Recursive Superintelligence Raises $650M to Build Self-Improving AI
UK-based AI research startup Recursive Superintelligence has emerged from stealth with a $650 million funding round at a $4.65 billion valuation. The company is focused on building AI systems that can autonomously improve themselves in an accelerating loop.
Team and Investors
The company was founded by former leaders from OpenAI, Google DeepMind, Meta AI, Salesforce AI, and Uber AI. CEO Richard Socher previously served as Chief Scientist at Salesforce, and co-founder Yuandong Tian was formerly Director of Meta FAIR. The round was led by GV and Greycroft, with AMD Ventures and NVIDIA also participating.
The Vision: Recursive Self-Improvement
Recursive Superintelligence is developing algorithms that enable AI to autonomously improve itself in a self-sustaining scientific discovery loop. The company targets a public launch of its "Level 1" autonomous training system in mid-2026, and plans to use the funding to secure large-scale compute infrastructure for running these systems.
Recursive self-improvement — where an AI's performance gains feed back into further capability improvements — is a central concept in long-horizon AI safety discussions. The raise comes amid unprecedented AI investment momentum, with AI capturing approximately $242 billion, or 80% of global venture funding, in Q1 2026.
Source: The Next Web, TechCrunch
Related Articles
Anthropic is putting CAD 10 million into Canadian AI research, with credits and partnerships spanning Amii, Mila, Vector and health institutions. The move links Claude distribution to safety, health and public-sector research.
Stanford HAI’s new report says the measurement gap is now part of the AI story, not a side note. U.S. private AI investment reached $285.9 billion in 2025, while documented AI incidents rose to 362 from 233 a year earlier.
Google said it is pairing new funding with AI-powered security tooling to help open source maintainers respond faster as AI increases both vulnerability discovery and attack pressure. The announcement combines a collective $12.5 million pledge through Alpha-Omega with wider use of tools such as Big Sleep, CodeMender, and Sec-Gemini.