One firm burned $500M on Claude in a month—because nobody set usage limits

an unnamed company allegedly incurred $500 million in charges for a single month of anthropic's claude usage because administrators failed to set limits on employee prompts, as reported by axios. the incident underscores a growing enterprise crisis where flat-rate pricing models lead to massive overages when staff use expensive llms for trivial tasks like checking the weather. executives from microsoft and uber note that without specialized ai oversight roles, companies waste money on inappropriate models and suffer from quality issues like biased outputs. organizations must cultivate internal expertise to distinguish between tasks requiring generative ai and those better served by traditional software to ensure roi.