Business spending on artificial intelligence slowed in August, with 56% of Ramp customers paying for these products, a rise of just 0.4% from the previous month.
This dip follows a similar pattern last year, where adoption stalled between August and October before recovering toward year-end. However, the sheer scale of investment in AI infrastructure by major labs and cloud providers relies on strong revenue returns. So far, usage has climbed sharply as software engineers adopted agentic coding tools. A slowdown there could drag revenue down with it.
Ramp’s figures likely overstate overall market uptake because its clients are heavily tech-focused. An ongoing US Census Bureau survey updated on August 23 shows only 22% of businesses report using AI. Ramp’s data is not necessarily representative, but it remains one of the few direct spending datasets available and may act as a leading indicator.
The timing matters. August is when much of the industry takes vacation, which could explain the quiet period. Yet, Ramp economist Ara Kharazian points to other warning signs for companies that rely on token spend.
The numbers
There has been a sharp drop in AI spend per employee among the top 1% of firms in his sample. That figure fell nearly 10% to $7,205. While part of this may be the vacation effect, it also reflects falling token costs. As OpenAI and Anthropic reduced prices, the average cost per million tokens dropped to $0.68. This is down from the 2026 peak of $1.15 recorded in March.
The data suggests that labs have not yet offset these price cuts with increased volume. The same pricing incentives are leading many customers to choose older, cheaper models like OpenAI’s ChatGPT 5.6-Terra and Anthropic’s Sonnet rather than the more powerful frontier releases. Employees at frontier labs have noted that much of the cost of training is recouped in the first weeks of a new model’s release. Slower adoption could threaten that dynamic.
Despite concerns about open-weight models threatening the frontier labs, only 6.4% of AI-spending businesses used model-serving or inference platforms in August. That share is growing steadily but not fast enough to drive broader business adoption.
“We are showing that competition between OpenAI and Anthropic is making AI more accessible, and also driving the price down for companies—and not just driving the price down, but driving spend down at the top 1% of companies that previously the market was expecting to drive much of the growth going forward,” Kharazian said.
This trend also helps explain the focus at AI labs on winning over non-technical users for AI co-working tools.
This data point could be a bad sign for model builders or hyperscalers with hundreds of billions of chips on order. But Kharazian notes it depends on who you are in the market. If your company is using AI, it is still a good position to be in.
