Recursive, an AI company building recursive self-improving superintelligence to automate knowledge discovery, announced a multi-year, $410 million agreement with Amazon Web Services (AWS) to run its automated AI research system on AWS as its cloud provider.
"Partnering with AWS gives us the infrastructure to run self-improving AI research at a scale very few teams in the world are capable of," said Richard Socher, CEO and co-founder of Recursive. "We've already shown our automated AI research system can outperform years of human optimization. As our research loops run autonomously, that advantage will compound. We’re excited to partner with AWS and make our system and its infrastructure the fastest at invention generation."
Recursive is building AI that conducts experiments on how to safely improve itself, in an open-ended process of automated scientific discovery. The company emerged from stealth in May at a $4.65 billion valuation, led by GV (Google Ventures) and Greycroft, with participation from AMD Ventures and NVIDIA. Recursive was founded by former research leaders from OpenAI, Google DeepMind, Meta AI, Salesforce AI, and Uber AI, spanning natural language processing, robotics, computer vision, reinforcement learning, open-endedness, and systems research. The founding team has helped drive major advances in open-ended algorithms, quality-diversity methods, AI-generating algorithms, self-improving coding agents, automated red-teaming, prompt engineering, world models, vision transformers, and AI scientists.
"Self-improving AI creates a compounding demand for compute — every research loop generates the next experiment. In just a few months, Recursive has produced results that outpaced years of community-driven optimization," said Jason Bennett, VP and Global Head of Startups and Venture Capital at AWS. "AWS gives them the elasticity to run those loops in parallel at massive scale, and the reliability, security, and purpose-built compute to turn months of sequential research into days of parallel discovery."
In its first public results, Recursive’s automated AI research system beat a two-year human leaderboard record and outperformed a live collaborative human-AI effort, setting new state-of-the-art results across three benchmarks (NanoChat, NanoGPT Speedrun, SOL-ExecBench). As part of the collaboration, Recursive and AWS plan to co-develop infrastructure purpose-built for large-scale automated AI research.
Socher added: "Recursive self-improving AI represents one of the biggest shifts ahead for humanity. The clearest path to superintelligence is to build AI that recursively improves itself. AI is code, and AI can code. If we can close that loop properly, we can automate the scientific method itself — the same process that has driven humanity's progress in discovery and invention for centuries. That is what we're building toward at Recursive, starting with AI research itself, and partnering with AWS is how we get the infrastructure to do it at scale."
