AI use rises as spending growth cools, Ramp data shows

Ramp’s lead economist, Ara Kharazian, told TechCrunch that the slowdown is visible among the top 1% of companies by AI spending. Their per-employee spend slipped just below the nearly $10,000 monthly levels reached at summer peaks. That pullback, he said, can slow revenue growth despite rising token volumes. Median firms spend $12–$15 per employee each month; the top 10% spend $600–$1,000.

Kharazian said the figures do not point to weakening demand so much as a more competitive market. OpenAI cut API prices in half in September for GPT-6 Sol and GPT-6 Luna. Anthropic reduced prices for Fable 5.1 and Opus 5.5, saying in a recent filing that the cuts lowered revenue growth by about $600 million. Both companies have also offered enterprise discounts. “It very much [has] become a buyer’s market,” Kharazian said.

Cheaper alternatives and more efficient use are adding to the pressure. Kharazian cited enterprise offerings from Meta and Nvidia, open models from companies in China and the United States, and Jev, a lightweight tool that routes requests among predetermined outputs. Businesses are also learning to limit use of expensive models, set employee budgets and shift tasks to cheaper models. Notion co-founder Akshay Kothari shared internal data that echoed the pattern: “tokens go brr. Costs go ahhh.”

Spending did pick up in October: token spend increased 8.9% from September 23 to October 23. But Kharazian said current products do not yet show that demand will return to breakneck growth. He is skeptical that personal agents will match coding tools’ economic impact. The spending outlook is drawing added scrutiny as Anthropic prepares for a possible IPO: Bloomberg reported its potential valuation fell from $2 trillion in early October to $1.2 trillion–$1.3 trillion. OpenAI also cut its proposed valuation to $1.25 trillion from $1.4 trillion and reduced its latest fundraising goals by about $10 billion.