A new report out today from cloud-native application security firm Sysdig Inc. reveals that cloud security is evolving at an unprecedented pace, as organizations surprisingly reduce detection times as cybersecurity systems evolve.

The finding comes from the Sysdig 2025 Cloud-Native Security and Usage Report, which also found an increase in artificial intelligence workloads, an increasing risk from machine identities and a shift toward real-time threat detection and open-source security adoption.

The report found that cloud attacks are now being detected, investigated and mitigated within 10 minutes, reflecting the growing emphasis on real-time security. Companies leveraging advanced detection tools were found to be now able to identify threats in as little as five seconds, enabling faster responses to potential breaches.

Machine identities are highlighted in the report as a growing concern, with their numbers vastly outpacing human identities in cloud environments. These identities are 7.5 times riskier than human accounts, increasing the attack surface for cybercriminals targeting cloud infrastructure.

At the same time, AI workloads were found to have surged by a stunning 500% over the past year, surprising as organizations have managed, at the same time, to improve their security posture. Public exposure of AI workloads has dropped by 38%, demonstrating a concerted effort to protect AI-powered applications from external threats.

Open-source security adoption was found to be on the rise, with tools such as Falco surpassing 140 million downloads. Organizations of all sizes were found to be embracing open-source solutions to detect runtime threats and enhance cloud security.

Vulnerability management has seen notable progress, with in-use vulnerabilities decreasing significantly. However, image bloat in containerized environments has quintupled, leading to inefficiencies and increased security risks.

Another area highlighted in the report was compliance, which is becoming a top priority, particularly for organizations in the European Union, where regulatory scrutiny is high. EU-based companies were found to be implementing stricter security policies compared to their global counterparts, reducing their risk exposure.

“The real-world data used to derive the report’s findings highlights the challenges and opportunities in modern cloud environments for the year to come,” the authors write in the conclusion. “This year’s analysis also underscores meaningful progress in vulnerability management and AI workload security while revealing a staggering imbalance between service and user accounts, regardless of how risky they may be.”

Image: SiliconANGLE/Ideogram