US Army Burns Through Annual AI Token Allocation in Weeks
After promising unlimited AI tokens, the US Army was forced to restrict usage of the Ask Sage platform as the annual token pool was exhausted within weeks, raising concerns about the sustainability of AI adoption.
Just over a month after the Department of Defense (DOD) boasted that nearly half of its 3.5 million employees were using AI at work, members of the Army's Combat Capabilities Development Command (DEVCOM) received an email informing them that they were burning through tokens and needed to limit use.
The email stated that although the Army CIO announced in May 2026 that unlimited tokens would be offered, by mid-June the Army CIO pool was exhausted and limits had to be re-established. The email added that while the Army had chosen to renew token usage at its current levels, it was unclear if the pool would be renewed after October 1.
The Army uses Ask Sage, a multimodal generative AI platform that allows users to run various large language models (LLMs), including Alphabet's Gemini, Meta's Llama, and OpenAI's ChatGPT. An anonymous Army employee told WIRED, "Apparently the whole Army burned through the whole year of tokens for just one service."
Ask Sage is used by the Army to power its enterprise LLM workspace and is accredited for Controlled Unclassified Information. It is also used by the DOD's Chief Digital and AI Office (CDAO) for acquisitions. According to the Army's website, Ask Sage was used for tasks such as reclassifying personnel descriptions.
Employees were given at least 200,000 tokens per month, with automatic allocations if they used up their initial allotment. Those who had signed up but were not regularly using Ask Sage received emails encouraging them to use more tokens.
The Army had access to 100,000,000 tokens as part of an annual subscription to an "enterprise pack." One token equates to about 3.7 characters. According to Breaking Defense, the DOD burned through some 20 billion tokens per day during the 38-day Operation Epic Fury in Iran.
It is unclear if the tokens used by regular DOD employees are drawn from the same pool as those used for classified information. This has not stopped the DOD's emphasis on AI. On Monday, The Intercept reported that the Pentagon cut staff at the Civilian Protection Center of Excellence, whose jobs involved preventing civilian casualties, and is developing an AI tool to speed up assessments instead.
The Army is not the first eager adopter of generative AI to rethink unlimited use. Meta encouraged employees to "tokenmaxx" but later quietly removed its token usage leaderboard and is now trying to curb use. Uber also saw its engineers burn through a year's worth of tokens in four months.
The Army employee noted that the generative AI tools have not been particularly useful for their work and have often been unreliable. One model even claimed to have completed a task it had not. The employee believes that while the tools might help with certain bureaucratic aspects of the US federal government, unthinking application will not lead to an effective, efficient, and trustworthy rollout.
