CrowdStrike report: AI agents slash exploit window from months to hours

CrowdStrike's 2026 Threat Hunting Report revealed alarming findings about AI agents' impact on cybersecurity timelines. According to CrowdStrike's senior vice president of intelligence Adam Meyers, the time window between a vulnerability's release and the availability of a working exploit has collapsed from months or years to mere hours. This acceleration is driven by AI agents' ability to rapidly identify and exploit vulnerabilities. The report also highlighted that AI-enabled attacks are now targeting trusted identities within organizations' AI supply chains rather than using traditional hacking techniques. Industry leaders, including Anthropic through its Project Glasswing initiative, are collaborating to provide early access to advanced models so the industry can better prepare defenses against emerging threats.