So, you know how everyone's been talking about AI as this amazing tool that's going to solve all our problems? Well, it turns out the bad guys got the memo too—and they're using it to break into things much, much faster.
That's the not-so-subtle takeaway from the 2026 Global Threat Report released Tuesday by CrowdStrike Holdings Inc. (CRWD). The cybersecurity firm, which tracks over 280 named adversary groups, paints a picture of a digital battlefield where artificial intelligence has decisively shifted from being a defensive asset to the offensive weapon of choice.
"This is an AI arms race," said Adam Meyers, head of counter-adversary operations at CrowdStrike. He's not being dramatic. The data shows the race is already on, and the pace is terrifying.
When Breakouts Happen in the Blink of an Eye
The most jarring metric in the report is what's called "breakout time." In cybersecurity speak, that's the window between an attacker's initial access to a system and when they start moving laterally through it to do real damage. It's the clock security teams are racing against.
That clock is now running at warp speed. The average eCrime breakout time collapsed to just 29 minutes in 2025. That's a 65% acceleration from 2024. Think about that: from first touch to full-scale intrusion in less time than it takes to watch a sitcom.
But the average is almost comforting compared to the extremes. The fastest recorded intrusion in CrowdStrike's data happened in a breathtaking 27 seconds. In one documented case, the attackers began siphoning data out just four minutes after getting in. At that speed, by the time a human defender gets a coffee and sits down at their desk, the attack is already over and the data is gone.
"Breakout time is the clearest signal of how intrusion has changed," Meyers explained. "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. Security teams must operate faster than the adversary to win."
A Rocky Market for the Cyber Sentinels
It's a bit ironic that this alarm bell is ringing just as the companies tasked with building the defenses are having a rough go of it in the market. CrowdStrike's stock is down roughly 22.76% year-to-date, caught in a broader selloff that has hit SaaS and cybersecurity names hard.
Part of that pressure was triggered last week when AI lab Anthropic unveiled Claude Code Security, a move that spooked some investors into wondering if big AI players might start eating the cybersecurity industry's lunch.
But not everyone on Wall Street is hitting the panic button. Wedbush Securities analyst Dan Ives has called the sector fears a "huge disconnected overhang." His view is that cybersecurity remains "the next frontier for the AI Revolution" and that companies like CrowdStrike, Palo Alto Networks Inc. (PANW), and Zscaler Inc. (ZS) stand to be the primary beneficiaries as AI's role in security expands.
In other words, the very technology making attacks more dangerous is also the key to building better defenses—and that's a massive business opportunity for the firms that can do it.
