Enterprise spending on artificial intelligence is showing signs of slowing even as companies process more AI tokens, according to new data from financial technology platform Ramp. The findings suggest falling model prices and improved cost management are allowing businesses to increase AI usage without corresponding increases in spending.
According to Ramp’s AI Index, token spending grew just 1.5% across August and September 2026, compared with a 42% monthly increase in July. Although spending growth recovered to 8.9% between September 23 and October 23, the data indicates a substantial slowdown from earlier growth rates.
AI Spending Remains Concentrated Among Heavy Users
Ramp lead economist Ara Kharazian said much of the industry’s revenue depends on a relatively small number of venture-backed technology companies that spend heavily on AI services. The top 1% of companies in Ramp’s dataset had been spending approximately $10,000 per employee each month before recording a modest decline in August and September.
Other businesses have continued increasing their spending, but their contributions remain comparatively small. Median companies spend approximately $12 to $15 per employee monthly, while those in the top 10% spend between $600 and $1,000.
Kharazian questioned whether businesses outside technology-intensive industries would eventually reach the spending levels recorded among the largest AI users. Although software developers and financial institutions have adopted computationally demanding AI applications, industries such as healthcare services and automotive retail may have different requirements.
Lower Model Prices Reduce Enterprise AI Costs
Ramp attributed much of the slowdown to stronger competition among AI model providers and companies becoming more efficient in managing workloads. In September, OpenAI reduced API pricing for its GPT-6 Sol and GPT-6 Luna models by approximately 50%, while Anthropic also introduced substantial pricing reductions.
Anthropic reportedly disclosed that lower prices for its Fable 5.1 and Opus 5.5 models reduced expected revenue growth by approximately $600 million. Both companies have also offered enterprise discounts to secure long-term customers, potentially limiting near-term revenue expansion.
Businesses are increasingly routing simpler requests to inexpensive AI models, establishing employee spending limits, and avoiding unnecessary use of advanced models. Kharazian also identified growing competition from Meta, Nvidia, open-model developers, and lightweight systems such as Jev.
Slowing Growth Raises Questions About AI Company Valuations
The spending data comes as Anthropic prepares for a potential public offering and investors reassess revenue expectations for leading AI developers. Bloomberg reported that Anthropic’s estimated IPO valuation had fallen from approximately $2 trillion in early October to between $1.2 trillion and $1.3 trillion, while OpenAI reportedly lowered its proposed valuation from $1.4 trillion to $1.25 trillion.
Kharazian said new AI applications could create additional demand, although their commercial potential remains uncertain. He questioned whether consumer AI agents could generate spending comparable to coding assistants, which have become one of the industry’s most commercially successful applications.
Featured image credits: Magnific.com
For more stories like it, click the +Follow button at the top of this page to follow us.