Employees are asked to use AI more, but budgets have been slashed. It is rumored that the daily quota of Apple engineers will be reduced from 300 to 100 yuan.

While Apple is promoting the integration of artificial intelligence (AI) into daily operations, it is also reducing the budget for employees to use AI to control costs.

Citing people familiar with the matter, Bloomberg reported that Apple last month quietly reduced the daily spending limit for some engineers using AI models from US$300 to US$200, and now plans to further lower the limit for some engineering teams, possibly to US$100.

These adjustments reflect the dilemma facing Apple: controlling costs while encouraging employees to use AI more deeply in their daily work. Apple is pushing to use AI to improve engineering processes and other internal operations, but it is also cutting spending on employees who use the technology.

Many non-engineering departments already have a daily spending limit of $100, but Apple is considering lowering the limit to $25 for some teams, people familiar with the matter said. Employees whose work requires higher quotas may apply to their superior supervisor for exception approval.

These budgets are used to pay for “tokens” consumed by the AI ​​model when processing instructions. Tokens are units of measurement that are used by models to process text and other data. AI service providers typically charge based on consumption, and more advanced models are typically more expensive. Complex tasks like writing programs, finding and correcting errors, etc. can consume a large number of words.

The restrictions also reflect broader changes in the tech industry. The industry initially encouraged employees to experiment with AI on a large scale, but now they are increasingly scrutinizing the costs. The difficulty is to save money without weakening the work efficiency brought by AI.

Reaction within Apple was mixed. Some employees said that few people used the $300 per day limit and that the original limit was too loose; Others believed that lowering the limit meant the company as a whole was saving costs and deepened concerns about layoffs.

Some Apple departments have hired external consultants to evaluate efficiency and operations, making employees more concerned about possible future restructuring. Bloomberg previously reported that new CEO Turner is pushing to streamline the organization and reduce management levels to make the company operate more efficiently.

Since some employees have not used the cap, it is still unclear how much money this adjustment will actually save. If Apple makes further cuts, the impact may be more pronounced. Employees who frequently use AI may have to apply for a separate increase in quota before they can continue using the equipment required for work.

On the other hand, Apple’s total expenses are still increasing. Apple said this week that research and development spending will increase 32% in fiscal year 2026, due to increased infrastructure and AI costs; Total operating expenses will increase by 22%, reaching US$75.5 billion, accounting for 16% of net sales.

Apple employees can use AI services like Anthropic’s Cloud, OpenAI’s ChatGPT, and Google’s Gemini. Some teams have been asked to switch to less expensive models or AI models that execute directly on the device.

Other technology companies are also reining in AI spending. Microsoft has tightened employee budgets for using AI, and Tesla has limited weekly spending to $200. Uber limits the use of personal AI program development tools to $1,500 per month; Meta also sets a Word budget and encourages employees to switch to internally developed AI tools.