As spending on AI tools balloons across the industry, Apple is moving to rein in costs from the inside out. The company has started capping how much its employees can spend on AI models each day, according to a Bloomberg report citing company sources.

The new limits align with the approach of CEO John Ternus, who took office on September 1 with a stated goal of building a leaner operational structure.

A $200 daily cap for engineers

gorsel_20261105_093006

Last month, Apple reduced the daily spending limit engineers have for AI models from $300 to $200. In some engineering teams, that figure is expected to drop as low as $100 in the coming weeks.

Outside engineering, where the daily limit already sits at $100, some teams could see it cut to $25. Anyone who needs a larger budget must obtain special approval from senior management. Apple has not commented on the matter.

What do the limits cover?

The amounts in question correspond to spending on "tokens," the data-processing units of AI models. Because coding and debugging consume large volumes of tokens, the caps hit engineers hardest.

Employees can keep using tools such as Anthropic's Claude, OpenAI's ChatGPT and Google's Gemini. However, some departments have reportedly been instructed to shift toward cheaper models or ones that run on-device.

Unease inside the company

The decision has drawn mixed reactions from staff. Some employees say they rarely exceeded the $300 cap in the first place, while others see the move — coming after more than 200 layoffs in August — as a sign of a broader cost-saving drive and possibly more job cuts to come.

Ternus is working with external consultants on efficiency reviews, aiming to flatten management layers and make Apple more agile. Yet despite the cutbacks, the company's overhead keeps climbing: according to its annual report, R&D spending rose 32% in fiscal year 2026, while total operating expenses grew 22% to reach $75.5 billion.