Anthropic's 'Step Change' AI Model Leak Sparks Cybersecurity Stock Selloff

A leaked report detailing a new, highly advanced AI model from Anthropic has triggered a sharp selloff in cybersecurity stocks. The premarket slump signals a return of the 'AI disruption' trade, as investors react to a Fortune report indicating the model "poses significant cybersecurity risks." The core tension stems not just from the model's capabilities, but from the manner of its revelation: descriptions were inadvertently stored in a publicly accessible data cache, exposing the development before its official announcement.

According to the report, Anthropic is actively developing and has begun testing this new model with early access customers. An Anthropic spokesperson explicitly called it a "step change" in AI performance and "the most capable we’ve built to date." The leak originated from a draft blog post that was available in an unsecured, publicly-searchable data store, which Fortune reviewed. This accidental disclosure provided the market with an early, unvarnished look at a potential paradigm shift in AI, directly linking its advancement to heightened cybersecurity concerns.

The immediate financial fallout is concentrated in the cybersecurity sector, where stocks are tumbling on the prospect that this new AI could outpace or undermine existing digital defenses. The incident underscores the market's acute sensitivity to AI developments framed as security threats. Furthermore, the data leak itself represents a significant operational security lapse for Anthropic, raising questions about internal controls even as it touts a breakthrough in external capability. The situation places Anthropic under dual scrutiny: for the power of its new model and for the breach that revealed it.