- The army previously encouraged workers to get behind the AI and offer them unlimited tokens
- Limits have been reinstated by the long-term future is less certain
- The Iran operation saw token consumption skyrocket
The US Army has reportedly been forced to reimpose limits on GenAI use after workers quickly used up the full authorization of tokens early, revealing that even one of the most well-funded agencies in the world can’t keep up with AI’s unpredictable and rising costs.
In May 2026, it was revealed that the army would provide users with unlimited tokens, but by mid-June this token pool had already run out.
Now an internal email, confirmed by The cableconfirmed that these limits have had to be reintroduced as a result, with usage renewed at current levels for the time being.
US Army struggles to keep pace with AI token consumption
However, the availability of tokens beyond October 2026 remains uncertain, and the Army is likely concerned about rising costs.
“Apparently the entire army burned through the entire year with tokens for just one service,” said an unnamed employee. The cable.
As for the complications, The cable reports that an annual enterprise package for Ask Sage, the Army’s AI platform of choice, contained 100 million tokens—enough for about 200,000 tokens per employee per month. However, users who used up their initial allocations were automatically allocated more, making the cap virtually meaningless.
To put volumes into perspective, the Department of Defense reportedly spent 20 billion AI tokens per day during its 38-day Operation Epic Fury in Iran.
More broadly, workers have been actively encouraged to use generative AI across the breadth of their roles, so the Army could figure out exactly where the technology could seriously improve productivity.
Aside from the military, this particular case is the perfect example of how aggressive AI pushes may not always take cost into account, and with many AI vendors now shifting to unpredictable consumption-based or output-based models over flat per-cap subscriptions. seat, it becomes increasingly difficult to predict and allocate budgets.
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews and opinions in your feeds.
