$81,000 on a meme shooter: fintech startup manager burned a week's budget on AI tokens chasing code
Fintech startup Slash, recently valued at $1.4 billion, received an unexpected lesson in financial discipline in the era of artificial intelligence. One of the company's executives, Nicolas Briante, spent over $81,000 on tokens for the Claude AI model in just one week. The reason for this extravagant expense was the development of a meme game called Brainrot Shooter.
The incident occurred after Slash's management urged employees to make more active use of AI tools for writing code. Briante took this call literally and devoted an entire day to creating a shooter featuring characters like Skibidi Toilet and Tung Tung Tung Sahur. However, he underestimated how quickly tokens are consumed when repeatedly loading the context of the entire codebase. Each request to the model cost money, and by the end of the day, the bill had grown to tens of thousands of dollars.
From Failure to Strategy
The company itself responded to the situation with humor, suggesting employees "play the game so it can be written off as marketing expenses." However, the story took an unexpected turn. In the first 48 hours, 6,912 people played Brainrot Shooter, with a total playtime of 8,986 hours and an average session of 1.3 hours. Peak online reached 437 concurrent players, and the company received three requests for ad placements. The finance department promptly reclassified the project from "expense incident" to "strategic initiative."
This case is not isolated. Earlier this year, Uber exhausted its annual AI budget in just four months, after which it imposed strict limits. And one large company using Anthropic's Claude received a $500 million bill for a single month, simply because it did not set token consumption limits for employees.
Expert opinion: The Slash story is a clear signal for the entire market. We are entering an era where "free" AI access for employees can turn into a financial disaster if API-level costs are not controlled. Companies urgently need to implement monitoring and usage-limiting systems for AI, otherwise "accidental" $80,000 bills will become the new norm.