The top 1% of US companies now burn $7,500 per employee every month on AI tools, according to the latest Ramp AI Index. That is almost half the salary of an average software engineer, just on software. AI spending has become a real budget line.
The Gist
- Top 1% firms spend $7,500 per employee per month on AI tools
- Median US company still spends only $11.38 per employee on AI
- AI spending in top firms grew 14.1% per employee last month alone
Have ChatGPT Recap This Article
ChatGPTWhat “AI-pilled” companies actually look like
The Ramp AI Index is a monthly report from Ramp, a corporate spending platform. It measures how much American businesses pay for AI tools like ChatGPT, Claude, Gemini, and the dozens of newer apps built on top of them. The latest edition is the one that put the $7,500 figure on the table.
“AI-pilled” is the term Ramp uses for the very top of that ranking. These are the top 1% of US companies in terms of AI spending intensity per employee. They are not the biggest companies. They are the ones that buy the most software per head, which usually means startups, fast-moving tech firms, and small consulting shops betting hard on automation.
To get a sense of scale, the top 10% of companies spend on average around $611 per employee per month. The median company in the index still spends just $11.38. The gap between top performers and everyone else has never been this wide in any productivity software category before.
What changes inside an “AI-pilled” team is not just the budget line. Workflows get rewired around AI tools. A marketing analyst might lean on three or four different paid AI subscriptions every day, from research assistants to image generators to meeting note takers.
That kind of intensity also creates new problems for finance teams. We saw last week how Uber had to cap employee AI spending after a 4-month budget blowout. When power users can spend hundreds on AI in days, finance teams need new guardrails fast.
Why this number sounds wild but is not (yet) reckless
Here is the comparison that puts the $7,500 in perspective. The average software engineer in the US earns around $16,000 per month. So even at the top of AI spending, companies are paying less for AI than they pay for a single engineer. The math is not insane yet.
The same logic shows up at Nvidia. One Nvidia executive has publicly said that “the cost of compute is now greater than the cost of his employees” for certain teams. Translated, this means some Nvidia engineers cost less than the GPUs they use to do their job. AI is starting to redraw the entire cost structure of building software.
Mercor, an AI hiring startup, says it now spends more on tokens for its internal agents than on payroll. A token is a small piece of text that AI models read or write. Each AI request costs tokens, and tokens cost money. When you run hundreds of agents at once, those tokens add up quickly.
That kind of figure used to belong to research labs only. Now it is showing up in early-stage startups that are barely a year old. The shift is happening that quickly. AI is no longer a productivity perk. It is becoming the core cost of doing business in software.
Spending growth tells the same story. The top 1% of companies grew their AI spending per employee by 14.1% in the last month alone. If that pace holds for a year, the $7,500 figure becomes roughly $40,000 by next summer. That would match half an engineer’s annual salary.
Keep learning on AI Noobies:
- Claude Fable 5 builds a video game from one prompt
- Google AI Plus now $4.99: cheaper AI is here
- Siri AI: Apple’s chatbot will finally hold a conversation
What this means for your job and your tools
The short term reading is encouraging for workers who use AI well. As long as a top employee earns roughly twice what their AI tool stack costs, the human still wins the math. Companies want both. They want the employee, and they want the tool stack that makes the employee twice as fast.
The medium term is where things get tricky. If AI spending keeps growing at 14% per month inside the top 1%, the gap between AI cost and salary closes fast. Some routine jobs suddenly become a budget choice. A finance director might compare a $50,000 AI agent license to a $80,000 junior analyst and pick the one that does not call in sick.
For everyday users, the trend also acts as a pricing signal. The fact that companies are willing to spend so much per seat tells consumer AI providers that there is room to push prices up. Some startups already do. Others go the opposite way to grab market share before the consolidation wave.
On the consumer side, the price war is already loud. Google AI Plus just dropped to $4.99 per month, which puts pressure on every paid AI subscription. For workers, that means the personal AI tools you pay for yourself are getting cheaper, even as your employer pays much more per seat for similar features at scale.
The practical takeaway is simple. If your company barely spends on AI tools, the gap with the leaders in AI spending is already big. If you are in the top 10%, you should already feel the difference in how fast new projects ship. The Ramp AI Index is a useful mirror to check where you sit, and where you might want to push your team next.
Follow the story on AI Noobies.
About the Author
