Stolen AI access leaves companies with cloud bills of up to $1 million
Our summary
Google's Threat Intelligence Group and Mandiant have uncovered a wave of attacks dubbed LLMJacking, where criminals hijack compromised cloud accounts to run AI models at the victim's expense, with some incidents generating over $100,000 in charges within a single day and total bills reaching as high as $1 million. Throughout 2026, Mandiant handled multiple such cases across financial services, retail, technology and transportation firms in North America, South America and Asia Pacific, with intruders gaining entry through leaked service account keys, GitHub tokens or active session tokens. Underground 'LLM Mixer' resale services such as Jet Night and newapi.pro buy and resell this stolen access to subscribers at a fraction of legitimate pricing, funding their operations entirely through the stolen credentials. In one notable intrusion, attackers abused a vulnerability tracked as CVE-2026-40217 in a LiteLLM gateway to escalate regular API keys into full administrator credentials, while others quietly re-enabled disabled AI services to start billing victims within minutes.
Help Net Security publishes this story on its own site; we link to it rather than reprint it.
Worried this affects your company?
Discuss a security assessment of your internet-facing systems. Scope agreed before testing.