The shift toward AI-driven offense coincides with a surge in critical vulnerabilities at 21 large software firms, including Apple, AWS, Microsoft and Google, where monthly reports jumped from under 100 to over 600 since spring. Approximately 87% of exploited bugs are now targeted on or before the day of public disclosure, up from 23% in 2020, and 86.7% of all 2026 CVEs were zero-days. OpenAI reacted to the trend by redirecting 25% of its production engineers to use its Astra model to find and patch gaps in its own internal architecture.
Wednesday, October 21, 2026
Story
Armadin Finds 90+ Fortune 500 Zero Days With AI Attack Swarms
Cybersecurity14 days ago
Mandiant founder Kevin Mandia has launched a security firm that utilizes artificial intelligence to proactively attack client networks to identify exploitable vulnerabilities. Armadin Security has uncovered more than 90 zero-day holes in the production systems of Fortune 500 companies since January using autonomous "drone swarm" agents that map services and routes from the outside. These models target logic flaws in custom applications without requiring source code access, allowing the company to notify security chiefs of remote code execution risks shortly after discovery.
Looking for more context or analysis?
Tilt chat provides reliable, tailored insights, alerts, and automations.