Anthropic has delayed the release of its latest model ("Mythos") after internal testing showed it could autonomously discover and exploit software vulnerabilities at scale, including potential zero-day attacks. Instead of a public launch, access is being restricted to a small group of vetted partners under a controlled cybersecurity program.
- First impact: market fear (short-term): Initial reaction was negative for cybersecurity stocks: Sharp but brief sell-off on concerns that AI could outpace existing defence tools i.e. risk of obsolescence for legacy / signature-based solutions
- Second impact: demand tailwind (medium-term): Sentiment quickly reversed as AI-driven threats must push towards higher enterprise & government security spend and Claude announced the model will be restricted to partners of the Project Glasswing. Cybersecurity becomes mission-critical infrastructure with AI-native platforms positioned as key beneficiaries.
Winners will be companies that:
- Integrate AI deeply
- Operate at platform scale
- Move from reactive → proactive security
Although Anthropic has not publicly disclosed a full partner list for its "Project Glasswing", the next table offers an overview of the Cyber Security industry key players with a last column highlighting "Yes/Probable" for those players that have been reported participation or strong likelihood based on enterprise partnerships and mentions in coverage:
Recommended reads:
https://www.axios.com/2026/04/08/anthropic-mythos-model-ai-cyberattack-warning
https://www.businessinsider.com/anthropic-mythos-latest-ai-model-too-powerful-to-be-released-2026-4
https://www.investors.com/news/technology/cybersecuity-stocks-anthropic-mythos-project-glasswing/
------------------------------
Carlos Salas
Portfolio Manager & Freelance Investment Research Consultant
------------------------------