Skip to content
Tech News
← Back to articles

Microsoft Tokenmaxxing King Spent $28,000 on AI in a Single Month as Company Begs Employees to Tone Down the Ludicrous Costs

read original more articles
Why This Matters

Microsoft's internal AI spending reveals a significant outlier, with one employee spending $28,000 in just a month, prompting the company to implement stricter controls and shift to more efficient AI models. This highlights the growing challenge for tech giants to balance innovation with cost management and responsible AI usage.

Key Takeaways

Sign up to see the future, today Can’t-miss innovations from the bleeding edge of science and tech Email address Sign Up Thank you!

Microsoft employees are supposed to be using AI as much as possible. But wait, not like that!

In internal tracker, one employee who worked in the Customer and Partner Solutions organization stood out from all the rest by admitting they spent an eyebrow-raising $28,000 on AI tools in just 28 days, Gadget Review noticed.

That sky high figure came from an internal employee compensation spreadsheet first leaked by Business Insider, which in 2026 gained a new column titled “AI $ Usage Per Month.” Only around 350 employees submitted data, but it showed that the company-wide median was roughly $300 per 28-day window, making our “tokenmaxxing” king a massive outlier.

Unsurprisingly, token usage varied wildly, according to Gadget Review, with a few other employees surging past the $10,000 mark, while others, sometimes in the same department, spent just tens of dollars.

It’s perhaps no coincidence, then, that Microsoft leadership recently began a crack down on this kind of spending.

In early August, Jay Parikh, the executive vice president of Microsoft’s CoreAI division, sent employees a memo telling them to cool it on the rampant AI usage, 404 Media reported.

“Tokenmaxxing is not what we are optimizing for,” Parikh wrote. “I want all of us focused on maximizing outcomes that move the needle for our customers and our business.”

“As such, we are updating our internal guidance and managing token spend with the same discipline we apply to every other critical resource,” he added.

On top of keeping closer tabs on how its employees used AI, Parikh also said that the company would switch to OpenAI’s GPT-5.6 Sol — a less token-hungry model — as its default for internal use, according to 404.

... continue reading