Security teams can’t keep up with how fast AI is cracking cloud security

Check Point’s 2026 Cloud Security Report warns that AI is overwhelming cloud defenses
While 77% have updated cloud strategies, only 26% have architectures capable of enforcing them
Researchers call for a unified, prevention-first architecture

Artificial intelligence is cracking cloud security at a rapid rate, and security teams just can’t keep up. This is one of the conclusions of the “2026 Cloud Security Report: Enter the AI Era”, a new in-depth report published by Check Point Software Technologies.

In the report, shared with TechRadar Pro earlier this week, Check Point claims that companies are aware of the risks AI poses in the wrong hands, but simply don’t have the means to deal with it. Apparently, in response to AI, 77% of organizations have updated their cloud security strategy this year, but only a quarter (26%) have the architecture to actually enforce it.

At the same time, AI is increasingly being weaponized in phishing and malware attacks at speeds that “traditional security models” cannot respond to.

Surpasses architecture

“The impact is already measurable: 78% of organizations reported confirmed or suspected AI-related security incidents over the past year,” Check Point said.

“AI adoption has outpaced the architecture built to manage it. Agents act inside live systems; data moves through external AI services, and most businesses still lack the visibility and enforcement to keep pace,” commented Stuart Green, Cloud Solution Architect at Check Point. “Visibility, control and security must be present at every layer of the stack AI workloads will operate in.”

There are numerous challenges for companies, especially cloud-native environments, the report further emphasizes. In addition to infrastructure misalignment (52% of AI workloads span hybrid environments, yet 64% confirmed their architecture needs to be redesigned), there are serious perimeter gaps (76% rated data center security as critical to AI, but only 35% said it can support current trends), as well as performance challenges (only 25% can fully impact the performance of AI traffic).

Finally, there are issues with operational complexity (88% said AI increased security complexity), as well as issues with limited visibility (54% experienced an AI-related security incident, with 24% saying they could not confirm due to lack of visibility).

To mitigate these risks, enterprises need a unified, prevention-first architecture across cloud, data center, SaaS and endpoints, says Check Point.