Software giant Atlassian has introduced "AI wallets" with monthly spending caps for its staff, becoming one of the first major technology companies to formally ration how much employees can spend on artificial intelligence tools. The move, reported by Guardian Australia on July 30, 2026, comes as runaway AI costs force a reckoning across the corporate world. For the latest AI business reporting and analysis, visit our breaking AI news.
How the AI Wallets Work
Under the new system, Atlassian employees in its research and development division receive a monthly allowance of between $500 and $2,000 that can be spent across four AI products, including Anthropic's Claude Code. According to an internal memo seen by Guardian Australia, workers receive notifications as they approach their spending limits, and usage is paused entirely once the money runs out.
Employees can request additional funds if they exhaust their wallet — and it is understood that Atlassian has not turned down any such request so far. A company spokesperson said the wallet system actually represented a boost in the amount employees could spend.
"Atlassian provides a significant budget for our builders to leverage multiple AI tools," the spokesperson said. "AI tooling budgets are set by role based on how different teams work."
The company, which recently cut 1,600 staff and cited AI as part of the reason, said it is transforming into an "AI-first company" by supporting employees who want to build and experiment with the technology.
Bucking the 'Tokenmaxxing' Trend
Atlassian's caps stand in stark contrast to a fad that swept through parts of the tech sector earlier this year dubbed "tokenmaxxing" — the practice of maximizing AI token consumption to squeeze as much AI-generated work as possible out of tools like ChatGPT and Claude. Some companies reportedly went so far as to introduce leaderboards ranking employees by how much AI they used in their work.
Tokens are the unit of measurement for how AI models process and generate text. OpenAI has said that a single token equals roughly four characters, and that the entire U.S. Declaration of Independence amounts to about 1,695 tokens. Pricing adds up quickly: OpenAI's flagship GPT-5.6 Sol model charges roughly $5 for every one million tokens, while Anthropic's Claude Fable and Mythos models charge around $10 for every one million tokens.
At that scale, heavy daily usage across thousands of employees can translate into staggering monthly bills.
The Cost Backlash Spreads
Atlassian is not alone in hitting the brakes. Ride-hailing company Uber reportedly blew through its entire AI budget in just four months, according to the Guardian's reporting. Amazon, meanwhile, has told employees to stop using AI "just for the sake of using AI" — a direct rebuke to the tokenmaxxing mindset.
The pullback reflects a growing realization across corporate boardrooms that enthusiasm for AI has outpaced measurable returns. A June 2026 survey by PureProfile, conducted on behalf of search company Elastic, polled 500 senior employees at Australian companies using AI. It found that 80% were concerned that high AI usage was being mistaken for genuine productivity gains. A further 32% reported that they had paused, cancelled, or wound back AI deployments due to cost.
Elastic's Australia and New Zealand manager, Jeremy Pell, said a monthly cap on AI spending was "smart" and argued that more organizations should adopt similar measures.
From Hype to Discipline
The shift from unbridled AI spending to budget discipline marks a turning point for enterprise AI adoption. In early 2026, the prevailing wisdom in much of the tech sector was that companies should use AI as aggressively as possible — the logic being that early experimentation would lock in a competitive advantage. Tokenmaxxing was celebrated as a badge of innovation.
By mid-summer, however, the bills began arriving. Companies that had encouraged employees to run AI on every task discovered that the costs scaled faster than the benefits. The result is a more measured approach in which AI is treated like any other resource: valuable, but subject to budget constraints and return-on-investment scrutiny.
Atlassian's wallet model offers one template. Rather than imposing a blanket ban or an unlimited free-for-all, it gives employees a defined budget and the ability to request more — putting the onus on workers to justify additional spending while still encouraging experimentation.
What It Means for the Industry
The episode carries lessons well beyond Atlassian. As AI tools become embedded in everyday workflows, companies face a delicate balancing act: encourage adoption and innovation, but avoid the trap of equating raw usage with productivity. The Elastic survey's finding that four in five respondents worry about confusing AI activity with actual results suggests this concern is now widespread.
For AI providers, the cost-consciousness shift is a double-edged sword. On one hand, enterprise demand remains strong — Atlassian, Uber, and Amazon are all still investing heavily in AI. On the other, customers are scrutinizing every token, which could pressure providers on pricing and force them to demonstrate clearer value.
The companies that emerge strongest may be those that, like Atlassian, build cost awareness directly into their workflows — turning the question from "how much AI can we use?" into "what is this AI actually worth?"
Stay Ahead of AI
The economics of artificial intelligence are reshaping how businesses adopt the technology. Read more AI news and analysis to keep up with the trends, deals, and strategy shifts defining the enterprise AI landscape.
