The analysis, based on telemetry from diverse enterprise environments, highlights a shift where AI agents often outnumber physical hardware. In some extreme cases, researchers identified 96 AI agents for every single device, with 35% of surveyed environments containing more autonomous agents than managed assets. This influx of non-human identities, combined with a 100% turnover rate of newly discovered devices within two weeks, creates significant visibility gaps that standard endpoint solutions fail to bridge.
AI Agents and Unmanaged Assets Create Massive Security Blind Spots
Nearly every modern enterprise network harbors attacker-relevant vulnerabilities, according to new research from Vectra AI. The 2026 State of Threat Exposure Management Report reveals that the rapid proliferation of AI agents and unmanaged devices has rendered traditional, static security assessments largely obsolete in the face of evolving digital perimeters.
Security teams face a landscape where 98% of environments contain at least one attacker-relevant exposure condition. These are not merely isolated vulnerabilities but complex chains involving legacy protocols, plaintext passwords, and expired certificates. Mark Wojtasiak, SVP of Market Research and Strategy at Vectra AI, noted that exposure management must transition from simple identification to a continuous understanding of how assets and identities interact. With 30% of devices remaining unmanaged due to deployment limitations, the report underscores that the ability to monitor behavior—rather than just inventorying assets—is the new baseline for organizational resilience.
Comments (0)
No comments yet. Be the first!