Stock Markets March 30, 2026

Leak of Anthropic’s Claude Mythos Spurs Drop in Cybersecurity Stocks; DA Davidson Backs Select Vendors

A leaked draft describing Claude Mythos’ cyber capabilities rattles the market, but analysts see durable demand for established security vendors.

By Jordan Park
Leak of Anthropic’s Claude Mythos Spurs Drop in Cybersecurity Stocks; DA Davidson Backs Select Vendors

Shares of cybersecurity companies fell on Monday after a leaked draft blog post from Anthropic described a new model, Claude Mythos (also called Capybara), as "far ahead of any other AI model in cyber capabilities." DA Davidson analysts say they do not expect frontier AI tools to displace incumbent security vendors, highlight continued CISO skepticism, and name CrowdStrike, Palo Alto Networks and Okta as well positioned to capture revenue from securing AI.

Key Points

  • A leaked Anthropic draft described Claude Mythos (Capybara) as "far ahead of any other AI model in cyber capabilities," triggering a sharp sell-off in cybersecurity shares.
  • DA Davidson judges it unlikely that frontier AI tools will displace vendors that provide real-time detection and response, and expects material financial benefits from securing AI only in the second half of the year at the earliest.
  • The firm named CrowdStrike, Palo Alto Networks, and Okta as the companies in its coverage best positioned to benefit from securing AI, assigning Buy ratings to all three.

Cybersecurity equities slid sharply on Monday after a leaked draft blog post from Anthropic outlined a new artificial intelligence model called Claude Mythos, also referred to as Capybara. The leaked material described the model as "far ahead of any other AI model in cyber capabilities," a characterization that prompted immediate market reaction.

Despite the market drop, analysts at DA Davidson maintain a cautious view about the risk posed to existing security vendors. In their assessment, it remains highly unlikely that tools emerging from frontier AI labs will completely displace cybersecurity companies that provide real-time detection and response solutions.

The firm said material revenue benefits from securing AI are unlikely to show up in corporate financials until at least the second half of this year. That timetable underlines DA Davidsons view that adoption and monetization of AI-focused security offerings will be gradual rather than immediate.

DA Davidson cited conversations held at the RSA conference as reinforcing their perspective. According to the analysts, most Chief Information Security Officers they spoke with remained highly skeptical of new security solutions coming out of frontier AI labs, and the firm does not believe CISOs are actively awaiting those offerings.

On product categories, DA Davidson argued that endpoint, cloud security, network security, and identity security solutions are highly defensible against potential competition from AI labs. These segments, the analysts said, benefit from established integration, operational maturity, and real-time response capabilities that are difficult to replicate quickly.

The firm cautioned, however, that headline risk will likely persist in the near term. Announcements or demonstrations from Anthropic or OpenAI could continue to drive volatile trading in cybersecurity stocks regardless of whether investors believe those solutions present a substantive threat to incumbent vendors.

DA Davidson defined a "material benefit" from securing AI as roughly 5% to 10% or more of net new annual recurring revenue attributable to AI security efforts. Based on that framework, the analysts identified three firms in their coverage as best positioned to benefit from the market opportunity created by securing AI: CrowdStrike, Palo Alto Networks, and Okta. DA Davidson assigned Buy ratings to all three names.


Context and market implications

The leak and ensuing market volatility highlight how announcements about advanced AI capabilities can produce immediate headline-driven moves in security stocks. At the same time, DA Davidsons analysis suggests that operational realities and buyer skepticism could limit near-term disruption to the incumbent vendor landscape.

Risks

  • Headline risk - Announcements from Anthropic or OpenAI could continue to produce volatile trading in cybersecurity stocks, affecting investor sentiment across the sector.
  • Adoption timing - Material revenue contribution from securing AI is unlikely to materialize until at least the second half of the year, leaving near-term financial uncertainty for vendors pursuing AI security opportunities.
  • Buyer skepticism - Chief Information Security Officers expressed high levels of skepticism toward frontier AI lab solutions at the RSA conference, which could slow procurement and deployment in enterprise cybersecurity budgets.

More from Stock Markets

Bernstein Sees Crypto Equities Approaching Cycle Low Ahead of Weak Q1 Results Mar 30, 2026 Alaska Air Warns of Larger Q1 Loss as Fuel Prices Spike and Regional Demand Softens Mar 30, 2026 White House Conducting Review of SEC Plan to Shift Corporate Reporting Cadence Mar 30, 2026 Sysco to Buy Jetro Restaurant Depot in $29 Billion Transaction Mar 30, 2026 Velo3D Stock Climbs After $9.8 Million Defense Agreement Mar 30, 2026