MW Palo Alto Networks and other cybersecurity stocks slide on fresh Anthropic fears. Investors may be overreacting.
By Christine Ji
Cybersecurity stocks tumbled Friday morning on the news of a new Anthropic model, but analysts argue that AI will ultimately be a tailwind for the sector
A draft blog post detailing Anthropic's new model was leaked on Thursday.
Shares of CrowdStrike, Palo Alto Networks and other cybersecurity stocks are among the worst performers in the S&P 500 index Friday on renewed concerns about competition from artificial-intelligence models.
But analysts aren't convinced that AI has disrupted the cybersecurity business model just yet.
A Thursday report by Fortune revealed that Anthropic is developing a new model called "Claude Mythos," which includes "dramatically higher scores on tests of software coding, academic reasoning, and cybersecurity." Fortune accessed the confidential files after an error in Anthropic's content-management system left a draft blog post and nearly 3,000 internal assets in a publicly searchable data cache.
Some investors fear the AI model will be so good at detecting threats that it will reduce demand for traditional cybersecurity offerings, similar to how advancements in AI coding tools have sparked fears in the broader software sector.
An Anthropic spokesperson reportedly told Fortune that the new model represented "a step change" in AI capabilities and was being trialed by "early access customers."
Investors starting selling on the news Friday "like clockwork," Bernstein analyst Peter Weed wrote, in reference to the trend of software stocks falling after Anthropic product developments. Shares of CrowdStrike (CRWD), Palo Alto Networks (PANW) and Zscaler (ZS) fell 6%. Shares of Okta $(OKTA)$ were down 7%.
But the selloff is grounded in reality, according to Weed. "This does not appear to reduce the potential cybersecurity sector tailwinds" from generative AI and large language models, Weed wrote.
Read: These 4 cybersecurity stocks are Wall Street's favorite AI-proof plays
This isn't the first time Anthropic has caused a slide in cybersecurity stocks. Last month, the launch of Claude Code Security triggered a similar reaction after the tool proved it could find and fix complex bugs faster than human experts.
But many analysts think AI will prove a net benefit to traditional cybersecurity vendors. The boom in AI agents has increased the attack surface for organizations, and heightened AI capabilities will only make bad actors more powerful. That should lead to a structural demand increase for cybersecurity services.
In fact, Anthropic cautioned that the model's more powerful cybersecurity capabilities will be able to "exploit vulnerabilities in ways that far outpace the efforts of defender," according to the leaked blog post. The post revealed that Anthropic is releasing the model in early access to cyber defenders to help them improve their infrastructure.
Weed emphasized that software flaws, which are the primary target of Claude's code-scanning abilities, are only a small source of actual intrusions. And most organizations, in his view, will remain model-agnostic, meaning they will need protection that covers multiple different AI agents and platforms.
In a Thursday note prior to reporting on the leak, Jefferies analyst Joseph Gallo wrote that Palo Alto Networks is largely insulated from AI disruption because up to 90% of its business relies on "deterministic" security, or binary decisions to block or allow traffic. AI models are "probabilistic," meaning they guess based on probability and don't provide high enough levels of accuracy.
"We continue to believe it is highly unlikely that tools from the frontier labs will displace cybersecurity vendors, particularly those that primarily provide solutions that do real-time detection & response," D.A. Davidson analyst Rudy Kessinger wrote Friday morning. However, he acknowledged that a material boost from AI won't show up in cybersecurity companies' results until the second half of the year "at the earliest."
Kessinger also added that "headline risk" will continue to create volatility among cybersecurity names in the near term. "Whether investors believe these solutions actually pose a threat to existing cybersecurity vendors is pretty much beside the point," he added.
Read on: Why CrowdStrike's stock just won another fan on Wall Street
-Christine Ji
This content was created by MarketWatch, which is operated by Dow Jones & Co. MarketWatch is published independently from Dow Jones Newswires and The Wall Street Journal.
(END) Dow Jones Newswires
March 27, 2026 11:36 ET (15:36 GMT)
Copyright (c) 2026 Dow Jones & Company, Inc.
Comments