An Nvidia executive recently remarked that computing costs have now surpassed employee salaries within parts of the company. Last week, the CEO of Mercor also revealed that the startup is spending more on tokens that power its internal AI agents than on staff headcount.
As businesses continue to allocate larger budgets to artificial intelligence, a growing question has emerged: Are companies now spending more on AI than on human workers?
According to new data from the Ramp AI Index, which tracks AI adoption among U.S. businesses, the answer is not yet. The research found that the top 1% of companies—described by Ramp as being heavily committed to AI—are spending roughly $7,500 per employee each month on AI-related services. While that figure is substantial, it still falls well below the average software engineer’s monthly salary of approximately $16,000
Those companies represent the most aggressive adopters. Across the top 10% of firms, AI spending averages about $611 per employee per month. For the median company, spending is much lower at around $11.38 per employee, roughly equivalent to the cost of a single enterprise software seat.
Even so, AI expenditures continue to trend upward. Among the most AI-focused firms, spending per employee increased by 14.1% over the past month alone. Whether that pace of growth can be maintained remains uncertain.
The report also notes that the companies spending the most on AI rarely rely on a single provider. Instead, many combine multiple frontier models and platforms, switching between services while taking advantage of lower-cost open-source alternatives whenever possible.
