Anonymous enterprise spent $500 million on Anthropic's Claude AI in one month due to unlimited employee access without usage caps

An unnamed company burned through $500 million in a single month on Claude after deploying it to thousands of employees without token usage limits, according to Axios reporting. Agentic AI tools (AI systems that autonomously complete multi-step tasks) can consume up to 1000x more tokens than basic chat queries, making costs spiral when unchecked.

Microsoft recently canceled most internal Claude Code licenses as part of what is described as 'the clearest enterprise-scale AI spending pullback so far in 2026.' Corporate leaders are questioning whether high AI spending delivers meaningful returns, contrasting with cases like Google Cloud customers facing $18,000 surprise bills and the OpenClaw project burning $1.3 million in OpenAI tokens monthly.

Employees at some organizations have engaged in 'tokenmaxxing'—maximizing AI usage to hit internal leaderboards rather than create business value. Amazon scrapped its AI usage tracking system after discovering workers were inflating consumption through pointless queries, including using advanced AI to check the weather; Uber's CEO noted there is no clear connection between extreme token consumption and shipping useful products.