Google’s Mandiant business warned that cybercriminals are selling access to large language models (LLMs) in a growing enterprise threat it calls “LLMJacking,” and said it had responded to multiple incidents so far this year across financial services, retail, technology, and transportation sectors. Mandiant said that unlike the gradual resource consumption associated with cryptojacking, LLMJacking can generate large spikes in fraudulent cloud charges in a matter of hours, and that across individual incidents it observed fraudulent charges associated with resource consumption exceeding $100,000 in a single day, with cumulative charges at some organizations reaching $1m before access was revoked. Threat actors bypass front-end chatbots to target backend model resources, usually via LLM API keys or exposed or stolen cloud identity tokens, including IAM credentials, Service Account keys, GitHub Personal Access Tokens and active session tokens, then escalate privileges and carry out internal reconnaissance to identify targets and more cloud credentials; they might also target vulnerabilities in LLM routing software to escalate privileges. Google said the ultimate objective is monetization and resource-hijacking of foundation models through token proxying, automated model streaming and GPU hardware provisioning, and it has seen financially motivated clusters across North America, South America, and Asia Pacific, plus a “division of labor” with “LLM Mixers” that proxy, aggregate and resell access.