Apple Reportedly Tightens Engineering AI Spending Limits

By Sophie Lindqvist

What changed in Apple's engineering budget?

Apple reportedly cut the daily ceiling on internal AI spending for its engineering teams from $300 to $200 in October, a 33% reduction, Bloomberg reported Wednesday.

Bloomberg also reported that more cuts are in the works. Unnamed sources said some allowances could fall to $100 a day. That would halve the October ceiling and equal one-third of the previous $300 limit.

Outside engineering, Apple is weighing a daily limit as low as $25, according to the report. That proposal has not been approved. The $100 figure also remains a plan rather than a disclosed policy.

The $25 proposal is 87.5% below the $200 ceiling, but the two figures cover different groups. The source does not establish that engineers and other employees face equivalent costs or workloads.

This distinction matters because a cap governs a budget, not necessarily actual consumption. The report gives ceilings and proposed ceilings. It does not show how much Apple or each group previously spent.

What sits behind the tighter limits?

Apple has worked to integrate AI into its engineering processes, according to the report, so tighter usage limits do not automatically signal a retreat from that work. The change may instead govern how freely engineers can use the systems already available to them.

Earlier this year, companies pushed staff to use AI as much as possible, a practice the report calls “tokenmaxxing.” The stated objective was “vaguely defined efficiency gains,” according to the report.

The report said the push faded as token bills arrived. Companies then turned toward token rationing as a more fiscally sensible response.

Persistent concerns center on a gap between the cost of building and running AI and the often unclear productivity gains it is supposed to deliver. Apple's own internal spending data could have clarified whether that concern applies at the company, but the report provides no such figures.

The October reduction to $200 therefore fits this broader corporate shift. The source offers no Apple-specific productivity data to test that broader rationale.

What does the report not establish?

TechCrunch has reached out to Apple for comment, but the report did not provide a company response. Neither the $100 nor the $25 figure should be presented as a final policy for the affected groups without that confirmation.

The report leaves several commercial variables unresolved:

• Apple's total internal AI bill.
• The number of teams affected by each cap.
• The exact date in October when the $200 ceiling took effect.
• Whether an allowance covers an individual, a team, a project, or another cost category.

That distinction matters because a $100 cap can mean very different commercial pressure depending on what it governs. A project-wide ceiling is not equivalent to a personal daily allowance.

The absence of total spending also blocks a reliable annual cost estimate. The source does not establish how many allocations exist or whether the ceiling applies every day.

No evidence in the report links lower caps to a reduction in Apple's broader AI investment. The stronger claim would be that Apple is cutting overall AI investment; the available evidence supports only a reduction in daily spending for internal AI use.

What could the proposed limits change?

If some allowances fall to $100, engineers would have half the October ceiling available for internal AI, potentially limiting frequent or concurrent experimentation. The commercial effect would depend on whether teams treat the new ceiling as a hard limit.

A $25 cap outside engineering would extend the same restraint to a broader group, although the proposal remains under consideration. The 87.5% reduction from $200 is arithmetically clear, but organizational differences prevent a direct like-for-like comparison.

For Apple, the practical test is whether lower daily budgets produce measurable engineering output without slowing deployment work. That test matters because the source provides no productivity metric, usage telemetry, or savings estimate capable of proving either outcome.

More broadly, the episode reflects a shift from maximum AI adoption toward financial discipline, as companies confront bills before they can demonstrate clear returns. The reported Apple cuts suggest that internal AI usage now requires a more visible cost-benefit case.

Until Apple provides realized spending, affected-team counts and productivity measures, the next signal will be whether the proposed $100 and $25 limits become policy for affected groups. That decision will determine how tightly Apple rations internal AI use and whether token rationing becomes a durable operating rule rather than a temporary budget response.