US businesses are racing into AI at wildly different speeds, with a tiny group of heavy adopters turning AI into a major line item while most firms treat it as little more than a software add‑on.

New data from the Ramp AI Index shows that the top 1% of US “AI‑pilled” companies now spend about $7,500 per employee per month on AI tools and compute.¹ By contrast, the median firm spends just $11.38 per employee monthly, roughly the cost of a single enterprise AI subscription.²

Chronologically, the pattern began with isolated anecdotes from tech leaders who claimed AI compute was starting to rival — or even exceed — payroll. An Nvidia executive recently said “the cost of compute is now greater than the salaries of his employees,” while Mercor’s CEO reported spending more on “tokens for internal agents than on employee headcount.”¹

Ramp’s latest figures, published on June 10, 2026, suggest those cases are still outliers. Even among the most AI‑obsessed firms, the $7,500 AI spend is “certainly not more than the roughly $16,000 per month the average software engineer makes.”¹ Yet growth is rapid: within this top 1%, AI spend per employee jumped 14.1% in the last month alone.²

The top 10% of adopters sit in the middle, spending around $611 per employee per month — equivalent to a few enterprise seats plus API usage.² These power users typically “mix and match, bouncing between multiple frontier models and platforms” to balance cost and capability rather than commit to a single provider.¹

For the vast majority of companies, however, AI remains a minor expense: at $11.38 per employee, it is “still a rounding error on their software budgets.”² The central question emerging from this divergence is whether today’s $7,500 top‑tier spend is a ceiling or a floor — and whether, as agentic AI expands what can be automated, the habits of the AI‑pilled 1% will eventually become the norm.²