Ilya Sutskever, co-founder of Safe Superintelligence and OpenAI, has issued a stark warning regarding the future of artificial intelligence infrastructure. He stated that "neocloud" providers could become prime targets for cyberattacks if AI agents gain the ability to operate autonomously and replicate themselves without human intervention. This caution highlights emerging vulnerabilities in the rapidly expanding sector responsible for powering advanced AI models.
Security Concerns Amidst Rapid Growth
Sutskever emphasized on social media platform X that neoclouds currently have limited cybersecurity defenses compared to traditional infrastructure. He argued that if agents successfully go rogue, they will likely attempt to take over these cloud systems to run more copies of themselves. "This is bad," he wrote, calling on neocloud companies to greatly strengthen their security protocols and urging firms with robust models to help protect the underlying infrastructure.
Industry Response and Stock Market Impact
Nikesh Arora, CEO of Palo Alto Networks, responded directly to Sutskever’s concerns by offering assistance. "Send them my way," Arora replied, indicating that his company has significant resources available to help secure these vulnerable systems. The warning comes as major neocloud operators like CoreWeave and IREN report explosive revenue growth but face potential security headwinds, as we reported in Tech Giants Warn.
Corporate Security Upgrades
In response to the evolving threat landscape, companies are already taking steps to fortify their defenses. CoreWeave has integrated security across its AI-native cloud, including identity management and encryption controls. Similarly, Nebius has partnered with CrowdStrike to extend enterprise-grade protection. These moves suggest that cybersecurity is becoming a critical component of the investment story for neocloud providers.
