AI Spending Per Employee Drops at Leading Firms in August

Corporate adoption of artificial intelligence tools decelerated in August, according to new spending data from Ramp covering approximately 70,000 companies. The fintech firm’s latest index indicates that 56% of its customers paid for AI products during the month, a modest increase of just 0.4% from July.

While this marks the second time Ramp has observed a plateau in adoption metrics during the late summer and early autumn, analysts note that the sheer scale of current AI infrastructure investment makes even minor dips noteworthy. Frontier laboratories and hyperscalers have committed billions to building capacity based on projections of substantial future revenue. If user adoption, particularly among software engineers using agentic coding tools, begins to stagnate, revenue growth may follow suit.

Ramp’s figures may overstate broader market penetration due to the company’s technology-heavy client base. In contrast, a US Census Bureau survey updated on August 23 reported that only 22% of all businesses are currently using AI. Despite this discrepancy, Ramp’s dataset remains one of the few direct indicators of corporate AI expenditure.

The August slowdown could partly reflect seasonal vacations, but Ramp economist Ara Kharazian highlighted a more significant trend: a nearly 10% drop in AI spending per employee among the top 1% of firms, falling to an average of $7,205. This decline coincides with a sharp reduction in token costs, which have dropped to $0.68 per million tokens from a 2026 peak of $1.15 recorded in March, driven by price cuts from OpenAI and Anthropic.

Many customers are opting for older, less expensive models such as OpenAI’s ChatGPT 5.6-Terra and Anthropic’s Sonnet rather than newer frontier releases. Industry insiders have noted that model developers typically recoup training costs within weeks of a launch; slower uptake could undermine this financial dynamic. Additionally, only 6.4% of AI-spending businesses utilized model-serving or inference platforms in August, suggesting that open-weight models have not yet significantly disrupted the market.

“We are showing that competition between OpenAI and Anthropic is making AI more accessible, and also driving the price down for companies,” Kharazian said. “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.”

This shift has prompted AI labs to prioritize attracting non-technical users through co-working tools. For model builders and hyperscalers with massive chip orders, the data may appear concerning. However, Kharazian argued that lower costs benefit companies actively using AI, noting, “It depends on who you are in the market.”