Business spending on AI tools slowed in August, according to data from Ramp covering about 70,000 companies. The share of Ramp customers paying for AI products reached 56%, up only 0.4% from the previous month.
The pattern is not entirely new. Ramp's AI index has shown similar late-summer pauses before, with growth often picking up again later in the year. This time, however, the slowdown comes as the AI industry continues to invest heavily in infrastructure and model development.
One notable shift came from the top tier of companies in Ramp's sample: AI spend per employee fell nearly 10% to $7,205. Ramp economist Ara Kharazian linked part of that drop to seasonal behavior, but also to lower token prices. As OpenAI and Anthropic reduced pricing, average token costs moved down to $0.68 per million tokens from a March peak of $1.15.
The data also suggests that many businesses are choosing older, less expensive models instead of the newest frontier releases. At the same time, only 6.4% of AI-spending companies used model-serving or inference platforms in August, showing steady but still limited expansion in that area.
Separate U.S. Census Bureau survey data indicates broader AI use remains more modest across businesses overall, which makes spending-based tracking an important signal for the market. For AI labs and cloud providers, the key question is whether lower prices will be matched by higher usage volume.
For companies adopting AI, cheaper access can widen experimentation and everyday use. Looking ahead, the balance between falling costs and rising adoption may shape the next phase of the AI economy.
