Apple is limiting daily spending on artificial intelligence within the company. According to Bloomberg, the limit for engineering teams was lowered from $300 to $200 per day in October, and further cuts—as low as $100—are now under discussion. Employees outside engineering departments could face even tighter restrictions.

This marks a notable shift for a company that has actively integrated AI into its workflows in recent years, including the development of its own chips. Apple now appears to be trying to balance the benefits of new tools against the rising bills for using them.

Engineers Face Budget Cuts

Bloomberg reports, citing sources, that Apple has already reduced the daily AI spending limit for engineering teams from $300 to $200. However, the budget review may not end there: the company plans to lower the spending allowance for some employees to $100 per day.

Outside engineering departments, restrictions could be significantly tighter. According to the publication, the company is considering setting a limit of just $25 per day for these employees.

For now, these measures concern AI usage budgets rather than an outright ban on the tools themselves. Apple has not officially commented on the information; TechCrunch has contacted the company for comment.

From AI Enthusiasm to Token Savings

The situation reflects a broader shift in technology companies’ attitudes toward generative AI. Earlier this year, many employers encouraged staff to use AI models as extensively as possible, hoping to boost productivity.

This approach became known as tokenmaxxing—the drive to consume as many tokens as possible, tokens being the units of text processed by language models. The assumption was that intensive AI use would speed up development, automate routine tasks, and improve team efficiency.

However, the enthusiasm proved short-lived. As the costs of processing requests grew, companies began taking a closer look at the expense of this practice. Instead of unrestricted AI use, employees are increasingly facing budget constraints.

AI’s Economic Benefits Are Not Always Clear Yet

The main problem is the gap between the cost of implementing artificial intelligence and the results it delivers for businesses. Using powerful models requires substantial computing resources, which means it can be expensive even for employees’ day-to-day work.

Meanwhile, actual productivity gains are not always clear. Companies do not yet have a definitive answer to whether the additional spending on AI is offset by time savings and better results.

Apple’s budget cuts show that even technology giants are beginning to move from experiments with the broadest possible use of AI models toward stricter cost controls.

This sends an important signal to the market: the spread of AI in corporate settings will depend not only on models’ capabilities, but also on how convincingly companies can demonstrate a financial return on their use.