AI Usage Is Rising, but Enterprise Spending Is Losing Speed

Businesses are using more AI tokens, but they are not spending much more to do it. Ramp data shows token spending grew just 1.5% across August and September 2026, even as token volumes continued to rise. That marks a sharp change from July, when token spending climbed 42% from the previous month.

The gap between usage and revenue growth matters for OpenAI and Anthropic. Companies are sending more work through AI systems, but falling prices, enterprise discounts, and tighter controls are limiting how much model providers earn from that activity.

More tokens, less money per token

Ramp’s October data shows a rebound, but it did not close the gap between token volume and spending. In the week ending October 23, total token spending rose 8.9% from the comparable week ending September 25. Token volume climbed 37.1% over the same period.

At the same time, the blended price per million tokens across OpenAI and Anthropic fell from $0.76 to $0.56. The Ramp AI Index lists that as a 26.8% decline. OpenAI’s price fell 42.4%, while Anthropic’s dropped 15.7%.

That helps explain the simple version of the trend: businesses can use more AI without paying more for it. Notion co-founder Akshay Kothari summed up the pattern in a post using six words: “tokens go brrr. costs go ahhh.”

Ara Kharazian, Ramp’s lead economist, said the slowdown is not a demand problem. Model price reductions and competition have created what he called a buyer’s market, with OpenAI, Anthropic, Meta, and Nvidia offering businesses more ways to get AI work done at lower prices.

OpenAI cut API prices in half in September for GPT-6 Sol and GPT-6 Luna. Anthropic also reduced prices for Fable 5.1 and Opus 5.5. One Anthropic decision slashed revenue growth by about $600 million, while both companies have offered enterprise discounts designed to lock in spending.

A small group drives most of the spending

The biggest risk sits in the concentration of AI spending. Ramp’s October 14 analysis found that 1% of firms drove 80% of spending on OpenAI and Anthropic. It also identified two consecutive months of virtually flat token expenditure.

In September, median monthly AI spending per employee reached $9,568 within the top 1% of firms paying for AI. The figure was $813 in the top 10% and $15.70 for the median firm. Ramp also found that median firms spend $12 to $15 per employee each month, while the top 10% spend $600 to $1,000.

The top 1% had been reaching close to $10,000 per employee during the summer peaks, Kharazian said. Their slight pullback in August and September was enough to slow revenue growth, even though token volumes kept rising. Spending among median firms and the top 10% grew, but not enough to make up the difference.

“Almost the entire spend on AI is from venture-backed, high-growth tech companies,” Kharazian said. That concentration creates a wider risk for the AI industry because a change in spending by a small group can affect many companies at once.

“That’s fine as long as AI spend continues to rise,” Kharazian said, “but if it starts to slow down because of something in the market or some other factor, they will all start to slow down simultaneously, and that could lead to other cascading effects.”

Customers are learning how to control AI costs

Advanced customers are not simply using fewer AI tools. They are managing them with more care. Kharazian said companies are setting defaults that keep expensive models away from long-running tasks, imposing employee budgets, and using routers to move work from standard or light models to cheaper light models.

They are also choosing simpler tools for simpler jobs. Kharazian pointed to Jev, a lightweight tool for choosing between a few predetermined outputs. “For a lot of AI use cases, you just need a simple Yes/No,” he said. Moving already cheap work to cheaper systems can reduce spending without reducing the amount of work handled by AI.

“As you get really advanced in using AI, you get better at figuring out how to control it,” Kharazian said. “You can use more AI, but use it for less.” That is good news for customers, but it creates new questions about the revenue paths of model companies.

Anthropic’s share of token spending rose from 51.3% in the week ending September 25 to 60.5% in the week ending October 23. OpenAI’s share moved from 44.2% to 35.1%. The broader Ramp Index reached 55.9% of Ramp businesses in September, with 43.8% paying Anthropic and 40.6% paying OpenAI.

Growth claims now face a tougher test

Ramp’s data contrasts with a Menlo Ventures press release that put enterprise AI spending at $176.8 billion, up 5.4 times in one year. That figure rose from $33 billion last year to $177 billion this year, with autonomous agents described as changing how businesses use and pay for AI.

The two pictures can exist at the same time. Enterprise AI spending can grow across a year while the latest monthly data shows weaker growth among the firms that account for most model revenue. Kharazian said, “A lot of the assumptions that we had about the trajectory of AI spend are starting to break down.”

The pressure is already showing in company valuations. Anthropic’s potential valuation fell from $2 trillion in early October to $1.2 trillion to $1.3 trillion. OpenAI lowered its proposed valuation to $1.25 trillion from $1.4 trillion and cut its latest fundraising goals by about $10 billion.

AI coding emerged as the most commercially viable product about a year and a half ago, helping Anthropic and OpenAI become trillion-dollar companies. Some now see personal agents as the next major growth engine, with OpenAI, SpaceXAI, Meta, and startups such as Instinct releasing their own products.

Kharazian remains skeptical that personal agents will deliver nonstop growth. “The projection is that there will be new products being introduced that will drive further growth,” he said. “But based on what we can observe today and the products available and what is commercially advanced, AI spend has started to slow down.”

Stay connected via Google News

Follow us for the latest AI News