Cybersecurity Stocks Dip as Anthropic’s AI Boosts Code Security
Cybersecurity software stocks experienced a sharp decline last week following the release of Anthropic’s Claude Code Security, an artificial intelligence tool designed to identify vulnerabilities in software code. Shares of companies including CrowdStrike, Cloudflare, Okta, SailPoint, and Zscaler all fell significantly on Thursday, wiping billions of dollars off their combined market value, according to reports.
The selloff was triggered by the unveiling of Claude Code Security, which differs from traditional security scanners by employing AI to “reason about code the way a human security researcher would,” as Anthropic described it. Existing scanners typically rely on pattern matching to detect known vulnerabilities. Claude Code Security, currently available as a limited research preview, analyzes how data flows through applications and identifies more complex flaws.
“There’s been steady selling in software, and today it’s security that’s getting a mini-flash crash on a headline,” said Dennis Dick, head trader at Triple D Trading, in a statement to Bloomberg News. “This kind of market is scary for investors, because things are just moving relentlessly to the downside as soon as you obtain a hint of disruption. It’s rational to be cautious, because people were saying a whereas ago that the software drop was overdone, and yet it keeps going down.”
The decline in cybersecurity stocks mirrors a similar trend observed last summer, when shares of companies like Salesforce and Workday also fell amid concerns about the impact of AI on the software market. Investors are increasingly worried that the rise of AI-powered code generation tools will reduce demand for traditional software products, allowing users to develop applications independently.
Anthropic’s new tool is not the first to raise these concerns. The launch of new AI tools from Google and OpenAI has also contributed to the recent volatility in the software sector. Though, recent research suggests that AI-assisted coding is unlikely to fully replace human developers in the near future. Studies have shown that AI models perform best when developers review their outputs at key stages, rather than operating autonomously. Without human oversight, these models can produce code that is less maintainable and contains security flaws.
Mark Allen, director of legal operations and strategy for legal and compliance at Zillow Group Inc, noted that many large companies already have enterprise subscriptions to AI models like Claude, OpenAI’s GPT, or Google’s Gemini. This could incentivize legal teams to build solutions directly on these platforms, potentially reducing the need for specialized legal tech vendors. Anthropic has also released a legal-specific plugin for its Cowork tool, capable of reviewing contracts, flagging risks, and automating workflows.
Despite the market reaction, Claude Code Security remains in a research preview phase and requires human review to apply suggested patches. Anthropic has provided expedited access to the tool for open-source maintainers. The long-term impact of the tool on the cybersecurity industry remains to be seen.