Firms introducing new AI controls amid 'tokenmaxxing' concerns
Atlassian has introduced new "AI wallets" with monthly spending caps as costs rise in other firms as a result of employees' surging AI use.
The Australian software firm introduced the wallets for employees in its research and development team, with a monthly budget between $500 and $2,000, The Guardian reported.
Employees can use the budget across four AI products, and are notified if they reach the limit of their wallets.
Usage stops when the budget runs out, according to the report, but employees can request additional funds.
A spokesperson for Atlassian told The Guardian that the wallet represented a boost in the amount employees could spend. It comes as the company seeks to redefine itself as an "AI-first" firm.
"Atlassian provides a significant budget for our builders to leverage multiple AI tools," the spokesperson told the news outlet. "AI tooling budgets are set by role based on how different teams work."
Rise of 'tokenmaxxing'
The new AI wallet comes as other firms across the world seek to regulate AI use among employees amid the rise of "tokenmaxxing."
Job-listing platform Built In defines "tokenmaxxing" as the practice of maximising AI usage, specifically by consuming as many tokens as possible with autonomous agents.
Tokens are the building blocks of text that AI models process, according to OpenAI. One token is equivalent to four characters, with a single paragraph taking up around 100 tokens.
OpenAI's GPT-5.6 Sol charges US$5 for every one million input tokens, The Guardian reported, while Anthropic's Claude Fable 5 and Claude Mythos 5 models are US$10 for every one million input tokens.
The widespread adoption of AI has sent token costs soaring across many firms, spurring a wave of restrictions on employee usage.
Amazon recently shut down an internal AI leaderboard that scored staff based on their activity on the AI platform Kiro, following reports that employees were tokenmaxxing.
Meta also recently warned employees that token budgets, allocations, and controls were underway as internal AI use in 2026 alone could reach billions.
