The companies going all-in on AI are now spending $7,500 per employee every month on it. That number comes from the Ramp AI Index, as reported by TechCrunch AI, and it gives us the clearest look yet at how far the most aggressive adopters are willing to push their budgets. Ramp calls these firms “AI-pilled,” and they sit in the top 1% of American businesses for AI spend.
What stands out here is how wide the gap is between the believers and everyone else. Most companies are barely spending anything. The leaders are spending like AI is already core infrastructure.
What the research measured
The Ramp AI Index tracks how American businesses adopt and pay for AI. By pulling real spending data across firms, it sorts companies into tiers based on how much they put toward AI per employee each month. That approach matters because it cuts through the hype. Instead of asking executives what they plan to do, it looks at what they’re actually paying for.
Here’s how the spending breaks down, according to TechCrunch AI:
- Top 1% (“AI-pilled”): about $7,500 per employee per month
- Top 10%: about $611 per employee per month
- Median firm: about $11.38 per employee per month
That median figure is roughly the cost of a single seat on an enterprise plan. So the typical company is buying one subscription and calling it a day, while the top tier is spending more than 600 times that.
Is AI replacing payroll yet?
This is the question driving the whole conversation. An Nvidia executive recently said compute now costs more than his employees’ salaries. Mercor’s CEO said last week the startup spends more on tokens for internal agents than on headcount. Those quotes make it sound like the tipping point has arrived.
The data says not yet. Even at $7,500 a month, the AI-pilled firms aren’t outspending humans. The average software engineer makes roughly $16,000 a month, so AI spend per employee is still less than half a single salary. Impressive growth, but not a replacement for payroll.
The trend is still climbing
Among the AI-pilled firms, spend per employee grew 14.1% last month alone, TechCrunch AI reports. That’s a steep curve. Whether it holds is the open question. Budgets face real pressure, and Ramp notes it isn’t clear the pace will continue.
One detail is worth flagging for anyone managing an AI budget. The top 1% don’t lock into a single provider. They mix and match, bouncing between multiple frontier models and platforms that give them access to cheaper open source models. That’s a deliberate cost strategy, not indecision. The power users have figured out that routing the right task to the right model keeps spend in check while still getting frontier performance where it counts.
What this means for you
A few practical takeaways from the numbers:
- Benchmark yourself honestly. If you’re spending around $11 per employee per month, you’re at the median. That’s a starting point, not a strategy. The $611 tier is where serious experimentation begins.
- Don’t marry one model. The biggest spenders treat models like interchangeable parts. Build your stack so you can swap providers and lean on open source for cheaper, high-volume tasks.
- Watch cost per outcome, not cost per seat. The leaders aren’t buying subscriptions. They’re funding agents and compute that do actual work. Measure what that spend produces.
- Expect the gap to widen. With AI-pilled firms growing spend double digits month over month, the distance between adopters and laggards is compounding fast.
A limitation to keep in mind: this index measures spending, not results. High spend doesn’t automatically mean high return, and Ramp itself isn’t sure the growth trend will continue. Money in is a signal of conviction, not proof of payoff.
The real story to watch is whether that 14.1% monthly growth keeps compounding or hits a ceiling as budgets tighten. For now, the AI-pilled are betting big and not slowing down. You can find the full breakdown in the original report at TechCrunch AI.