Recursive Superintelligence Secures $410 Million Compute Deal with AWS
On Tuesday, the AI company Recursive Superintelligence revealed a $410 million compute agreement with Amazon Web Services. The startup, which came out of stealth in May backed by $650 million in funding, is dedicated to developing open-ended self-improving systems, an approach to AI research that demands significant computing power. The multi-year deal is designed to offer flexibility as the company scales those systems.
Recursive's $410 million commitment represents the majority of the company's total fundraising to date. However, during a call with TechCrunch, CEO Richard Socher emphasized that he expects this to be just the beginning. He described the announcement as "likely going to be one of the smallest compute deals we're going to sign in the next few years."
Focus on Self-Improving AI Systems
Recursive's focus on self-improving AI systems means that much of the budget that would typically go toward hiring and operations is instead directed at compute resources. The company aims to automate its own product development process.
"For us, it's less about headcount and more about agent count," Socher said.
Unlike major AI labs that often use hybrid investment arrangements, Amazon's involvement does not include an equity investment component. But the sheer size of the commitment allows AWS to dedicate significant resources to supporting Recursive's specific needs. This could help attract other foundational AI companies in the future.
Stay ahead of the AI curve
The most important updates, news, and content — delivered weekly.
No spam. Unsubscribe anytime.
"Part of the agreement is that we're going to co-develop infrastructure purpose-built for these types of company," said Jason Bennett, vice president for startups and venture capital at AWS.
The Promise of Recursive Self-Improvement
Recursive self-improvement (RSI) has long been viewed as a potential turning point for AI. Some experts predict a rapid acceleration in progress once AI can improve itself without human intervention. As more labs and companies pursue this idea, the specific requirements have become less clear. Some researchers expect an imminent breakthrough, while others see self-improvement as more of a gradual process.
For Recursive, the goal is to apply RSI capabilities to create actual products. Socher expects to release the first examples before the end of the year.
"We are excited to build like really amazing products that people can use, and you will see those within a few months, not within a few quarters or years," Socher said. "In October or so, you'll see some actually tangible, useful things that you'll be able to play around with."