Corporate spending on artificial intelligence has hit its first real speed bump. New data from Ramp, the corporate card and spend-management company, shows that business spending on AI tokens grew just 1.5 percent across August and September combined — after jumping 42 percent in July.
The slowdown is not because companies are using less AI. Token volumes kept rising through the summer. The problem, for the AI labs at least, is that customers are getting the same work done for less money, as price cuts and smarter budgeting squeeze the revenue out of every token.
Volumes up, dollars flat
Ramp tracks payments across more than 70,000 US businesses, giving it one of the clearest windows into how companies actually spend on AI. Its lead economist, Ara Kharazian, told TechCrunch the stall is a pricing story, not a demand story.
"Almost the entire spend on AI is from venture-backed, high-growth tech companies," Kharazian said — the startups and scale-ups that were the earliest and heaviest adopters of frontier models.
The concentration is extreme. Just 1 percent of companies account for 80 percent of all business spending on OpenAI and Anthropic, according to Ramp, and that cohort is where the slowdown bit hardest. The most AI-heavy firms — the "AI-pilled" top 1 percent — spent slightly less in August and September, landing just below the roughly $10,000 per employee per month they reached at their summer peaks. That dip alone was enough to slow total revenue growth, Kharazian said, "despite the fact that we're seeing rising volumes."
Below the top tier, the numbers are far smaller. Median companies spend roughly $12 to $15 per employee per month on AI tokens, while the top 10 percent spend $600 to $1,000. Spending growth among those groups continued, but not enough to offset the stall at the top.
Price cuts are eating the labs' revenue
The main culprit is deflation. OpenAI cut API prices in half for two of its GPT-6 models in September, and Anthropic made big cuts of its own — Anthropic said in a filing that the price reductions slowed its revenue growth. Customers are also routing work to cheaper models and capping budgets, squeezing more output from every dollar.
October brought a modest rebound, with token spend up 8.9 percent from September 23 to October 23, according to Ramp. But Kharazian, who writes a series of analyses under the title "Cracks in the AI Thesis," said the headwinds on spending growth are strong — and that the era of hypergrowth may be over.
What it means for Anthropic's IPO
The timing is awkward for Anthropic, which is preparing a stock market listing that could be the largest in AI history. Bloomberg reported on Friday that buyer interest in the offering has softened, with the potential valuation sliding to $1.2 trillion to $1.3 trillion from the $2 trillion estimated in early October.
The Ramp data suggests investors may be wrong to assume AI companies will keep growing at breakneck speed. If the heaviest users of AI are already spending less per employee — even while using the technology more — the labs' revenue growth will depend on winning new customers rather than selling more to the ones they have. For an industry built on the promise of exponential growth, flat spending is a number Wall Street will not ignore.



