Tokenmaxxing Is Blowing Corporate AI Budgets
Uber Technologies burned through its entire 2026 AI budget by April, the clearest early casualty of a workplace trend called tokenmaxxing.
Tokenmaxxing means ramping up AI tool consumption as a performance signal. It spread through tech companies as employees used high usage rates to demonstrate productivity to management, turning token consumption into an informal KPI.
Every AI request burns digital tokens, the billing unit for large language models, and the bill only arrives after the task is complete, per MarketWatch.
Agentic AI compounds the problem by autonomously handling tasks like booking appointments or writing code, spinning up dozens of simultaneous processes, each racking up charges.
Token prices rose roughly 60% since late February, driven by a surge in agentic usage alongside a shortage of computing memory and storage.
"All the costs are really starting to skyrocket," said Mark Barton of tech consultancy Omniux.
Productivity Has Not Kept Up
Uber's operations chief said the spending has produced no noticeable improvement in consumer features or productivity, per the same MarketWatch report.
Meta's CTO Andrew Bosworth pushed back internally, writing that "nobody should be using AI tools just for the sake of using them," according to The Wall Street Journal as cited by AFP.
Amazon.com tried measuring AI engagement with a company-wide leaderboard, but workers began inflating their scores, making the data unreliable.
Nvidia VP Bryan Catanzaro told Axios that his team's AI costs had exceeded human labor costs for months, according to Forbes.
Microsoft recently revoked Claude Code access from developers across Windows, Microsoft 365, and Surface teams, redirecting them to its own GitHub Copilot CLI tool, with sources telling The Verge the move was partly financial.
Smarter Spending Is Replacing the Binge
Some companies are switching to open-source or specialized models to contain costs, with budget alternatives priced as low as $0.05 per million tokens versus $15 for flagship models.
Global tech companies have announced $740B in capital expenditures this year, a 69% jump over 2025, according to Morgan Stanley as cited by Forbes.
Roughly 117K workers have been laid off at 164 tech companies so far this year, per Layoffs.fyi, but the tokenmaxxing backlash is making human labor look like the more predictable budget line.
