Atlassian Caps Staff AI Spending at $2,000 a Month While Other Tech Firms Cheer Unlimited Use

Key points

• Atlassian, the workplace software company behind tools like Jira and Confluence, has capped each employee's monthly AI spending at $2,000.
• The company introduced virtual "wallets" to track and limit how much individual staff members spend on AI tools.
• Atlassian cut 1,600 jobs in 2024, citing AI as part of the reason.
• Some tech companies have reportedly gone the other direction, encouraging unlimited AI use and even ranking employees by how much of it they consume.
• The trend of pushing employees to use AI as heavily as possible has been nicknamed "tokenmaxxing" inside the industry.

Atlassian has given every employee a digital spending account, capped at $2,000 per month, for AI tools. That figure covers the "tokens" (units of text or data that AI systems process, essentially the fuel that powers AI tools) the employee burns through in a given month. Once the wallet runs dry, that's the limit. The Guardian reported the move as a deliberate break from what other technology companies are doing.

We first reported on the tokenmaxxing trend on 17 July 2026 in "AI Token Spending Is About to Blindside Corporate Bosses", when a prominent tech investor warned that enormous AI bills were quietly accumulating before executives even noticed.

What is "tokenmaxxing" and why does it matter?

Tokenmaxxing means encouraging employees to use AI tools as heavily as possible, on the theory that heavier use produces more productivity. It's the opposite of rationing. Some companies have reportedly introduced leaderboards ranking staff by how many AI tokens they consume, turning heavy usage into a badge of honour.

The problem is cost. Every prompt sent to a large language model (the technology behind chatbots like ChatGPT and Claude) carries a small price tag. Multiply that across hundreds of employees running queries all day and the bills get very large very fast. As our 28 July 2026 story on Google's ballooning AI costs showed, even the biggest players are finding those numbers harder to absorb than they expected.

Atlassian has decided those bills need a ceiling.

What does this mean for workers?

$2,000 a month per person isn't a tight squeeze for most roles. Frequent everyday AI use typically costs well under $100 a month. The cap is more a guardrail against runaway experimentation than a hard restriction on normal work.

What has changed is visibility. Employees now have a budget line for AI, which means usage is no longer invisible to management. That alone is a shift worth watching closely.

The broader picture is harder to ignore. Atlassian laid off 1,600 staff and named AI as a contributing factor. Now it's also putting a price ceiling on the AI tools the remaining employees use. Both decisions point the same way: the company wants AI to cut costs, not add them. A researcher we covered on 24 July argued that making individuals faster with AI often does little for the organisation as a whole if team structures don't change, which makes Atlassian's cost-first framing look at least as plausible as the productivity-first bet.

Other tech firms have taken the opposite wager, betting that flooding employees with AI access will generate productivity gains that outrun the spending. The industry is quietly watching to see who's right.

Common questions

Could this kind of AI spending cap affect how people do their jobs?

For most employees, a $2,000 monthly cap leaves plenty of room for everyday AI tasks. It would mainly affect staff running large-scale or experimental AI workloads, who'd need to prioritise which tasks they hand to AI tools.

Why do AI tools cost money per use?

Most AI assistants run on large language models hosted by outside providers. Those providers charge a small fee for every query processed, similar to paying per text message. Heavy use across a big workforce adds up fast.

Is tracking employee AI use becoming normal?

Several large technology companies are now measuring how often staff use AI tools, whether to control costs or assess productivity.