A new investment narrative is taking shape in the artificial intelligence sector, according to comments cited by The Information. Jasjeet Sekhon, chief strategy officer at Google DeepMind, said one of the core assumptions behind the industry’s unprecedented capital spending is that future AI systems may achieve recursive self-improvement, or RSI. In his view, that idea is starting to emerge as a fresh thesis after AGI, or artificial general intelligence.
Sekhon described RSI as a scenario in which AI systems analyze and improve their own capabilities, develop more efficient algorithms, refine model architectures, and help drive stronger next-generation systems. He said infrastructure spending on the current scale needs a higher-order technological breakthrough to justify its long-term value, and RSI is one of the directions now drawing attention.
He also said the industry is moving beyond simply scaling models and compute, toward more autonomous AI agents and intelligent systems. At the same time, he noted that RSI remains exploratory. While current models can already generate code, use tools, and assist research and development, fully autonomous iteration with limited human involvement still faces technical, safety, and controllability challenges.
According to The Information, Google DeepMind Chief Strategy Officer Jasjeet Sekhon said one of the core ideas behind the AI industry’s unprecedented wave of capital spending is a bet that future AI systems will be capable of recursive self-improvement, or RSI.
RSI refers to AI systems being able to analyze and optimize their own capabilities, design more efficient algorithms, improve model architectures, and keep strengthening the next generation of AI systems. Sekhon said RSI is emerging as a new central investment narrative for the AI industry after AGI, or artificial general intelligence.
Why infrastructure spending is tied to RSI
Sekhon recently said infrastructure investment on this scale in AI needs a more advanced technological breakthrough to support its long-term value, and RSI is one of the key directions. He said the industry is shifting from simply expanding model size and compute investment toward exploring more autonomous AI agents and intelligent systems.
If AI can continuously improve itself, that could change software development, scientific research, and the way companies operate, while raising return expectations for AI infrastructure spending.
Still at an exploratory stage
For now, RSI remains at an early stage. Sekhon said that although AI models already have capabilities such as code generation, tool use, and research assistance, major technical, safety, and controllability hurdles remain before they can truly achieve autonomous iteration and capability jumps without heavy human involvement.
As tech giants continue to spend hundreds of billions of dollars building AI data centers and computing infrastructure, the market is watching whether those investments can eventually push AI from large-scale training into a new phase of self-evolution.
This article was originally published by Bit.Fan. For more cryptocurrency news and market insights, visit www.bit.fan. Disclaimer:
The market information, project data, and third-party content displayed on this platform are for industry information sharing only and do not constitute any form of investment advice or return commitment.
Cryptocurrency trading carries high risks. Users should fully assess their risk tolerance and make independent decisions. All profits, losses, and legal responsibilities are borne by the users themselves.
