Stay informed with free updates
Simply sign up to the Artificial intelligence myFT Digest — delivered directly to your inbox.
A four-month-old start-up founded by former engineers at Google DeepMind and OpenAI has raised at least $500mn, as investors bet on the group building a new kind of AI that teaches itself.
Recursive Superintelligence has been valued at $4bn, excluding the new capital, in a funding round led by GV, formerly Google Ventures, with contributions from chipmaker Nvidia, according to people familiar with the financing.
After securing an initial $500mn investment, the round was so oversubscribed that Recursive could raise as much as $1bn in total, the people added.
Recursive’s founders include San Francisco-based Richard Socher, a highly regarded AI researcher who previously worked as Salesforce.com’s chief scientist, and London-based Tim Rocktäschel, a professor of AI at University College London who until recently was a principal scientist at Google DeepMind, where he worked on projects including the Genie interactive world model.
The company, which has about 20 staff, also includes former OpenAI researchers Josh Tobin, Jeff Clune and Tim Shi, as well as others from Google and Meta. “It’s a ridiculously strong team,” said one person close to the start-up.
Recursive ultimately hopes to create an AI system that can continuously improve itself without human intervention, according to people familiar with its plans. However, the concept remains at the research stage and has not yet been proven to work over extended periods.
The deal is the latest sign of continued venture capital enthusiasm for AI start-ups, even after OpenAI and Anthropic have amassed war chests running into tens of billions of dollars as they look to take on Big Tech giants Google, Meta and Microsoft.
Start-up investment hit an unprecedented $300bn in the first quarter of 2026, driven by massive deals for OpenAI, Anthropic, xAI and Waymo — far exceeding any previous quarterly total, according to Crunchbase, which tracks VC deals.
Bloomberg reported in January that Recursive was looking to raise hundreds of millions of dollars in funding.
It is one of several new AI research labs to spin out from OpenAI, Google and Meta in recent months. These include Thinking Machines Lab, Safe Superintelligence, Ineffable Intelligence and Advanced Machine Intelligence Labs.
VCs hope that these start-ups can be more nimble in tackling novel approaches to AI, at a time when OpenAI and Anthropic are focused on fuelling the growth of their existing platforms ChatGPT and Claude.
Recursive has not yet officially announced its existence, but a filing at the UK’s Companies House register shows that it was incorporated in London at the end of last year.
GV declined to comment. Nvidia and Socher did not respond to requests for comment.
Additional reporting by Cristina Criddle
