Apple Cuts Employees' Daily AI Budgets as Ternus Pushes to Rein In Costs

Last month Apple lowered the daily AI spending cap for some engineers to $200 from $300, according to the report, and it is considering further cuts that could bring some engineering teams down to $100 a day. Many non-engineering groups already have a $100 daily limit, and Apple has considered capping some teams at $25. Staff who need more can ask senior leaders for an exemption.

The budgets cover tokens, the units of text AI models process and charge for. Engineers get the largest allowances because AI coding and debugging use a lot of tokens. Apple staff can use Anthropic's Claude, OpenAI's ChatGPT and Google's Gemini, among other models, and some teams are being told to switch to cheaper models or to models that run on the device, according to the report.

The savings may be small. Some employees told Bloomberg that few people used their full $300 daily allowance. Apple declined to comment.

The limits are part of a wider push on costs under John Ternus, who became chief executive on September 1, 2026. Ternus has pushed the company to move faster and run leaner and has removed layers of management, according to reports. Some divisions have also hired outside consultants to review how efficiently they work. Together with the AI caps, those reviews have left some staff worried about a restructuring. Apple cut more than 200 jobs in August. No new layoffs have been reported.

Overall spending is still rising. In its annual report, filed this week, Apple said research and development costs rose 32 per cent in fiscal 2026, driven by higher infrastructure and AI costs. Operating expenses rose 22 per cent to $75.5 billion.

Apple is not the only company limiting staff AI use. Microsoft introduced AI budgets at division level in July. Tesla caps staff AI spending at $200 a week, Uber limits each AI coding tool to $1,500 a month per user, and Accenture caps Claude use at $29 a month, according to published reports.

The question for AI labs is what these caps mean for demand. Enterprise use of coding agents has driven much of the recent revenue growth at Anthropic and OpenAI. If more large customers start rationing tokens per employee, labs could find it harder to grow revenue simply from heavier use.