Particle.news

Apple Cuts Employee AI Token Allowances as New CEO Pushes Cost Cuts

The change seeks to rein in unpredictable model billing by steering teams to cheaper or on‑device models while allowing exemptions for mission‑critical work.

Overview

  • Apple reduced some engineers’ daily AI spending caps from $300 to $200 in October and is weighing deeper cuts that could bring certain engineering teams to $100 per day.
  • Many non‑engineering groups already face $100 daily caps and Apple has considered lowering some of those limits to as little as $25 with a formal exemption process for needed work.
  • The budgets are measured in AI “tokens,” the units providers charge for processing text and data, and Apple is directing teams to cheaper third‑party models or local on‑device inference to reduce per‑token costs.
  • Employee reaction inside Cupertino is mixed: several workers say few used the full $300 allowance so savings may be limited, while consultant reviews and the new leadership’s push to slim management have stoked layoff and restructuring worries.
  • The move follows a broader industry shift to tighter internal AI controls and comes as Apple’s fiscal 2026 R&D and infrastructure spending rose sharply, with R&D up about 32 percent.