Software company Atlassian is looking to better track employees’ AI spending by introducing a “wallet” with a monthly cap of $2,000 per employee, as costs explode at other tech companies.
Atlassian’s move, which recently cited AI as part of the reason for cutting 1,600 employees, bucks a trend in other tech industries to encourage employees to use technology as much as possible, a practice known as “tokenmaxxing.” Some companies are reportedly implementing leaderboards of the most AI-enabled employees in their operations.
A token is a measurement of the response that the AI gives to a prompt. According to OpenAI, one token is about four characters, and something like the U.S. Declaration of Independence is worth 1,695 tokens. OpenAI’s flagship model, GPT-5.6 Sol, costs US$5 per million tokens, while Anthropic’s Claude Fable and Mythos models cost US$10 per million tokens.
With tokenmaxxing, costs add up quickly.
Uber reportedly burned through its AI budget in four months, and Amazon reportedly told its employees to stop using AI just for the sake of using AI.
While Atlassian has never advocated token maxing or offered unlimited AI budgets, the Australian company this month introduced an “AI wallet” for staff in its research and development teams.
An internal memo obtained by Guardian Australia said employees keep between $500 and $2,000 a month in their wallets, which can be used across four AI products, including Claude Code. Employees receive notifications when they approach their wallet limit, and usage is suspended when they run out of money.
Employees can request additional funds. Atlassian is understood to have not turned down any requests for additional funding so far.
A company spokesperson said the company is transforming into an “AI-first company” by helping people build and experiment with technology.
“Atlassian provides significant funding to enable our builders to take advantage of multiple AI tools,” a spokesperson said.
“AI tools budgets are set by role based on how different teams work.”
They also said the wallet represents an increase in the amount of money available to employees.
A June PureProfile survey of 500 senior Australian staff from companies using AI, commissioned by search AI company Elastic, found that 80% were concerned that high usage was being mistaken for increased productivity.
We found that 32% reported pausing, canceling, or scaling back their AI deployments due to cost.
Jeremy Pell, ANZ manager at Elastic, said it was “sensible” to have a monthly cap on AI spend and more organizations should do this.
“Currently, only 9% of Australian organizations have restrictions on the use of tokens or APIs for AI agents or autonomous workflows, so those adopting this practice are outliers,” he said.
Arun Chandrasekaran, distinguished vice president analyst at research firm Gartner, said wallets are an easy way to “incentivize the right behavior” and prevent people from using AI effectively.
He said costs are exploding due to AI agents performing tasks autonomously on behalf of users, and these expenses are a key issue for businesses.
“All of a sudden, these systems are all trying to perform independent tasks, spawning small agents, creating their own prompts, and starting requests to the model,” he said.
“So, while the price of AI models has fallen over the past three years, the amount of tokens that AI agents in particular are starting to send to the models has increased significantly.”
Chandrasekaran said companies are looking for ways to reduce costs and are looking at using models that are less powerful for simple tasks, or using open weights, or open source models, where users can download the models and run them on their own systems.
