Companies are using more artificial intelligence, but their spending is no longer increasing at the same pace. Data from Ramp shows AI-related spending grew just 1.5% in August and September, compared with 42% month-on-month growth in July, even as token usage continued to rise.
Spending picked up in October, increasing 8.9% between September 23 and October 23. Ramp's data suggests that the overall trend is being shaped not by falling demand, but by lower prices and companies becoming more selective about which models they use.
Lower costs, smarter usage
Price reductions from OpenAI and Anthropic, along with a wider choice of lower-cost and open models, are making AI cheaper to deploy. Businesses are also managing usage more closely, directing tasks to less expensive models and setting budgets for employees.
Spending varies considerably between companies. The median firm in Ramp's data spends about $12-$15 per employee each month on AI, while the top 10% spend roughly $600-$1,000. The heaviest-spending 1% had been approaching $10,000 per employee during the summer, but their spending eased slightly in August and September.
The shift comes as Anthropic prepares for a potential IPO, prompting fresh attention to how AI companies can sustain revenue growth when usage expands faster than customer spending. Ramp's findings point to an increasingly cost-conscious market, where access to AI is spreading while the expense of each task declines.
As AI tools become more efficient and affordable, future growth may depend on new practical uses as much as on rising usage--potentially broadening adoption across more industries.
