Apple reportedly trims internal AI budgets, cutting engineering caps from $300 to $200 a day

Apple is reducing how much its engineering teams can spend on AI each day, according to a Bloomberg report published Wednesday and relayed by TechCrunch. For a company that has spent the past two years wiring AI into its engineering processes, it is a notable change in direction. Apple did not respond to a request for comment.

The reported numbers are specific. In October, Apple lowered the daily AI spending cap for engineering teams from $300 to $200. More cuts are in the works, with some allowances expected to fall to $100 a day, and outside engineering the company is reportedly weighing a daily limit as low as $25.

Those caps govern the token costs engineers generate while using AI models, coding assistants and internal tools. The mood earlier this year was the opposite: a wave of “tokenmaxxing,” in which companies pushed staff to use as much AI as possible and treated consumption as a proxy for efficiency.

Then the bills arrived. Once model usage becomes a recurring expense, the calculation changes, because every autocomplete carries a price while the productivity gain it produces is hard to isolate. A hard daily ceiling turns a vague argument about return on investment into a number a manager can budget against.

It is worth separating internal frugality from strategy. The report describes a squeeze on internal usage budgets, not a change to Apple's product roadmap or a retreat from shipping AI features. The company has been beta testing AI capabilities on Apple Watch for rewinding and recapping conversations. Spending less on internal tooling and continuing to ship consumer AI are different decisions.

Because the reporting rests on unnamed sources and Apple has not commented, the scope, timing and any team-level exemptions remain unclear. The direction, though, is visible: AI tools are moving out of an open trial period and into procurement.

The signal extends beyond Cupertino. When one of the wealthiest companies in tech, with the resources to build its own models, starts capping its engineers' AI budgets, procurement elsewhere tends to follow. Per-seat, per-usage and per-team allowances are becoming the standard way large organizations manage AI spend.

Three things to watch: whether Apple ties those allowances to project or performance budgets, whether the tightening reaches the tools it offers developers, and whether other large platforms disclose similar cost controls. The productivity promise of AI still stands; it now has to pass a finance review as well.