OpenAI and Anthropic face AI spend slowdown to 1.5%

OpenAI and Anthropic face a slowdown in customer spending, with Ramp data showing AI token spend rose 1.5% across August and September, TechCrunch reported November 6.

For Anthropic, which Atlas360 reported November 6 is preparing a Nasdaq IPO by late November, the Ramp report cuts against expectations for rapid revenue expansion: token spend grew 42% month on month in July.

AI tokens are units of text processed by large language models, and token spend is a proxy for paid usage. Ramp’s data shows companies processed more AI work even as spending growth slowed, TechCrunch reported.

“Almost the entire spend on AI is from venture-backed, high-growth tech companies,” Ara Kharazian, Ramp’s lead economist, told TechCrunch. “That’s fine as long as AI spend continues to rise, 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.”

Kharazian said the top 1% of AI-spending firms slipped below the roughly $10,000 per employee per month they reached during summer peaks. Median firms spend $12 to $15 per employee per month, while the top 10% spend $600 to $1,000, according to the TechCrunch report.

The gap matters because the heaviest users may be large enough to shape total revenue growth. “A slowdown among that cohort is sufficient to slow down revenue entirely, despite the fact that we’re seeing rising volumes,” Kharazian told TechCrunch.

October showed a rebound, with token spend up 8.9% from September 23 to October 23. Kharazian still said spending faces pressure because customers have more suppliers, cheaper models and better tools to limit usage.

Price cuts are one source of pressure. OpenAI cut application programming interface prices in half in September for GPT-6 Sol and GPT-6 Luna, TechCrunch reported. Anthropic has also reduced pricing for Fable 5.1 and Opus 5.5, and the company said in a recent filing that the move cut revenue growth by about $600 million.

Kharazian framed the issue as competition, not weaker demand. He pointed to enterprise offerings from Meta and Nvidia, open models from China and the United States, and Jev, a lightweight model for selecting among preset outputs that has gained traction on Ramp.

“For a lot of AI use cases, you just need a simple Yes/No,” Kharazian said. “So you take your already cheap models, you shift them over to Jev, and now your already cheap stuff gets even cheaper.”

Customers are also imposing tighter controls. Kharazian said advanced AI users are more likely to set defaults that avoid expensive models for long tasks, cap employee budgets and route work to cheaper alternatives.

That leaves Anthropic entering its planned listing with two competing narratives: rising model usage on one side, and evidence that buyers are reducing the revenue per unit of usage on the other. Atlas360 reported November 6 that Anthropic faces investor pressure to lower its IPO valuation from $2 trillion to $1.5 trillion, with a potential delay tied to broader technology-market volatility.