CrowdStrike CEO Rejects Anthropic's Call to Slow AI Development
George Kurtz argues that global AI progress is unstoppable, shifting the focus from pacing development to securing it.
CrowdStrike CEO George Kurtz has rejected a call for an immediate slowdown in artificial intelligence development, arguing that such efforts are futile. The disagreement follows a high-profile warning from Anthropic CEO Dario Amodei regarding the catastrophic risks of unchecked AI acceleration.
In September 2026, Amodei urged the industry to adopt a strategy of "pacing" AI development. According to reports from Politico and Nairametrics, Amodei cited the potential for AI to facilitate cyberattacks, bioterrorism, and widespread economic disruption, suggesting that the current race is moving faster than humanity can effectively control. He called for immediate coordination among AI labs to mitigate these existential threats.
The Futility of Pacing
Responding to these concerns, George Kurtz contended that the frontier of AI development will continue to move at its own speed regardless of calls for restraint. Speaking via CNBC, Kurtz argued that the rest of the world will not slow down, making any attempt to centrally pace the technology an impractical strategy. He emphasized that because the global nature of AI development is so decentralized, a voluntary slowdown by a few labs would not eliminate the inherent risks.
A Shift Toward Defense
This clash highlights a fundamental divide in the tech industry between those advocating for centralized safety controls and those favoring a "defense-in-depth" approach. While Anthropic positions itself as a safety-first organization that believes development must be slowed to ensure security, CrowdStrike represents the view that ubiquity is already a reality. From this perspective, the only practical path forward is to strengthen the cybersecurity infrastructure surrounding the technology rather than attempting to halt its progress.
The Path Forward
Kurtz maintains that the primary responsibility of the cybersecurity community is to ensure that AI evolves securely and safely. Rather than focusing on the speed of the "frontier," he argues the industry must prioritize the tools and protocols necessary to defend against the capabilities these models provide. As AI labs continue to release more powerful systems, the debate now shifts toward whether safety can be managed through policy and pacing or if it must be solved through technical resilience and constant vigilance.