Business AI adoption growth stalls in August, Ramp spending data shows

Adoption growth nearly flat in August

Business adoption of AI tools lost momentum in August, according to TechCrunch's reading of spending data from the payments company Ramp. Of the roughly 70,000 companies Ramp tracks, 56% paid for AI products during the month, an increase of just 0.4% over the previous month.

This is not the first time the indicator has flagged a pause. Ramp's AI index showed little or no growth in adoption between August and October last year, before growth resumed as the year closed. August is also a holiday month across much of the industry, which could account for the slowdown on its own.

Spending per employee fell sharply at the biggest spenders

Among the top 1% of firms in Ramp's sample, AI spending per employee dropped nearly 10% in August, to $7,205. Vacation schedules may explain part of that, but Ramp economist Ara Kharazian pointed to a second factor: falling token prices. As OpenAI and Anthropic have cut prices, the average cost of a million tokens has declined to $0.68, down from a 2026 peak of $1.15 in March.

According to TechCrunch, the data suggests the labs have not yet made up for those price cuts with growing volume. Many customers are also choosing older, cheaper models, such as OpenAI's ChatGPT 5.6-Terra and Anthropic's Sonnet, rather than the newest frontier releases. Employees at frontier labs have said much of the cost of training a model is recovered in the first weeks after its release, and slower uptake could threaten that pattern.

Open-weight inference remains a niche

For all the discussion of open-weight models as a threat to the frontier labs, TechCrunch reports that only 6.4% of businesses spending on AI used model-serving or inference platforms in August. That share is climbing steadily, but not quickly enough to drive the dynamics of broader business adoption.

Ramp's numbers may flatter the market

Ramp's customer base skews technical, so its figures could overstate adoption across the wider economy. An ongoing US Census Bureau survey of AI adoption, updated on 23 August, found that just 22% of businesses report using AI at all. Ramp's dataset is not necessarily representative of the market, but it is one of the few direct spending datasets available and may work as a leading indicator.

Why it matters

The enormous investment in AI infrastructure now underway at frontier labs and hyperscalers rests on the assumption that revenue will keep growing quickly enough to repay it. Usage has climbed steeply so far, driven largely by software engineers taking up agentic coding tools, and any slowdown in adoption would feed straight into slower revenue.

Kharazian frames the shift as a matter of where you sit: competition between OpenAI and Anthropic is making AI more accessible and cheaper for companies, but it is also pushing spend down at the top 1% of firms that the market had expected to power much of the industry's future growth. That pressure helps explain why AI labs are working hard to court non-technical users with co-working tools.

For businesses buying AI, falling prices are plainly good news. For model builders and hyperscalers with vast chip orders outstanding, one weak month may prove to be a summer blip, but it is exactly the kind of signal that the economics of the buildout depend on holding steady.