Apple has recently reduced the daily AI token spending limits for many of its engineers, scaling back from a previous $300 cap to $200 in some cases. Additional cutbacks are being considered across multiple teams, with some roles potentially seeing as little as $100 per day. Teams outside of engineering already working under the lower limit may see those caps drop even further, potentially to $25. Employees who require greater usage for their work will be able to request exceptions through senior leadership.
What’s Changing at Apple?
The changes reflect a broader push under CEO John Ternus to reduce internal expenses. While engineers had access to daily allowances of up to $300 for AI-powered tools, that amount has now been trimmed to $200 in certain areas. The firm is also discussing additional reductions, aiming to limit some teams to $100 per day. Many non-engineering groups already operate under $100 daily caps, and the company is exploring dropping those to $25 for certain roles.
Not all is being cut indiscriminately: teams that rely on large-scale token usage — such as for coding or research work — may still be granted higher allowances on a case-by-case basis. The company hasn’t yet clarified which AI tools or services are impacted, but employees regularly use platforms like Anthropic’s Claude, OpenAI’s ChatGPT, and Google’s Gemini. In some cases, Apple is encouraging use of more cost-efficient models or on-device AI to reduce third-party service costs.
Why Now?
This move comes as cost pressures grow across the tech sector. While token usage had become a way to measure employee output in some companies, soaring bills have forced many to reconsider the practice. Apple itself reported a 32% increase in R&D spending for fiscal year 2026, reaching $45.8 billion — in part due to rising AI and infrastructure costs. Tightening token budgets appears to be part of Ternus’s strategy to rein in spending without widespread cuts.
Employee reactions are mixed. Some report rarely using their full previous allowance, while others see the cuts as evidence of escalating belt-tightening. For some, it’s a sign cost containment may lead to broader reductions; for others, it’s just adjusting expectations to align with organizational priorities.
For context, last year Apple encouraged its workforce to lean into AI more generally—highlighting its growing role across teams and highly emphasizing its potential impact. Tokenmaxxing — exceeding token usage aggressively — was becoming commonplace, partly due to increasingly capable tools like Claude Code, OpenAI Codex, and Cursor. But the steep costs associated with large-scale use have prompted many, including Apple, to pull back.
What to watch: whether other large tech firms follow Apple’s lead in cutting token budgets, what internal exemption requests may reveal about team workflows, and whether the push toward cheaper or local AI models gains traction within Apple’s organization.
