The leadership team at Apple is pulling back on how much money its workers can spend on external AI services. The company recently decided to reduce the daily allowances that engineers and other team members use to access third-party artificial intelligence models like ChatGPT and Claude. This move aims to keep growing research and development costs under control while still encouraging staff to use the technology.
The company lowers daily spending limits for engineering teams
Previously, engineers at the tech giant had a daily budget of $300 to spend on AI tokens. These tokens dictate how much data an employee can process through a large language model. Now, that daily cap has been reduced to $200. Reports indicate that further cuts are planned, which could bring the engineering limit down to $100.
For other departments, the reductions might be even more severe. Some teams could soon face a daily usage budget of just $25. For staff members who rely heavily on AI to write code, review software bugs, or process large amounts of data, these smaller budgets might change how they handle their daily tasks. However, workers can reportedly request a temporary increase to their allowance for specific projects that need more computing power.
Don’t miss the best of The Mac Observer
Set us as a preferred source and our Apple reporting ranks higher in your Google Search results and Discover feed — one tap, no account changes.
Rising research expenses force a shift toward running local models
This financial decision comes after a period of heavy investment. The company saw its research and development expenses jump by 32 percent during the 2026 fiscal year. A big part of that increase came from new server infrastructure and other related tech costs. Overall operating expenses also rose by 22 percent.
To balance the books, the company wants its employees to be more careful about which models they choose for their tasks. Workers are encouraged to use cheaper third-party options when possible. Because modern Mac computers are powerful enough to run many smaller models locally without relying on cloud servers, the business hopes staff will shift some of their workflow directly to their own machines.
While the initial experimental phase of this technology allowed for free spending, tech companies are now looking closely at the actual return on their investments. By tightening these daily limits, the business hopes to maintain a healthy balance sheet without completely cutting off access to helpful new tools.
Discussion