Companies are putting limits on employees’ AI spending, but a spending cap alone cannot tell a finance team whether that usage is wasteful or valuable. PYMNTS reports that Apple is cutting AI budgets for some internal teams by as much as a third, while Tesla and Uber have set employee allowances.
Watch Desk analysis
What happened
The reported limits include $200 a week at Tesla and $1,500 a month at Uber. PYMNTS says Apple’s policy allows exceptions for employees who need higher allowances. The article frames the change as a shift towards managing AI’s usage-based costs as adoption grows.
Its research offers some context, though not a clean verdict on returns. PYMNTS says 62% of US enterprises with at least $1 billion in annual revenue reported spending more than $10 million on new forms of AI in the previous 12 months. In separate research, it says only 5% to 10% of surveyed firms in healthcare, financial services and media had achieved full payback on recent AI investments, despite 90% to 100% rating returns positively.
Why it matters
A flat allowance can curb costs, but it may also restrict people doing valuable work. The more useful measure is what it costs to complete a workflow, including the AI service, staff time and oversight, rather than how many prompts someone uses. PYMNTS also points to the possibility that AI costs one department money while producing value elsewhere in the business.
That makes the choice less tidy than “cap it” or “let it run”. A business can save on AI charges while leaving an inefficient process untouched, or spend more without showing that the extra usage pays its way.
Our read
Treat a cap as a budget guardrail, not evidence that an AI tool is either too expensive or worth its bill. The sensible next step is to measure complete workflows and make exceptions possible when teams can show what the extra usage achieves. The spreadsheet, alas, still has a job.
What to watch
- Whether companies publish how they measure AI costs and productivity gains.
- Whether spending limits include quick exceptions for high-value work.
- Whether reported AI returns translate into full payback, not just positive impressions.
Discussion spark: Should companies cap employees’ AI spending by default, or require evidence of workflow value before setting limits?
Sources and evidence
- Source update (5 November 2026, 20:30 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.