CrowdStrike, Okta lead cyber selloff after Anthropic’s Claude update

Cybersecurity stocks were hit hard on Friday as investors digested a new move from Anthropic that pushes Claude further into a job long dominated by specialist security software.

CrowdStrike and Okta led the decline, with other cybersecurity names moving lower as investors weighed recurring strategic questions.

The answer is showing up in prices with​ CrowdStrike slipping as much as 6.5%, Cloudflare dropping more than 6%, Okta plunging 5.7%, Zscaler falling 3.5% and SailPoint shedding 6.8%.

The Global X Cybersecurity ETF declined as much as 3.8%, extending its year-to-date loss to 14%.​

Why Claude spooked the market

Anthropic said its new capability, “Claude Code Security,” scans codebases for security vulnerabilities and suggests targeted software patches for human review, and it is being released in a limited research preview.

In simple words, it reads a company’s software like a security analyst would, flags potential weak spots, and drafts fixes, but leaves the final decision to humans, rather than auto-applying changes.​

That “human-in-the-loop” design matters because false alarms are expensive in security, and bad fixes can create new problems.

Anthropic says the tool tries to reduce that risk by re-checking its own findings in multiple stages, assigning severity ratings so teams can focus, and adding a confidence rating so reviewers can judge how much to trust each alert.​

Also Read: Anthropic lands $30B at $380B valuation as AI funding hits new extreme

AI as a product, not just a feature

Friday’s selloff also fits into a broader market narrative that’s been building for weeks.

The investors are increasingly wary that AI-native tools could chip away at software companies’ pricing power by offering “good enough” alternatives embedded in the coding workflow.

The fear is not that security spending disappears, but that some work shifts from standalone security products toward AI-assisted scanning and remediation that feels more like a built-in utility than a separate subscription line item.​

To be clear, Anthropic is positioning Claude Code Security as a defensive tool and stresses that nothing is applied without human approval.

It also frames the rollout as a controlled preview, limited to Enterprise and Team customers, with expedited access for maintainers of open-source repositories.

But markets often trade the direction of travel, not the fine print, and the direction here is that frontier AI models are moving from writing code to policing it.​

For the cybersecurity vendors that sold off, the near-term question is whether this is a headline-driven reset or the start of a tougher competitive conversation in earnings calls.

Investors will be watching for signs that incumbents can defend their moat by tying AI into broader platforms like endpoint protection, identity, network security, and response, where outcomes, data and real-time threat intel still matter as much as code analysis.