Cybercriminals Hijack LLMs, Causing Up to $100,000 Daily in Cloud Charges via 'LLMJacking'

Mandiant reports a surge in 'LLMJacking' incidents where cybercriminals hijack corporate access to Large Language Models (LLMs) at scale, leading to fraudulent cloud costs exceeding $100,000 per day for some victims. Attackers exploit unsecured endpoints and stolen cloud identities, with one incident involving over 500,000 LLM calls in under eight hours. This new threat highlights the rapid financial impact of AI-related cyberattacks.