Ramp AI Index: Top US firms spend $7,500 per employee monthly on AI
While the most aggressive adopters, termed 'AI-pilled', are burning through significant compute budgets, costs remain below average engineering salaries, though growth rates signal intensifying competition for resources.
The most aggressive adopters of artificial intelligence in the United States are spending approximately $7,500 per employee each month on technology, according to the latest data from the Ramp AI Index. This figure represents the expenditure of the top 1% of US businesses, a group Ramp describes as "AI-pilled" due to their high adoption rates.
Despite the substantial outlay, this level of spending remains below the estimated average monthly salary of a software engineer, which stands at roughly $16,000. The data suggests that while enterprises are allocating significant capital to compute and model access, human labour costs still outweigh artificial intelligence operational expenses for the time being.
The disparity in spending is pronounced when examining broader adoption metrics. The top 10% of firms spend approximately $611 per employee monthly, while the median firm spends just $11.38. This median figure is comparable to the cost of a single seat on an enterprise software plan, highlighting that most organisations are still in the early stages of integration compared to their high-spending peers.
Spending among the top tier is accelerating, with expenditure growing by 14.1% per employee in the previous month. The Ramp AI Index notes that these high-volume users tend to utilise a mix of multiple frontier models and platforms. This strategy often involves leveraging cheaper open-source models to manage costs while maintaining access to advanced capabilities.
The surge in consumption comes amidst warnings from industry leaders about rising compute costs. An Nvidia executive recently stated that the cost of compute has surpassed the salaries of his employees, and Mercor’s CEO noted that his startup is spending more on tokens for internal agents than on headcount. While these comments highlight the rapid inflation in token costs, the broader data from Ramp indicates that such scenarios are not yet the norm across the wider market.
It remains unclear whether the 14.1% month-on-month growth observed among "AI-pilled" firms will sustain itself. As enterprises continue to blow through token budgets, the long-term sustainability of current spending levels relative to compute cost inflation is a key variable for investors and policy makers to monitor.
