"AI is both the weapon and the target," CrowdStrike counter adversary division senior VP Adam Meyers told reporters, summing up a rapid shift in how attackers operate and what they aim at.
CrowdStrike's headline figures: 89% rise, 290 tracked adversaries, and a 2.5x shift to AI agents
CrowdStrike's annual Threat Hunting Report documents an 89 percent increase in attacks by AI-enabled adversaries in 2025, and the vendor now tracks more than 290 adversary groups, having added about ten this year. The firm's threat-hunting team is logging AI agent–triggered leads at 2.5 times the rate of human-triggered threats, a volume increase the report says is true across both government-backed actors and financially motivated criminals.
Famous Chollima: North Korean crews weaponize fake companies and supply chains
CrowdStrike identified a North Korean crew it tracks as Famous Chollima — described as a sub-unit operating under the Lazarus Group umbrella — as demonstrating "the most advanced AI usage" over the second half of 2025 and first half of 2026. According to the report, Famous Chollima created "entire fake companies with AI-generated websites, GitHub accounts, and email infrastructure to support insider threat operations." The campaign included a January–February supply-chain attack that targeted cryptocurrency and blockchain companies by publishing trojanized repositories, primarily hosted on GitHub, that contained legitimate-looking project files alongside hidden malicious scripts which executed automatically when developers opened the repositories.
Built by Nubivance.
OSINTSights' secure edge-first architecture, AI content pipeline, and serverless ops are designed by Nubivance. We do this for clients too.
Talk to us →Supply-chain poisoning and mass dependency compromises: Altered Spider and suspected Stardust Chollima activity
The report ties multiple supply-chain techniques to recent intrusions. CrowdStrike notes AI supply-chain compromise was the second most common MITRE ATLAS technique for initial access. A financially motivated crew tracked as Altered Spider (also known elsewhere as TeamPCP) targeted developers' AI tools, compromising more than 300 software dependencies in a single day, harvesting credentials and secrets, and then pivoting into cloud environments for theft and extortion. "It gives you a sense of how quickly they can move throughout that environment," Meyers said, adding that Altered Spider "hits the endpoint in seconds and within minutes, they're inside of the cloud."
Separately, threat hunters suspect another Lazarus offshoot — tracked as Stardust Chollima or Sapphire Sleet — was behind a March Axios supply-chain attack, and Amazon has attributed four npm compromises over the past 18 months to the same North Korean crew, the report states.
Patching windows slamming shut: PoCs weaponized in 24–48 hours and a surge in CVEs
CrowdStrike warns AI is accelerating exploitation of newly disclosed vulnerabilities. From January to June, 88 percent of exploitation observed by CrowdStrike using public proof‑of‑concept (PoC) code occurred within 48 hours of the code's release; China-linked groups such as Vault Panda and Genesis Panda moved even faster, launching attacks within 24 hours of disclosure. "Vulnerabilities are weaponized through the use of AI," Meyers said, arguing the "30-day patch window, which frankly, was aspirational, is completely obsolete. We're down to 24-hour, 48-hour patch cycles."
Meyers also highlighted the scale of the vulnerability stream: "In 2025, there were something like 48,200 CVEs that were registered," he said, and as of last week "we're already... at 43,000 for this year. We're not even into August yet, and we're already coming very close to the number from last year." June alone saw more than 7,600 software bugs reported and tracked through CVEs, the report notes.
What this means for technologists, developers, and cloud operators
- Technologists and security teams: Expect faster incident windows. CrowdStrike's data implies organisations face 24–48 hour patch cycles and a surge of AI-agent activity that produces 2.5x more leads than human-triggered incidents — a capacity and prioritization challenge for detection and response.
- Developers and open-source maintainers: Repositories and dependencies are active attack surfaces. The report recounts trojanized GitHub repos and mass dependency compromises, including more than 300 dependencies in a single day, underscoring the risk to CI/CD pipelines and developer workstations.
- Cloud operators and enterprises: Credential theft and cost-harvesting are real, immediate threats. CrowdStrike documented a token thief making roughly 200,000 API requests in two minutes and describes campaigns that harvest secrets, then pivot into cloud environments for theft and extortion.
CrowdStrike's account makes plain two tied trends: attackers increasingly use AI to amplify speed and scale, and attackers are turning AI systems themselves into high-value targets — via poisoned packages, trojanized repos, and CI/CD compromises. The firm's verdict is stark and specific: the traditional 30‑day patch cadence is untenable when proof‑of‑concepts and weaponization appear within a day or two, and when adversaries can weaponize both code and AI at machine speed. How organisations adapt operationally to those compressed windows — and whether defensive tooling and supply‑chain hygiene can keep pace — will determine how costly this next wave becomes.
