Apple has cut the daily amount some of its engineers may spend on artificial intelligence, lowering the allowance from $300 to $200 in October, with deeper reductions still under discussion. The move puts a price on the AI tools the company’s staff use every day, and it shows that even the world’s most valuable company is now counting what its AI habit costs.
What changed in October
In October, Apple quietly lowered the daily spending cap for some engineers using AI models to $200, down from $300, according to people with knowledge of the change. The company is now weighing further cuts for some engineering teams, potentially reducing their daily allowances to as little as $100.
Teams outside engineering already work to a $100 daily cap, and Apple has considered cutting that to $25 for some of them. Staff who need more than their allowance to do their jobs can seek an exemption from senior leadership, so the caps act as a gate rather than a hard stop.
What the money buys
The allowances pay for tokens, the units of text and other data that AI models process when they handle a request. Apple staff use outside models including OpenAI’s ChatGPT, Anthropic’s Claude and Google’s Gemini, and heavy coding work can burn through tokens fast, which is why engineers carry the largest allowances.
Apple declined to comment on the change, and it is not clear exactly which tools fall under the caps. That gap matters, because the practical effect depends on whether the limits touch the coding assistants engineers lean on most or sit across a wider set of services.
For engineers, the arithmetic is stark. At the old $300 cap, a single engineer could run up roughly $6,000 in a working month; at $200, that falls to about $4,000. Multiplied across thousands of engineers, the allowances explain why finance teams now treat AI tokens like any other cloud bill.
Why Apple is counting tokens now
The cuts sit inside a wider cost push under chief executive John Ternus, who took over on September 1 and has moved to run Apple leaner. Research and development spending rose 32 percent in fiscal 2026, driven by higher infrastructure and AI costs, so the token bill lands on top of an already steep climb.
Apple is also not alone in pulling back. Earlier this year many companies pushed staff to use AI as much as possible, then watched the bills arrive and started rationing instead. Amazon told its engineers that more tokens are not better work, and the same rethink is now spreading across the industry.
The pattern reaches beyond tech. Goldman Sachs pulled AI coding tools from some bankers after a review of productivity and risk, a reminder that the token question is now a board-level one. For Apple, the tension is sharper, because it is cutting AI spending while trying to put AI at the centre of its operations.
Our read: the bill is the story, not the budget
The headline numbers flatter the saving. Few workers were using their full $300 daily allowances, so the actual money recovered is probably small. The real story is that AI spending has become a line item to manage, not a blank cheque, and that shift changes how companies judge the tools they rushed to adopt.
That is a healthy correction, but it carries a risk of its own. Caps set too low push engineers toward cheaper models or off-the-books workarounds, and the productivity Apple hopes to protect could quietly leak away. The exemption route helps, but only if managers grant it quickly when the work genuinely needs the tokens.
Apple’s own ambitions make the tension visible. The company is asking staff to weave AI into daily work while capping the tokens that work consumes, and the balance will be judged by output, not by the size of the budget. If the caps slow the coding assistants engineers rely on, the saving could cost more than it returns.
Why It Matters
If Apple is rationing tokens, every other company will feel licensed to do the same. The move tells the industry that the era of unlimited AI use is over, and that vendors will now have to prove their tools earn the spend rather than simply counting on usage to grow.
There is a second thread worth watching. Apple filed its annual report on November 3 warning that AI agents can act on their own in ways users never cleared, so cost is not the only reason to slow down. Managing AI inside a giant company is now about money and control alike, and Apple is tightening both at once.
Background
Ternus took over as chief executive on September 1, and his cost push has already reshaped how Apple runs its engineering teams. He has drawn a line on AI: useful, not loud, and the new caps fit that message — spend on AI where it pays, and nowhere else.
Source: Reuters report on the cuts and TechCrunch report
0 Comments