On November 4, Bloomberg reported, citing people familiar with the matter, that Apple last month cut the daily AI spending allowance for some engineers from $300 to $200, and is considering lowering it further to $100; non-engineering teams could see their caps fall from $100 to $25. Apple is asking employees to use AI more while tightening what they may spend on it — the reason is cost. Apple did not respond to a request for comment.
How much was cut, and will it go further?
The allowance is charged per token and covers models employees use daily, including Claude, ChatGPT and Gemini, with coding consuming it fastest. The October change has already taken effect: some engineering teams went from a $300 daily cap to $200; the option under discussion would take engineering teams to $100 and non-engineering teams from the current $100 to $25. Employees who run out can ask senior executives for an exemption. Bloomberg also flagged a detail: few employees ever used their full $300 allowance, so how much this tightening will actually save is unclear — the signal matters more than the savings.
Where the cost-cutting comes from
Tightening AI budgets is not an isolated move. Apple's annual report, filed on November 3, shows research and development spending of $45.8 billion in fiscal 2026, up 32% year on year, driven largely by AI and infrastructure; total operating expenses rose 22% to $75.5 billion. John Ternus, who became chief executive in September, is pushing a broad cost-cutting effort aimed at making the company faster and leaner, and reports have already described cuts to engineering program manager roles. Notably, Apple is not cutting AI investment itself: the R&D bill keeps growing, on computing power and infrastructure; what is being tightened is the scattered spending on employees' everyday calls to outside models. In other words, the company keeps spending big on AI, while casual use at the individual level now has to be justified.
For ordinary users, this change will not be visible in the short term: it targets the model-usage allowances of Apple's own employees, not consumer products and services. But it reflects a widening trend: as AI moves from novelty to an everyday productivity tool, companies are starting to manage token bills the way they manage travel expenses. Microsoft, Tesla, Uber and Meta are all tightening internal AI experimentation spending. What to watch next at Apple is twofold: whether the new $100 cap really lands, and whether employees shift toward cheaper models or models that run locally as a result.
