PoeLLM malware turns exposed AI servers into cryptomining botnet
· Cybersecurity
Researchers are warning about a financially motivated malware family dubbed PoeLLM that targets exposed AI and large language model infrastructure to deploy cryptocurrency miners and expand a botnet. The campaign has infected more than 3,400 servers, according to the reports, and turns compromised machines into scanners and exploit launchpads for further attacks. The targeting is notable: as organizations rush LLM and GPU infrastructure into production, misconfigured or unauthenticated AI services become high-value, high-bandwidth assets for attackers. The incident fits a broader pattern of criminals opportunistically harvesting compute from whatever new infrastructure class is widely deployed and poorly secured.
Why it matters:It marks AI and LLM servers as a new, attractive target class for financially motivated attackers, signaling that compute infrastructure security now directly affects AI operations.
Key points
- The campaign has infected more than 3,400 servers, researchers say.
- Compromised hosts are used as scanners and exploit launchpads to grow the botnet.
- The malware is financially motivated and deploys cryptocurrency miners.
Sources (original reporting)
More stories from this day
- Attackers hijack ccTLD registries, obtain Google domain certificates
- FBI warns FortiBleed attacks still lock out FortiGate admins
- Anthropic opens reduced-guardrail Claude access to vetted cyber teams
- Atlassian critical flaw exploited within hours of public disclosure
- Nvidia RTX Spark Laptops Launch, Surface Laptop Ultra From $2,599
- Samsung posts record quarterly profit on AI memory chip demand