Apple has begun restricting how much its engineers can spend on artificial intelligence services each day, reducing allowances from $300 to $200 last month and signalling plans for even steeper reductions ahead, according to Bloomberg. The move reflects broader efforts by new chief executive John Ternus to streamline operations, even as the company's research and development expenditure climbed 32% in fiscal 2026 owing to infrastructure and AI investments.
The daily spending caps apply to tokens—the computational units that language models consume when processing requests. Tasks like code generation and debugging can rapidly deplete token allowances. Engineers currently have access to Anthropic's Claude, OpenAI's ChatGPT and Google's Gemini, though some teams are being encouraged to migrate toward lower-cost alternatives or models that execute locally on devices.
Beyond the engineering division, staff in other departments already operate under a $100 daily ceiling. Apple is weighing whether to reduce this further to $25 for certain teams. Those requiring additional capacity can petition senior management for special approval. The company has declined to offer any public statement on the matter.
Employee sentiment within the organisation remains divided, Bloomberg's sources indicate. Some workers note that relatively few colleagues exhausted their previous $300 allowance. Others interpret the reductions as part of a larger austerity programme and express concern about potential workforce reductions, particularly following Apple's elimination of more than 200 positions in August.
Since taking the helm on 1 September, Ternus has prioritised operational efficiency and removed management layers. Certain business units have also brought in external consultants to assess their operational effectiveness.
Despite these constraints, Apple's overall financial outlay continues to expand. The company disclosed in its latest annual report that research and development costs jumped 22% to $75.5B, driven substantially by heightened infrastructure and AI expenditures.
Apple is far from alone in implementing such restrictions. Microsoft established division-specific AI budgets in July, Tesla restricts weekly spending to $200 per employee, and Uber caps individual AI coding tools at $1,500 monthly. Accenture has imposed even tighter controls, limiting Claude usage to $29 per month. Across the technology sector, companies are implementing spending caps as AI service costs continue to mount.
Source: The Next Web
