Apple Is Rationing Its Own AI: What It Means for AI Platform Buyers
4 November 2026
Apple spends more on AI than almost anyone, and even it has started rationing. Bloomberg reports that Apple cut the daily AI budget for its engineering teams from $300 to $200 in October, with further cuts to $100 a day on the way. Outside engineering, the company is reportedly weighing a limit as low as $25 a day.
The number matters less than what it reveals: the cost of running AI has grown faster than the value it produces. If you are comparing AI platforms, that gap is exactly what their pricing pages are trying to hide.
Based on reporting by TechCrunch.
What Apple actually changed
Apple's engineers can now spend $200 a day on AI tools, down from $300. Sources told Bloomberg the cuts will continue, with some teams dropping to $100 a day. Outside engineering, the daily allowance could fall as low as $25.
None of this is about Apple being short of money. It is about a company realising that unlimited AI spending does not produce unlimited returns.
The bill that came due
Earlier this year, companies chased what can only be called 'tokenmaxxing': pushing staff to use AI as much as possible in the name of efficiency. The wave did not last. When the token bills arrived, the efficiency gains were harder to find than the invoices.
The result is token rationing: a budget, a cap, and a hard look at what each AI call actually returns. Apple is not the first to do this, but it is the loudest signal yet.
Every AI product has a token bill
Apple's cuts are not unique to Apple. Every AI feature runs on tokens, and tokens cost money. A chatbot answer, a trading signal, a 'free' analysis: each one sends a request to a model somewhere, and the provider pays for it.
When a platform sells you 'unlimited AI' for a flat fee, it is betting that you will use less than it costs to serve you. That bet works until usage spikes, and then the platform adds caps, throttling, or a price change. The budget Apple is now setting internally is the same arithmetic every AI platform has to solve.
How to spot a fake 'unlimited AI' claim
A platform that promises 'unlimited AI' is promising something even Apple will not fund for its own staff. Before you believe the claim, run these four checks.
1. Read the terms, not the headline. Look for 'fair use', daily caps, token limits, or throttling after a certain volume. Apple's internal cap is $100 to $200 a day; platforms use the same language in their terms.
2. Check what happens after heavy use. If the platform does not say what happens when you generate a lot, assume it throttles you or moves you to a paid tier.
3. Look for a real price list. A credible AI platform prices its AI usage openly, or at least admits limits exist. Vague copy is the tell.
4. Search for existing complaints. If users report sudden throttling, removed features, or price hikes, the 'unlimited' promise has already failed for someone else.
The verdict
Treat 'unlimited AI' as a red flag, not a feature. A credible AI platform prices its AI usage openly, or at least admits the limits exist. If the marketing promises infinity and the terms mention caps, the platform has already told you which one to believe.
Apple rationing its own AI is the clearest public proof that AI's unit economics are still unsolved. The same arithmetic applies to every AI trading platform and every 'free' AI tool you are shown.
Before you sign up to any AI platform, ask the question Apple is now asking every day: what does this actually cost, and what does it actually return? If the platform cannot answer both clearly, treat its 'unlimited AI' claim as a cost problem waiting to surface.