NextFin News - The top 1% of U.S. firms classified by the Ramp AI Index as “AI-pilled” are spending about $7,500 per employee each month on AI, according to TechCrunch.
That figure has quickly become a shorthand for how fast enterprise adoption is moving, but it is not a broad corporate average. It reflects the heaviest users, not most companies. TechCrunch notes that the amount is still below the roughly $16,000 a month the average software engineer makes. The comparison is imperfect, but it helps explain why managers often present AI as a productivity tool rather than a direct substitute for payroll.
TechCrunch also reported that an Nvidia executive recently said compute costs now exceed employee salaries. Last week, Mercor’s chief executive said his startup was spending more on tokens for internal agents than on headcount. Together, those comments describe the same shift: AI is becoming a recurring operating expense, not just another software purchase.
Whether that spending reflects healthy adoption or uncontrolled experimentation depends on who is paying the bill and what they are getting for it. For well-capitalized startups and software companies, heavier token usage can mean AI is being built into customer support, coding, sales, or internal operations at real scale. For firms still struggling to find a reliable return on investment, the same spending can look like an expanding burn rate. Token costs are variable, and that changes the budget math. Unlike a one-time software license, each new workflow, agent loop, or model call can raise the bill. In AI, successful usage can bring cost pressure with it.
The companies spending the most are often under the greatest competitive pressure, where speed matters and experimentation is rewarded. Those firms may be willing to pay more for AI because the alternative is moving slower than rivals. In that sense, $7,500 per employee per month is better read as a snapshot of the leading edge than as a benchmark for the typical American enterprise. It does not support the idea that AI has already become more expensive than labor across corporate America.
The comparison with the average software engineer’s monthly pay matters because AI is being marketed most aggressively to software-heavy organizations, where the productivity case is easiest to make. If a company can spend thousands of dollars a month on AI and cut the time spent on coding, debugging, or support, the economics can work. If AI tools are simply added on top of existing teams without reducing manual work, the spending becomes additive rather than substitutive. That is when executives start asking why the AI bill is rising faster than the savings.
There are also limits to what a single index can show, even one from a data-driven company like Ramp. Adoption measures can tilt toward companies with modern procurement systems, better visibility into card and vendor spending, or heavier use of trackable AI services. They can miss private contracting, embedded enterprise licenses, or usage folded into broader cloud bills. They also do not fully capture the value side. A company spending more on AI may still be improving margin, speeding up product release cycles, or shortening customer response times. The headline cost shows intensity, not efficiency.
AI usage is becoming more tied to day-to-day operations, and that tends to produce recurring costs. This is a different stage from the early demo period, when companies ran small pilot projects and mostly paid for experimentation. Now the spending is tied to production workloads, internal agents, and always-on software behavior. It also forces budgeting decisions. Once AI is part of a workflow, the question shifts from whether to use it to where to cap it.
For investors, the key point is that AI economics are becoming more uneven, not less. Some companies will use the technology to limit headcount growth. Others will use it to push faster revenue expansion. Some will absorb the cost as the price of staying competitive. The result is likely to be uneven margins across sectors and within the same industry. The $7,500-per-employee figure shows how intensely the most committed users are spending, but it does not prove that AI spending is broadly eclipsing human labor.
Explore more exclusive insights at nextfin.ai.