Apple is cutting daily spending caps for engineers using external AI models from $300 to $200. The company wants to manage high infrastructure costs as research expenses surged 32% in fiscal year 2026. Many staff now use smaller, cheaper models or on-device processing to keep their work within these new limits.
Apple is clamping down on internal spending for artificial intelligence tools, cutting daily allowances for its engineers despite positioning the technology at the absolute centre of its global operations.
According to a Bloomberg report, the Cupertino-headquartered technology company recently lowered the daily spending cap for select engineering teams using external AI models from $300 to $200. The austerity measures are expected to intensify, with internal discussions pointing toward further reductions that could see daily caps drop to as low as $100 for specific technical departments.
The move highlights an industry-wide paradox: tech conglomerates are aggressively marketing AI to consumers while simultaneously micro-managing the massive token infrastructure costs incurred by their own employees.
A delicate balancing act in Silicon Valley
The budget restrictions strike at the heart of Apple's broader operational strategy. Under the leadership of Chief Executive Officer John Ternus, the iPhone maker is pushing to modernise its internal workflows, debugging practices, and software development pipelines using generative tools. However, this mandate now clashes directly with stricter corporate guardrails.
For non-engineering divisions, where a $100 daily limit was already standard, allowances could plummet to a mere $25. While senior leadership will reportedly permit high-volume users to apply for exemptions, the restrictions have triggered mixed reactions internally. Some staff view the move as a pragmatic adjustment, noting that few routinely exhausted the previous $300 limit. Others fear it signals aggressive corporate restructuring, especially as outside consultants are brought in to evaluate team efficiencies.
The financial reality driving these decisions is stark. In its fiscal year 2026 financial results, Apple revealed that operating expenses surged 22% to $75.5 billion—equivalent to 16% of total net sales. Research and development expenses alone skyrocketed by 32%, heavily driven by infrastructure investments and data costs. To mitigate this, Apple is increasingly directing employees to deploy smaller, less expensive models or leverage local, on-device processing.
The high price of digitalt tokens
The restrictions target "tokens"—the fundamental units of text and data that AI models process. Complex engineering tasks, such as writing software code, consume vast quantities of these tokens, and third-party vendors charge premium rates for access to advanced systems. Apple employees currently have access to a suite of market-leading platforms, including OpenAI’s ChatGPT, Google’s Gemini, and Anthropic’s Claude.
Apple is far from alone in scrutinising its digital consumption. A wave of token budgeting is sweeping through the tech sector as platforms look to prove the return on investment of these tools.
Microsoft has similarly tightened internal user limits, while Meta Platforms has steered its workforce toward internally developed models. In the automotive and transport sectors, Tesla has enforced a strict $200 weekly limit per employee, and Uber has placed a $1500 monthly ceiling on individual coding accounts.