Apple Reportedly Tightens AI Token Budgets for Employees

By Priya Raman

Apple reportedly reduced the daily AI spending allowance for some engineers from $300 to $200 in October, according to a Bloomberg report published Wednesday. The $100 reduction equals one-third of the previous limit, and Apple is considering further cuts that could lower selected engineering teams’ daily caps to $100.

The company also is considering lowering a $100 daily cap for some non-engineering groups to as little as $25. Employees needing broader access can request exemptions from senior executives, making senior approval part of the process for exceeding a standard allowance.

These limits cover AI tokens, the units of data processed by artificial intelligence models. OpenAI’s ChatGPT, Anthropic’s Claude and Google’s Gemini use tokens to process prompts and generate responses, so increased consumption can become a significant operating cost. Apple employees use the tools for research, coding and other workplace tasks, making consumption a direct budget issue.

Apple’s caps sit within a wider change across technology companies. Firms initially encouraged staff to experiment with AI models and coding assistants, then began asking whether broad access created costs larger than the productivity returned.

What does Apple’s spending shift change?

Apple has been working to embed generative AI into its engineering processes, including efforts to use the technology in chip development. The reported caps mark a shift from encouraging employees to experiment with models and coding assistants toward managing how extensively teams use those services.

The constraints may not apply uniformly across Apple’s workforce. Employees who need greater access can seek an exemption, which places senior executives between a standard allowance and higher use.

For the affected non-engineering groups, $25 would represent a 75% reduction from the current $100 cap. That would make routine experimentation subject to a much tighter budget and make the exemption route more consequential.

The move does not necessarily signal a retreat from AI. It reflects a narrower operating question: how much access employees need, and whether productivity gains justify the expense.

Why are employers tightening AI budgets?

Bloomberg described an earlier-2026 wave as “tokenmaxxing,” encouraging employees to use AI tools as extensively as possible in pursuit of productivity improvements. As token-related bills increased, companies began reconsidering whether unrestricted access was financially sustainable.

The issue is the gap between AI infrastructure costs and measurable productivity improvements. AI tools can help employees complete some tasks more quickly, but companies are still assessing whether those efficiency gains cover the costs of developing, deploying and operating the technology.

That changes adoption decisions from a simple expansion target into a cost-and-value test. Employers must decide where AI produces measurable value, how heavily each team should use it and when higher consumption warrants an exception.

How should employers evaluate access and value?

For Apple, proposed reductions may make employees more selective about AI services, particularly when work involves expensive models. Staff can ask senior executives for an exception when a task requires usage beyond the standard cap.

The practical adoption questions are specific: Which tasks finish faster with AI? Which models justify their token cost? Which teams can show a need for more than the standard allowance? How do those gains compare with infrastructure, deployment and operating expenses?

Apple is not alone in revising internal AI policy. Bloomberg recently reported that Goldman Sachs restricted access to some AI coding tools after a review of productivity and risk management. That move puts cost beside two other tests: whether the tools deliver practical benefits and whether their use creates risks that require tighter controls.

The report gives no Apple-wide AI bill, productivity gain or count of affected employees. It supports a change in spending rules, not a conclusion that Apple’s AI investments have become more or less productive. Apple’s pending cuts will test whether controlled token consumption preserves measurable productivity while exceptions preserve higher-cost use for priority teams.