Companies are using more AI than ever, but Ramp’s latest data suggests they are not spending proportionally more on it. The payment company’s report shows AI token spending flattened in late summer even as usage volumes kept rising, a sign that price cuts, discounts and more disciplined customers may be capping revenue growth for major model providers such as OpenAI and Anthropic.
Ramp sees AI spending lose momentum
According to Ramp, token spend grew just 1.5% across August and September. That was a sharp slowdown from July, when spending rose 42% from the previous month. October brought a partial rebound, with token spend up 8.9% from September 23 to October 23, but Ramp’s lead economist Ara Kharazian said the broader trend still points to weaker spending growth.
The key detail is that usage is still increasing. Kharazian said the slowdown is not being driven by demand falling off. Instead, companies are getting more efficient about how they use AI, while vendors are cutting prices in a competitive market.
Why the data matters for Anthropic
The timing is awkward for Anthropic, which is preparing for an IPO. Ramp’s findings suggest investors should be careful about assuming AI companies can keep compounding revenue at the breakneck pace seen earlier in the boom. If customers continue to use more tokens but spend less per unit, the economics that powered the industry’s rapid rise could become harder to sustain.
“Almost the entire spend on AI is from venture-backed, high-growth tech companies,” Kharazian told TechCrunch. “That’s fine as long as AI spend continues to rise, but if it starts to slow down because of something in the market or some other factor, they will all start to slow down simultaneously, and that could lead to other cascading effects.”
He said the top 1% of firms — the most “AI-pilled” users — have already “stuttered slightly” in August and September, slipping just below the $10,000 per employee level they reached at summer peaks. In his view, even a slowdown among that small group could be enough to pressure overall revenue growth, despite rising usage volumes.
The market is becoming a buyer’s market
Ramp’s data shows spending growth among median firms and the top 10% of users, but not enough to offset the cooling at the very top. Median firms spend about $12 to $15 per employee each month, while the top 10% spend roughly $600 to $1,000. That leaves the biggest AI customers doing most of the heavy lifting for model-company revenue.
Kharazian said the current market looks increasingly like a buyer’s market, especially after price cuts from leading vendors. In September, OpenAI cut API prices in half for GPT-6 Sol and GPT-6 Luna. Anthropic responded with significant reductions to pricing for Fable 5.1 and Opus 5.5, a move the company said in a recent filing reduced revenue growth by about $600 million.
Both companies have also been offering enterprise discounts aimed at locking in usage, but Kharazian said those concessions are hitting revenue targets rather than simply widening adoption.
More competition, cheaper models
He argued that the competitive environment is very different from the old Big Tech playbook, where a company could subsidize usage until rivals disappeared and prices later rose. In AI, he said, that kind of lock-in is harder because there are too many alternatives.
Among the pressures he pointed to:
- Meta and Nvidia offering enterprise products with inexpensive or open source models
- Other open model providers from China and the United States
- A new tool called Jev, which helps choose between a few predetermined outputs and is appearing more often among Ramp’s top performers
For some use cases, Kharazian said, customers do not need a complex model at all. A simple yes-or-no decision can often be handled by cheaper systems, and routing work to lighter models can reduce costs even further.
Customers are also getting more efficient
The slowdown is not only about lower list prices. Companies are learning how to control AI usage more carefully, which further squeezes spend. Kharazian said advanced users are more likely to set defaults that prevent employees from using expensive models for long-running tasks, impose budgets, and route work from standard models to even cheaper ones.
That shift can be good for customers, but less so for model makers. As Kharazian put it, “You can use more AI, but use it for less.”
Notion co-founder Akshay Kothari recently posted internal data that echoed Ramp’s findings, summing up the trend with a succinct line: “tokens go brrr. costs go ahhh.”
Valuation pressure is already showing up
The spending data arrives as investor sentiment around frontier AI companies appears to be cooling. Bloomberg reported Friday that Anthropic’s potential valuation dropped from the $2 trillion figure estimated in early October to about $1.2 trillion to $1.3 trillion. The company is also reportedly expected to receive a high “junk” rating from credit agencies.
OpenAI has also reportedly lowered its proposed valuation to $1.25 trillion from $1.4 trillion and cut its latest fundraising goals by about $10 billion. Together, those moves suggest the market is asking tougher questions about how quickly AI revenues can grow from here.
What could still revive spending
Kharazian does see one possible path back to faster growth: new AI products that prove commercially important in the way coding agents did over the past year and a half. That class of tools helped turn Anthropic and OpenAI into trillion-dollar companies in investor eyes.
Some investors believe personal agents could be the next breakout category. OpenAI, SpaceXAI and Meta have all released personal agents, alongside startups such as Instinct. Kharazian is skeptical that those products will match the economic impact of coding tools, at least at current levels of maturity.
For now, his conclusion is cautious. “A lot of the assumptions that we had about the trajectory of AI spend are starting to break down,” he said. “The projection is that there will be new products being introduced that will drive further growth. But based on what we can observe today and the products available and what is commercially advanced, AI spend has started to slow down.”
Source: Original report
Was this helpful?
Explore more: AI Automation Services More Business & Startups Tech News
Last Modified: November 6, 2026 at 10:32 pm
0 views
