Companies are increasing their spending on AI tools, agents, and compute, with many now spending more on AI than their employees.
An Nvidia executive recently said the cost of compute for his team is now higher than employee costs. Mercor CEO Brendan Foody also said last week that the startup spends more on tokens for internal AI agents than on employee headcount.
However, the same is not true for every company yet, with some still spending more on employee salaries.
Fresh research from the Ramp AI Index shows a wide gap between normal business AI spending and the spending habits of the heaviest users. Ramp describes the top 1% of firms as “AI-pilled” and says these companies spend about $7,500 per employee per month on AI, which converts to over Rs. 2 million.
That is a high figure, but it is still below the roughly $16,000 per month earned by the average software engineer, according to the report.
The highest-spending companies are still unusual cases.
Ramp says the top 1% of firms spend about $7,500 per employee per month on AI. The top 10% spend much less, at around $611 per employee per month.
The median company spends only $11.38 per employee per month, which is close to the cost of a basic enterprise AI seat.
That gap shows that most businesses are still using AI at a limited cost, while a small group of companies is spending heavily on advanced tools, tokens, and internal agents.
AI spending is still rising among the biggest users.
Ramp’s data shows that AI spending among the top 1% of firms grew 14.1% per employee last month.
It is not yet clear whether this pace will continue. The report says these companies often move between multiple frontier models and platforms that also give them access to cheaper open-source models.
That means their spending may change depending on model performance, pricing, compute availability, and internal use cases.