Apple is reportedly tightening its internal AI spending. TechCrunch AI reports, citing a Bloomberg story published Wednesday, that Apple lowered the daily AI spending cap for its engineering teams from $300 to $200 in October, and more cuts are coming. Some allowances could fall to $100 a day, while staff outside engineering may get a limit as low as $25. TechCrunch says it has asked Apple for comment, so none of this is officially confirmed yet.
The short version
- What changed: Apple’s daily engineering cap went from $300 to $200 in October.
- What’s next: Some allowances may drop to $100 a day, and Apple is weighing a limit as low as $25 for non-engineering staff.
- Why it matters: One of the world’s richest companies is rationing AI use. Expect that to give other finance teams cover to do the same.
- What to do: If your team relies on AI tools, start tracking cost per task now. It’s better to have those numbers before someone asks for them.
What the caps look like over a year
Here’s my rough math, assuming about 250 working days a year:
- $300 a day: about $75,000 a year
- $200 a day: about $50,000 a year
- $100 a day: about $25,000 a year
- $25 a day: about $6,250 a year
Two caveats apply. These are ceilings, not actual spend, and the report doesn’t say how many people hit them. TechCrunch’s summary also doesn’t spell out whether each cap applies per person, though the word “allowances” suggests it might.
Even so, the direction is clear. A drop from $300 to $100 would be a 67% cut to the ceiling.
From tokenmaxxing to token rationing
This is a sharp turn from where the industry was just months ago. TechCrunch describes a “tsunami of tokenmaxxing” earlier this year, with companies pushing staff to use AI as much as possible in the name of “vaguely defined efficiency gains.”
Tokens are the units AI providers typically bill by, so more usage means a bigger invoice. That’s especially true of coding agents, which can burn through a lot of tokens on a single task because they keep rereading code and retrying steps.
Then the bills arrived. According to TechCrunch, the frenzy was short-lived, and companies decided that rationing tokens made more financial sense.
What stands out here is who’s doing it. Apple has worked to build AI into its engineering processes, and it isn’t short on cash. If it’s watching the meter this closely, smaller companies have even less reason to treat AI usage as an open tab.
The real question is productivity
The caps point to a bigger problem. TechCrunch notes that concerns persist about the gap between what AI costs to build and run and the often unclear productivity gains it’s supposed to deliver.
That gap is hard to ignore when AI spend sits on the budget as a daily, per-team line item. Token costs are easy to count. Time saved, bugs avoided and features shipped sooner are much harder to measure.
I don’t read this as Apple giving up on AI. A $200 daily cap is still real money. It looks more like a company moving from “use as much as you can” to “show me what we got for it.”
What to expect next
If other big companies follow Apple’s lead, here’s what I’d watch for:
- Tiered access: Engineers get bigger budgets than everyone else, as Apple’s reported $100 and $25 limits already suggest.
- Smarter model use: Teams send simple tasks to cheaper models and save the expensive ones for hard problems.
- Harder ROI questions: Vendors will need better answers than “people feel faster.”
- Pricing pressure: Customers with fixed budgets will care a lot more about cost per task.
Practitioners should get ready for the same shift. Knowing which AI workflows pay for themselves is about to matter as much as knowing how to use the tools.
The big unknown is whether Apple’s cuts improve its numbers without slowing its engineers. We may not hear that part anytime soon, but more details are available in the original TechCrunch AI report.