Company Blew $500M on Claude Because Nobody Set a Spending Limit - MRKT3.0
Skip to content
Ethical AI Policy & Analysis
Company Blew $500M on Claude Because Nobody Set a Spending Limit
Grace Sharp
June 2, 2026
Takeaways
An American company reportedly spent $500 million on Anthropic’s Claude AI in one month after failing to impose usage limits on employee licences.<br>Widespread adoption without oversight mirrored Uber’s experience, where AI tools exhausted annual budgets rapidly due to high productivity and token consumption.<br>This highlights risks of tokenmaxxing culture, prompting cost cuts and redundancies while underscoring the need for governance and spending controls.
One unlucky American company reportedly ran up $500 million on Claude in a single month. The only thing that went wrong was that nobody ticked the box marked “limit.”<br>A company in corporate America spent half a billion dollars on AI in a single month. Not through a rogue model, not through a security breach, not through anything that would make a headline about the dangers of AI. It spent it the way you rack up charges on a free trial you forgot to cancel: automatically, invisibly, and at a scale nobody was tracking. By the time anyone looked, the number had nine digits .<br>The tool was Anthropic’s Claude . The source, according to Axios , is an AI consultant who says one of their clients did exactly that – handed staff AI licences with no usage caps and then left them to it. No limits. No dashboard. No one watching the meter while employees ran coding marathons and agent workflows through the night. The whole thing came down to paperwork.<br>There is one important caveat. That figure comes from a single consultant’s account. No company has been named, no auditor has confirmed it, and it could easily be inflated, misremembered, or buffed up for the retelling.<br>That said, the number is plausible. And this is far from the only company facing the problem.<br>CEO waking up to $500M Claude API bill pic.twitter.com/Twcsgaq4oF<br>— Sahil (@sahill_og) June 1, 2026
Companies Are Learning the Hard Way About AI Costs<br>Uber is one of the clearest examples so far. Per reporting in Fortune and The Information , the company rolled out Claude Code to roughly 5,000 engineers last December. Adoption jumped from 32% to 84% in a matter of months. The tool worked extremely well. By April, just four months into the financial year, Uber had already spent its entire 2026 AI budget, with its heaviest users running up bills close to $2,000 a month each.<br>Uber’s CTO, Praveen Neppalli Naga , later said: “I’m back to the drawing board because the budget I thought I would need is blown away already.”<br>In other words, Uber was not burned because the tool failed. It was burned because the tool worked too well to afford.<br>Microsoft has reportedly cancelled most of its internal Claude Code licences, partly due to cost, quietly steering engineers toward its own cheaper alternative instead. So the mystery half billion dollar bill is not an outlier. It is the same story, just with a much larger price tag.<br>Welcome to Tokenmaxxing<br>This is where it stops being an accident and starts being a culture.<br>For two years, executives have been telling staff to use AI as much as possible . In some companies, that instruction was built directly into performance incentives. Meta reportedly folded AI usage into performance reviews. Amazon built an internal leaderboard ranking how much staff used AI tools, then quietly scrapped it after noticing employees were setting agents loose on pointless tasks purely to climb the rankings.<br>The internet has a name for this behaviour: tokenmaxxing . Burning the maximum number of tokens because the system rewards consumption rather than outcomes.<br>The use cases are often absurd. One CTO told Axios that staff were using premium AI models to check the weather, something your phone already does for free. Former Microsoft chief AI officer Sophia Velastegui put it more bluntly. She explained that people tend to use AI to automate the tasks they personally dislike, rather than the ones that actually create value. Handing out licences to everyone and hoping for the best rarely produces transformation. It usually just creates an expensive new way to avoid tasks employees find boring.<br>tokenmaxxing on the company card >>> https://t.co/3hYwki3pjo pic.twitter.com/JNlxsB68pS<br>— rohan (@rohandevs) June 1, 2026
Now apply that logic to the mystery company. Thousands of uncapped accounts. Gentle pressure from above to use AI more. People measured on how much they consume rather than what they produce. Suddenly the nine-figure bill stops looking like a freak accident, and it starts looking almost inevitable.<br>So, What Happens When the Bill Arrives?<br>When the bill lands, the answer depends on who you ask. And that is exactly the problem.<br>For many companies, the easiest lever is to cut costs elsewhere. One executive told Axios that for many firms, cutting staff may be the only realistic...