Artificial intelligence (AI) cyber threats have reached a critical turning point, with an 89% year-over-year increase in attacks by AI-enabled adversaries, according to a report issued Tuesday from cybersecurity services provider CrowdStrike.
The CrowdStrike 2026 Global Threat Report revealed that AI is both the accelerant and the target. According to the report, adversaries exploited more than 90 organizations by manipulating generative AI tools with malicious prompts to produce commands designed to steal credentials and cryptocurrency.
Additionally, adversaries exploited vulnerabilities in AI development platforms to maintain access, deploy ransomware, and set up fake AI servers posing as trusted services to capture sensitive data.
As AI accelerates attacks, CrowdStrike’s research the average time between initial access and lateral movement to another system fell to 29 minutes, a 65% increase in speed from 2024. The fastest observed breakout time was just 27 seconds.
“This is an AI arms race,” said Adam Meyers, head of counter adversary operations at CrowdStrike, in a Feb. 24 press release. “Breakout time is the clearest signal of how intrusion has changed. Adversaries are moving from initial access to lateral movement in minutes.”
“AI is compressing the time between intent and execution while turning enterprise AI systems into targets,” Meyers said. “Security teams must operate faster than the adversary to win.”
Notably, CrowdStrike found that 82% of detections were malware-free. Instead, attackers moved through trusted systems and pathways, making their activity appear normal.
Activity connected to the People’s Republic of China increased 38% in 2025, according to the report. Sixty-seven percent of all exploited vulnerabilities by China-connected actors delivered immediate system access, while 40% targeted internet-facing edge devices.
“In the agentic era, defending against AI-accelerated adversaries, and securing AI systems themselves, requires operating at machine speed,” George Kurtz, CrowdStrike CEO and founder, said in the report’s foreword.
“The CrowdStrike 2026 Global Threat Report reflects this reality. It provides the intelligence defenders need to understand how adversaries exploit trust, accelerate with AI, and move across domains to remain evasive,” Kurtz said.
For federal agencies and industry partners, the findings arrive as the White House and agency leaders push to expand AI adoption in line with the president’s AI Action Plan.