Apple slashes internal AI spending as token costs mount

Apple has reduced daily AI spending limits for its engineering teams from $300 to $200, with further cuts planned to $100 for some groups and as low as $25 for non-engineering staff, Bloomberg reported. The move reflects a broader industry shift from "tokenmaxxing"—pushing staff to use AI heavily for vague efficiency gains—to token rationing as companies grapple with the bills. Concerns continue about the gap between AI's operational costs and its often unclear productivity benefits.

Key points

• Apple lowered daily AI spending caps for engineering teams from $300 to $200 in October 2026
• Further cuts are planned: some engineering allowances will drop to $100/day, non-engineering to as low as $25/day
• The shift follows an earlier industry trend of "tokenmaxxing" where companies encouraged maximum AI usage for efficiency
• Industry-wide concerns persist about the cost-to-productivity ratio of AI tools
• Apple had been integrating AI into engineering processes but is now tightening budgets

How to use it

• Establish clear AI spending caps and monitor token usage before costs accumulate
• Track actual productivity gains against AI spending to justify budgets and adjust limits based on measured ROI
• Consider tiered allowances by role, as different teams may require different capacity levels
• Prepare for potential budget scrutiny as the industry recalibrates AI spending expectations

Has your organization set AI spending limits, or are you still in the "try everything" phase?