# Atlassian Caps Employee AI Spend at $2,000 Monthly Amid Tokenmaxxing Trend > Source: > Published: 2026-07-29 17:23:47+00:00 **July 29, 2026**, (Inside AI) — Software firm **Atlassian** has begun capping employee spending on AI tools through a new “wallet” system, setting monthly limits between **$500** and **$2,000** per person. The move, detailed in an internal memo, affects the research and development team and covers usage across four AI products, including **Claude Code**. Employees receive alerts as they near their cap, and spending is halted when the wallet empties. Requests for additional funds are allowed, and so far, none have been denied. The policy marks a sharp departure from the “tokenmaxxing” trend at other tech firms, where staff are encouraged to maximize AI usage, sometimes with internal leaderboards tracking the most prolific users. Atlassian’s pivot comes as AI costs soar across the industry. **Uber** reportedly exhausted its AI budget in just four months, and **Amazon** has told employees to stop using AI gratuitously. The company, which recently cut **1,600** jobs partly due to AI, says the wallet system is about enabling smarter experimentation, not restriction. **“Atlassian provides a significant budget for our builders to leverage multiple AI tools,”** a company spokesperson said. **“AI tooling budgets are set by role based on how different teams work.”** They added that the wallet actually increases spending power for some roles. The move underscores a growing tension: while AI model prices have fallen for three years, the volume of tokens consumed by autonomous agents is exploding. **OpenAI** charges **$5** per million tokens for its **GPT-5.6 Sol** model, and **Anthropic** charges **$10** per million for **Claude Fable** and **Mythos**. Under tokenmaxxing, these costs multiply fast. According to a **June PureProfile** survey of **500** senior Australian staff at AI-using companies, commissioned by search AI firm **Elastic**, **80%** worried that high usage was mistaken for productivity. **32%** had paused or scaled back AI deployments due to cost. **Jeremy Pell**, Elastic’s ANZ manager, called Atlassian’s cap “smart” and noted that only **9%** of Australian organizations limit token consumption for AI agents. **“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 VP analyst at **Gartner**, said wallets can “incentivise the right behaviour” and curb ineffective AI use. He explained that cost explosions are driven by AI agents that spawn sub-agents, creating cascading token consumption. Companies are now exploring cheaper models for simple tasks and open-weight alternatives to rein in expenses. ## Agentic Sprawl Drives Unchecked Spending Chandrasekaran elaborated on the mechanics: **“You suddenly have these systems that are all trying to do independent tasks that are spawning smaller agents, that are creating their own prompts and initiating requests for the model,”** he said. **“So while the 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.”** This agentic sprawl is reshaping enterprise AI economics. A [recent paper on arXiv](https://arxiv.org/abs/2501.05409) documents how multi-agent systems can balloon token usage by an order of magnitude compared to single-model calls, reinforcing the need for governance. Atlassian’s wallet approach mirrors early experiments in cloud cost controls, but applied to generative AI. ## From Tokenmaxxing to Fiscal Discipline The term tokenmaxxing emerged as a playful metric of AI enthusiasm, but its financial toll is now clear. [Gartner research](https://www.gartner.com/en/documents/6015767) indicates that by **2028**, more than half of enterprises will implement similar AI consumption limits. Atlassian’s move may signal the end of the AI spending honeymoon, as firms balance innovation with accountability.