Atlassian tightens tracking of staff AI use as other technology firms encourage ‘tokenmaxxing’
Atlassian has introduced monthly "AI wallets" capping how much each employee can spend on AI tools, in a move against the industry trend of "tokenmaxxing" – encouraging staff to use as much AI as possible, with some firms reportedly running leaderboards to reward heavy users. The shift comes as Atlassian recently cited AI as a factor in cutting 1,600 jobs, and highlights growing concern across the tech sector that unchecked AI spending is spiralling without corresponding productivity gains.
Under the scheme, research and development staff receive between $500 and $2,000 a month to spend across four AI products, including Claude Code, with usage paused once funds run out, though staff can request more and Atlassian is understood never to have refused such a request. By contrast, Uber reportedly exhausted its AI budget in just four months, while Amazon has told employees to stop using AI purely for its own sake. A June survey of 500 senior Australian staff found 80% feared high AI usage was being mistaken for productivity, and 32% had scaled back AI deployments due to cost, with analysts noting only 9% of Australian organisations currently impose any such spending limits.
In short
- Atlassian caps staff AI spending via new monthly "wallets"
- Move contrasts with rivals' unlimited "tokenmaxxing" AI use
- Surveys show growing cost and productivity concerns over AI
Steelman: both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Monthly AI budgets can make costs visible and encourage staff to choose tools where they add practical value, especially when heavy use has not yet produced clear productivity gains. Allowing employees to request more funding preserves flexibility while giving managers a way to assess whether spending is worthwhile.
The case against
Caps may discourage experimentation and slow useful work when staff reach their allowance, particularly if the most valuable applications are difficult to predict in advance. Heavy usage can also help employees build skills and discover productive uses, so limits risk treating a temporary rise in costs as evidence that AI is not delivering benefits.
Explained simply
For the past year or two, many technology firms have been encouraging their employees to use artificial intelligence tools extensively. Some companies have created competitions or leaderboards to reward staff who use AI most heavily, reflecting optimism that increased AI usage would boost productivity and innovation.
However, there is growing doubt about whether this approach is working. Surveys of workers and business leaders have found widespread concern about whether heavy AI usage actually translates to productivity improvements, and spending on these tools is mounting without clear evidence of corresponding benefits.
Growing scepticism is changing how companies approach AI in the workplace. The question of whether encouraging unlimited AI use actually improves business performance has become a significant issue across the technology sector, particularly as firms grapple with rising costs and the need to demonstrate real value from their investments.