On February 20, 2026, Anthropic launched Claude Code Security, an AI-based tool to scan codebases, identify security weaknesses, and provide patching solutions. The Claude Code preview caused a panic that resulted in billions in lost market capitalization among cybersecurity stocks.
CrowdStrike shares decreased by 8%, reaching approximately $388.87, while Okta experienced a 9.2% decline and Zscaler saw a 5.5% drop in its stock price. That demonstrates the increasing investor anxiety about AI technology developments that threaten to disrupt established cybersecurity frameworks.
Some analysts believe that Claude Code Security will reduce the need for human-centered threat detection and remediation services, which might explain a decline in the stocks of some companies. Among them are some major symbols from the stock screener — like IBM, which has been hit hard by Anthropic's announcement of Claude's ability to maintain COBOL-based systems.
Meanwhile, amid the sector’s decline, quantum computing stocks appeared among the top stock gainers — an interesting contrast given the technology’s potential use in cybersecurity.
Anthropic positions its tool as a defensive aid, which helps developers identify hidden logic errors that manual reviews cannot always detect, following more than a year-long internal testing and collaborations (like the one with Pacific Northwest National Laboratory).
However, investors interpreted it as a competitive threat to established vendors, heightening concerns about Anthropic’s growth. The company already recorded $14 billion in annual revenue by February 2026, and Claude Code accounted for $2.5 billion.
Analysts offer mixed views. Barclays deemed the sell-off "incongruent," stating that the tool protects developers by providing security features, which will not be in direct competition with companies like CrowdStrike, Cloudflare, or Palo Alto Networks.
However, the reaction reflects the "SaaS apocalypse" narrative, with Anthropic and OpenAI previews driving a reassessment across the industry. We’ve seen similar signs after Claude's legal plugin and Cowork launches.
T. Rowe Price sees this as an inflection point, a signal that traditional tech patterns are no longer safe. The broader software sector, including JFrog (down nearly 25%), suffered amid fears of AI disrupting pay-per-seat models.
Broader Implications for Cybersecurity
The Claude Code Security study shows that AI technology leads to two separate effects in cybersecurity. It protects systems while creating new security threats. AI tools can improve flaw detection with their ability to notice vulnerabilities more efficiently than humans.
The situation contains numerous potential dangers. The use of AI technologies creates a lower barrier to entry for cybercriminals who can now execute automatic cyberattacks.
A Chinese government unit used Claude Code in 80-90% of their operations in 2025 to hack 30 different organizations through automated reconnaissance and exploitation methods requiring minimal human assistance. The "agentic AI" system enables attackers to increase their operational range via phishing activities and data extortion operations.
AI-generated code introduces vulnerabilities: a study showed up to 44.73% insecure outputs in languages like C#, with flaws persisting across models. More than 30 security weaknesses exist within AI-integrated environments, allowing hackers to carry out data breaches and remote code execution through prompt-injection attacks.
The "vibe coding" increases organizational vulnerability, resulting in four times as many security breaches in certain industries. The extensive data processing operations pose a threat to data privacy, while the models themselves remain open to both poisoning and jailbreaking attacks.
The global AI cybersecurity market is projected to grow from its current value to more than $130 billion by 2030, increasing the risks for the industry.
The launch of Claude Code Security has probably received an excessive response from the market. Nevertheless, investors consider cybersecurity-related activities to be about to become obsolete. The release is surely a turning point.
The sector certainly requires organizations to develop new solutions that include human-AI hybrid systems and workflows that provide complete control and visibility to protect against autonomous threats and safeguard code development.
