Bloomberg reported on Wednesday that Apple is tightening the daily financial ceiling its engineering divisions can allocate to artificial-intelligence tools. The policy shift marks a notable departure from the company's earlier strategy of embedding AI broadly across product development and internal workflows, signaling a more cautious approach to AI expenditure.
In October, the company lowered the per-engineer daily cap from $300 to $200, according to the same Bloomberg source. The adjustment reduced the amount of compute and model-access tokens that individual developers could consume each day, aiming to curb unchecked spending while still permitting essential experimentation within product teams.
Sources familiar with the plan tell Bloomberg that Apple intends to introduce additional tiers of restriction. Some engineering groups may see their daily allowance fall to $100, while non-engineering functions could be limited to as little as $25 per day. The tiered model reflects a broader effort to align AI usage with budgetary constraints across the organization.
Earlier in the year, many technology firms experienced a surge of what insiders called “tokenmaxxing,” a rapid escalation in AI model usage driven by vague promises of efficiency gains. Companies encouraged staff to tap AI services extensively, hoping to accelerate development cycles, but the practice quickly ran into financial reality as usage fees accumulated.
When the monthly invoices arrived, firms found the cost of token consumption far exceeding initial expectations. The sudden expense prompted many to scale back, instituting internal quotas and stricter approval processes. Apple’s recent budget caps appear to be a direct response to this fiscal pressure.
Analysts continue to question whether the monetary outlay on AI tools translates into measurable productivity improvements. The lack of clear metrics makes it difficult for companies to justify unrestricted spending. Apple’s incremental tightening of daily limits may signal a broader industry trend toward more disciplined, cost-aware AI adoption.