An enterprise customer of Anthropic’s Claude accumulated a $500 million bill in a single month after failing to set any spending caps or usage limits on employee accounts. Engineers running long autonomous coding-agent sessions, large-context prompts, and unattended automated workflows consumed enormous volumes of tokens without any governance controls in place to catch the runaway usage before the bill arrived. The incident is cited as a stark example of enterprise AI cost-governance failure as organizations scale up agentic AI usage without corresponding spending guardrails.