The growing use of generative AI inside technology companies is bringing a new expense under scrutiny: how much employees spend on accessing these tools. Apple is reportedly tightening the amount its workforce can use on paid AI services, with some non-engineering employees potentially facing daily limits of just $25.
According to a Bloomberg report, the company has already reduced the daily AI spending allowance for some engineering employees from $300 to $200 in October. Further cuts are now being considered for certain engineering teams, potentially bringing their allocation down to $100 per day.
The proposed restrictions could also affect employees outside engineering. Some non-engineering teams currently have a daily allowance of $100, which could be reduced to $25. Employees requiring greater access to AI services can reportedly request exemptions from senior executives.
These budgets cover AI token usage across services such as OpenAI’s ChatGPT, Anthropic’s Claude and Google’s Gemini. Tokens measure the volume of data processed by AI models, meaning consumption can vary depending on how frequently employees use the tools and how complex their requests are.
Apple’s move reflects a wider reassessment of the costs associated with integrating AI into everyday workplace operations. While companies have encouraged employees to experiment with these tools, extensive usage can generate significant expenditure. The question is increasingly shifting from how much AI employees use to whether that usage delivers measurable value.
Several major technology and financial companies have introduced similar restrictions. Bloomberg has reported that Goldman Sachs limited access to certain AI coding tools after evaluating their productivity benefits and risk management implications.
Amazon has also cautioned software engineers against treating higher token consumption as an indicator of productivity. Its position comes amid concerns around ‘tokenmaxxing’, a practice involving excessive AI token usage that does not necessarily translate into meaningful improvements in output.
Other companies are adopting different approaches to managing the expense. Microsoft has introduced tighter controls on employee AI spending, while Tesla has established a weekly spending ceiling of $200. Uber has capped monthly expenditure on individual AI coding services at $1,500 per employee. Meta, meanwhile, has introduced token allowances and encouraged employees to use its internally developed AI tools.
For Apple, the reported reductions suggest that access to paid AI services may increasingly be governed by departmental budgets and individual requirements, rather than unrestricted employee usage.
The reported cuts highlight a growing tension in enterprise AI adoption: companies want employees to use advanced tools, but widespread access comes with recurring costs. As spending controls become more common, businesses may increasingly evaluate AI adoption through productivity gains and measurable outcomes rather than usage volumes alone.
