Cybersecurity Stocks CrowdStrike, Palo Alto Hit By New Anthropic Claude Worries

• Anthropic is developing 'Mythos,' a new Claude AI model featuring computer security tools, and has launched 'Claude Code Security' to detect vulnerabilities and 'Cyber.AI' system for threat detection and response
• Partnerships with Accenture and Databricks will create agentic security systems operating at 'machine speed' to combat increasingly sophisticated AI-powered cyberattacks
• The IBD computer security group ranks only 186 out of 197 tracked groups as cybersecurity stocks have underperformed in 2026 amid AI disruption concerns

Cybersecurity Stocks Decline on Anthropic AI Competition Concerns

Key Developments

Cybersecurity stocks fell Friday following reports that AI model maker Anthropic is developing new software tools that could disrupt the industry. The company's new Claude AI model, code-named "Mythos," features advanced computer security capabilities unveiled at the RSA conference in San Francisco.

Major Companies Affected

Stock Declines: CrowdStrike Holdings, Palo Alto Networks, Zscaler, and Fortinet all experienced losses. The IBD computer security group ranks only No. 186 out of 197 groups tracked, reflecting significant 2026 underperformance.

New Partnerships: Anthropic announced collaborations with Accenture and Databricks. Accenture will develop services using Anthropic's Cyber.AI system for threat detection and response. Databricks is launching "Lakewatch," an agentic security analytics system built on Anthropic's platform. Competitor Snowflake was also mentioned in the context of data analytics software.

Market Implications

This marks the second wave of concern over Anthropic's competitive threat—the initial pressure occurred in early February with the release of "Claude Cowork" coding tools. Cybersecurity stocks have broadly underperformed in 2026 amid fears of AI model disruption across software, finance, healthcare, and legal services.

The industry focus is shifting toward "agentic" AI technology—autonomous, goal-driven agents that execute multistep tasks independently, representing an evolution beyond conversational AI co-pilots. Both hackers and security firms are deploying these new AI tools.

Citi analyst Fatima Bolani noted the systems enable "machine speed to machine speed combat" for security breaches, allowing faster investigations through AI agents operating on real-time data. The competitive landscape suggests traditional cybersecurity companies face increasing pressure from AI-native platforms.