Software firm Atlassian has introduced monthly AI spending caps of up to $2,000 per employee in its research and development team, tightening tracking of AI costs amid an explosion in spending across the tech sector. The move, detailed in an internal memo seen by Guardian Australia, bucks the trend of other companies encouraging unlimited AI use, a practice dubbed "tokenmaxxing."
AI wallets and spending limits
Employees in Atlassian's R&D division now receive between $500 and $2,000 each month in an "AI wallet," usable across four AI products including Claude Code. Notifications alert staff as they near their limit, and usage pauses when funds run out. Employees can request additional funds, and Atlassian has reportedly never refused such a request so far.
Atlassian recently cited AI as part of the reason behind cutting 1,600 staff. A company spokesperson said the firm is "transforming into an AI-first company" by supporting building and experimentation with AI. "Atlassian provides a significant budget for our builders to leverage multiple AI tools," the spokesperson added. "AI tooling budgets are set by role based on how different teams work." The wallet represents a boost in the amount employees could spend.
Tokenmaxxing and cost explosion
Other tech companies have encouraged employees to use as much AI as possible, with some reportedly introducing leaderboards for the most AI usage. Tokens measure AI responses: OpenAI says one token equals about four characters, and the US Declaration of Independence amounts to 1,695 tokens. OpenAI's GPT-5.6 Sol charges $5 per 1 million tokens, while Anthropic's Claude Fable and Mythos models cost $10 per 1 million tokens. Under tokenmaxxing, costs add up quickly.
Uber reportedly blew through its AI budget in four months, and Amazon has told employees to stop using AI just for the sake of it. A June PureProfile survey of 500 senior Australian staff at companies using AI, conducted for search AI firm Elastic, found 80% were concerned that high usage was mistaken for productivity gains. 32% had reported pausing, cancelling, or winding back AI deployments due to cost.
Industry perspective on AI cost management
Elastic's ANZ manager, Jeremy Pell, called monthly AI caps "smart" and said more organisations should adopt them. "Right now, only 9% of Australian organisations currently have any limits on token or API consumption when it comes to AI agents or autonomous workflows, so any organisation that adopts this practice is an outlier," he said.
Arun Chandrasekaran, a distinguished vice-president analyst at Gartner, said wallets are a simple way to "incentivise the right behaviour" and stop ineffective AI use. He noted cost explosions are driven by AI agents that autonomously undertake tasks, spawning smaller agents and creating prompts. "While AI model prices have been falling for the last three years, the volume of tokens that particularly the AI agents are starting to send to the models is significantly increasing," he said. Companies are exploring ways to cut costs, including using less powerful models for simpler tasks and open-weight or open-source models.
