Real Spending, Not Just Talk

Ramp publishes an AI Index based on actual money that companies spend, not on what executives say they plan to spend. The company looked at real purchases made by more than 70,000 U.S. businesses that use Ramp's corporate payment system. Every month, they measure how many of these companies are actively paying for AI tools and services.

This matters because surveys that ask executives what they intend to do often tell a misleading story. During AI hype cycles, executives tend to sound more enthusiastic than their actual spending reflects. During quieter periods, the spending can be happening steadily even when nobody talks about it much. Real transactions do not lie the same way a survey checkbox can.

The Biggest Spenders Are in a League of Their Own

The most surprising finding is what the top one percent of companies are spending on AI. These firms are paying $7,500 per employee each month just for AI tools and services. That works out to $90,000 per employee per year.

For context: a company with 50 employees spending at this rate would be putting $4.5 million a year toward AI tools. That is roughly what you would pay for several senior engineers in an expensive city — except this money goes to AI subscriptions and services instead of salaries.

What kinds of companies spend this much? Ramp does not name them. But at this spending level, these are not casual users. They are not just paying for ChatGPT accounts. Instead, they are likely running AI as a core part of how the business works — using it heavily through APIs, building custom AI systems, or using specialized AI software built for their industry.

One thing to keep in mind: this per-employee number does not tell us whether every worker in the company uses AI heavily, or just a small AI specialist team. A startup with ten AI experts could have the same spending-per-person as a larger company where one team has been given an AI budget. The numbers look the same even though the setup is different.

How Many Companies Use Paid AI Tools

The other key number is simpler: what percentage of the 70,000 companies Ramp tracks are paying for at least one AI tool in a given month. This number is more useful for understanding whether AI is becoming mainstream in business.

Most reporting on AI adoption focuses on huge companies or cloud giants making big announcements. But Ramp's data covers small and medium-sized businesses — the companies that often get left out of the headlines. When Ramp publishes its monthly percentage of companies using paid AI, it is showing whether AI is reaching ordinary businesses or staying concentrated among tech-forward early adopters.

We have seen this before with earlier technology shifts. In the mid-2000s, cloud computing adoption started the same way: a handful of startups jumped in immediately, then mid-sized tech-savvy companies followed within a couple of years, and finally, most conventional businesses took many years to make the move. The Ramp index, tracked month by month, is like an early signal for where AI sits on that same adoption curve — and whether the adoption is speeding up.

What This Data Misses

Ramp's data only tracks money flowing through their payment system. It does not include companies that use other payment processors, wire money directly to vendors, or run AI systems on their own computers and servers. Some large companies buy cloud computing capacity (from Amazon AWS, Microsoft Azure, or Google Cloud) through special pricing deals, and those charges might not show up in Ramp's data either.

The data also only covers the United States, so it does not tell us much about AI adoption elsewhere in the world. For people trying to understand the U.S. market specifically — investors, policymakers, or companies selling AI tools — a panel of 70,000 firms is substantial and reliable enough to be useful.

Why This Matters

The $7,500-per-employee-per-month figure tells us something important: for some organizations, AI has moved beyond being a nice productivity tool. It has become a core piece of how they operate, like electricity or networking. At that spending level, it is not a small line item — it is a major investment.

For most technology organizations that are not yet spending at that level, the real story to watch is the subscription penetration number: the percentage of the 70,000 companies that move from zero paid AI services to at least one. That month-to-month movement is where mainstream AI adoption is actually happening.

The Ramp AI Index does not tell us which AI tools are best, which companies will win or lose, or whether the money companies spend on AI actually makes them more productive. Those are harder questions that spending data alone cannot answer. What it does show is a clear, up-to-date picture of how widely commercial AI has spread across U.S. businesses — based on what companies are actually paying for, not what they say they plan to do.