AI Token Waste Hits Six Figures and Corporate Budgets

Some employees have turned AI token usage into an expensive joke. One Slash fintech worker blew through over $80,000 just to build a game called “brainrot shooter.”

Nicolas Brilliante, the culprit, shared a screenshot showing $81,267 spent on AI tokens. He called it a “genuine accident” and admitted he underestimated his own ability. Later, he wondered if he’d become a case study in AI spend gone wild.

Slash encouraged people to play the game with a cheeky nudge: “Pls play it so we can write this off as a marketing expense.” That’s one way to soften a six-figure blunder.

Slash isn’t alone. Uber burned through its entire annual AI budget by April. The company’s CTO confirmed widespread use of coding tools drained funds faster than expected.

Routine tasks are the real culprits. Converting PDFs to slides and other mundane jobs generate the largest token bills. Leaked Accenture audio revealed non-engineers drive this surge. Justice Kwak, Accenture’s head of AI strategy, said, “It’s actually not our engineers that are driving the token consumption.”

Everyday workers automating boring office chores rack up bigger AI bills than engineers pushing frontier tech. This trend forces companies to rethink AI spending.

Now, firms are implementing token governance, quotas, and chargeback models. The message is clear: AI spend must be treated like a budget, not an open tab.

Vendors selling AI tools face scrutiny over usage-based pricing. Organizations want to avoid surprises from runaway token counts.

Uber’s $32 billion market cap cost dwarfs its AI token budget, but even giants struggle to control AI expenses. Anthropic reportedly raised more than $32 billion in six months—more than Uber’s valuation—and companies still fret over token usage by ordinary employees.

AI’s promise is real. But left unchecked, token costs can spiral out of control on the most trivial tasks.