American businesses spent more on artificial intelligence in October than in any month before it — yet the pace of growth has slowed to a crawl. The Ramp AI Index, published on November 11, shows corporate AI spending rose just 1.65% from September to a new record, a sharp comedown from the boom months of the summer.

The index tracks payments from more than 85,000 US companies that use Ramp's corporate cards and bill-pay platform, giving it a real-world view of what businesses actually pay for AI tools, rather than what vendors report as revenue. Its message for October is that usage is still climbing, but spending is not keeping up.

Prices are falling faster than usage is rising

The main reason is price. The average cost of processing a million text tokens fell 21% during October, as OpenAI and Anthropic cut prices and competed harder for business customers. Token volumes kept growing far faster than the money spent on them — a pattern Ramp's lead economist, Ara Kharazian, has been warning about in his "Cracks in the AI Thesis" series.

"The risks I identified in the Cracks series remain," Kharazian wrote in the November update, arguing that the slowdown is not a demand problem. Businesses are using more AI than ever; they are simply paying less for each unit of it.

The split between the two leading vendors underlines the point. Business spending on OpenAI grew 2.1% month over month in October, while spending on Anthropic grew just 0.5%. Even so, Anthropic still leads on paid adoption: 44.8% of the businesses Ramp tracks now pay Anthropic, up 1.1 points, against 40.6% for OpenAI, which was flat. SpaceX's AI offering was the fastest climber, reaching 6.3% paid adoption, up 2.2 points.

Adoption is broadening, but spending stays concentrated

Overall, 56.8% of the businesses on Ramp's platform now pay for at least one AI service. But the money remains highly concentrated. The median company spent just $16.32 per employee per month on AI in October, while the top 1% of spenders paid a median of $10,200 per employee per month, up 9% from September.

Where the money goes is also shifting. Software development teams still account for 72% of all tracked AI spending, though their share has fallen from the summer highs. The fastest-growing users are customer support and customer success teams, which drove 13.3% of classified AI spending in October — up from just 0.8% in June — as companies put AI agents in front of customers.

What it means for the AI boom

The October numbers extend a trend that began in late summer. Token spending surged 42% in July, then grew only 1.5% across August and September combined. October brought a partial recovery in token spending, up 8.9% in the week ending October 23, while token volumes jumped 37.1% over the same stretch.

For the AI industry, the gap matters. If businesses keep consuming far more AI while paying only slightly more for it, the model makers' revenue growth will depend on finding new customers and new uses — not on charging more for the same work. Ramp cautions that its data covers only its own customer base, which skews toward larger, tech-savvy firms, and misses free tools and personal accounts used for work. Even with that caveat, the October print is the clearest sign yet that the AI spending boom is entering a slower, more price-sensitive phase.