 2024; Arthur Mensch left Google DeepMind in 2023 to start French start-up Mistral and raised \u20ac105mn<\/a> in a seed funding round that year.<\/p>\n