Businesses are using more AI but paying about the same for it. New data from the spend-management company Ramp shows token spending grew just 1.5% across August and September, compared with 42% growth in July alone, and the numbers landed on November 6 as reports said investors now value Anthropic’s planned IPO well below earlier expectations.
What the Ramp data shows
According to TechCrunch’s account of the report, token spend, meaning what companies pay for the units of text that AI models read and write, flatlined over late summer even though token volumes kept rising. More usage at a lower price per token produces flat revenue for the labs that sell it.
Ramp’s lead economist, Ara Kharazian, told TechCrunch that almost all AI spending comes from venture-backed, high-growth technology companies. The most intense users, the top 1% of firms, were spending close to $10,000 per employee at their summer peak and dipped slightly below that in August and September. Median companies spend $12 to $15 per employee per month, and the top 10% spend $600 to $1,000. Growth in those groups continued, but not enough to offset the top 1%.
October looked better: token spend rose 8.9% between September 23 and October 23. Kharazian still described the headwinds on spend growth as strong, and he has argued in his “Cracks in the AI Thesis” blog series that the era of hypergrowth may be over.
Why spending stalled while usage rose
Kharazian points to competition rather than weak demand. In September OpenAI cut API prices by half for GPT-6 Sol and GPT-6 Luna. Anthropic answered with significant price reductions on Fable 5.1 and Opus 5.5, a move the company said in a recent filing cut its revenue growth by about $600 million. Both companies also offer enterprise discounts meant to lock in customers.
Buyers are also getting more efficient. Kharazian says sophisticated users set defaults that block expensive models on long-running tasks, impose budgets on employees and use routers that push work from standard models to lighter ones. Meta and Nvidia now sell enterprise offerings built on cheap or open models, and a newer tool called Jev, which picks between a few predetermined outputs, is climbing the Ramp rankings for simple yes/no jobs.
How it fits with the IPO and funding news
The timing matters because of the capital markets. Bloomberg, as relayed by TechCrunch, reported that Anthropic’s potential valuation fell from roughly $2 trillion in early October to $1.2 to $1.3 trillion, and a separate TechCrunch brief says some prospective buyers see it as low as $1.1 trillion. We cannot independently verify those figures; they come from anonymous sources.
OpenAI, meanwhile, has reportedly raised $11 billion so far at a $1.25 trillion valuation, part of a round that could total about $20 billion, well short of the $30 billion first reported in late September. We covered the earlier stages in our stories on OpenAI’s lowered funding target and on reports that Anthropic could face a junk credit rating.
What it means for buyers and for the labs
For companies buying AI, a buyer’s market means leverage: price cuts, committed-spend discounts and credible cheaper alternatives. Anthropic had already started giving enterprise customers token spending reports, which help customers see where tokens go and, by extension, where they can cut.
For the labs, the question is whether volume growth can outrun price declines. Ramp’s data covers only its own customer base, which skews toward startups and technology firms, so it is a signal rather than a market census.
What to watch next
- Whether Ramp’s November data confirms the October re-acceleration or the late-summer flatline.
- Anthropic’s final IPO pricing and how it compares with the reported $1.1 to $2 trillion range.
- The close of OpenAI’s round, expected in the coming weeks, and its final size.
- Further API price cuts or new enterprise discounts from OpenAI, Anthropic, Meta and Nvidia.
- Spending data from non-tech sectors, which Ramp’s economist says is where growth must come from.
Key takeaways
- Ramp reports business AI token spend grew only 1.5% in August and September, versus 42% in July.
- Usage kept rising, so lower prices and smarter buying explain the flat spend.
- OpenAI halved GPT-6 Sol and Luna API prices in September; Anthropic cut Fable 5.1 and Opus 5.5 prices too.
- Valuation reports for Anthropic and OpenAI rest on anonymous sources and are unconfirmed.
Frequently asked questions
What did Ramp’s data say about AI spending?
Ramp found token spend grew 1.5% across August and September after growing 42% in July, even though the volume of tokens used kept increasing. October showed an 8.9% rise from September 23 to October 23.
Why is AI spending flat if usage is up?
Price cuts by OpenAI and Anthropic, enterprise discounts and more careful buying, such as budgets and routing work to cheaper models, mean companies pay less per task.
Does this mean the AI boom is over?
Not necessarily. Ramp’s data covers its own customers, mostly venture-backed tech firms, and October showed growth returning. It does raise questions about how fast lab revenue can grow.
