Editorial illustration for Ramp Data Shows AI Spending Stalling as Anthropic IPO Nears
AI Spending Stalls Despite Rising Token Usage
Ramp Data Shows AI Spending Stalling as Anthropic IPO Nears
Something strange is happening in the numbers behind the AI boom. Companies are feeding more tokens through OpenAI and Anthropic's models than ever, but the dollars aren't keeping pace. Payments platform Ramp tracked corporate AI spending through August and September and found growth nearly flat at 1.5%, a sharp drop from July's 42% month-over-month jump.
Usage went up. Revenue didn't follow.
The timing is awkward. Anthropic is heading toward an IPO built on the assumption that enterprise AI spending climbs in a straight line, and Ramp's numbers suggest that line just bent. Cheaper models and buyers who've gotten better at negotiating are likely squeezing margins across the industry, even as more employees and more workflows touch AI tools every month.
Ramp's lead economist, Ara Kharazian, has been watching where that spending actually concentrates, and it's not spread evenly. A small slice of venture-backed, high-growth tech companies accounts for almost all of it. That concentration is exactly what makes the recent stall worth paying attention to.
Why this matters
Ramp's numbers land right as Anthropic's bankers are presumably building growth charts for its IPO roadshow, and the gap between usage and spend is the kind of thing public-market investors will poke at. A 1.5% rise in token spend over two months, against 42% growth in July alone, is not a rounding error. It reads like price competition and procurement discipline finally catching up with the hype.
For founders building on top of OpenAI or Anthropic APIs, this is good news on margins: your customers aren't going to tolerate open-ended AI line items forever, so vendors have to compete on price. For researchers and developers, it's a reminder that token volume isn't revenue, and growth narratives built on usage charts can mislead. The dentist's office or car dealership example in the piece is the real tell: horizontal AI demand has a ceiling that software and finance don't.
Anyone valuing Anthropic, or betting on sustained AI infrastructure spend broadly, should be watching whether this flattening is a blip or the start of a real repricing.
Related Reading
Browse all Anthropic news →
Business & Startups Nordic pilot adds Gemini for Education, NotebookLM to boost AI literacy
Business & Startups Kling launches Video O1, all-in-one model with MVL bridge using transformer
Business & Startups DeepSeek Seeks More Capital Weeks After USD 7B Funding Round
Related Topic Claude gains shared context in Excel, PowerPoint; Microsoft adds Copilot Cowork
AI Tools & Apps
Related Topic Anthropic: Hackers Stole Claude Tokens Via Compromised Session Key
LLMs & Generative AI
From Archives OpenAI Could Be IPO Ready in 2027, Says President Greg Brockman
From Archives Perplexity launches Teammates AI agents for work, running 24/7 like OpenAI's Dots
From Archives Anthropic adds Claude bank access via Plaid to track spending
LLMs & Generative AI
From Archives Inside Anthropic's AI speedup and OpenAI's faster Images 2.6 model
LLMs & Generative AI
Common Questions Answered
What did Ramp's data reveal about the disconnect between AI token usage and corporate spending?
Ramp tracked corporate AI spending through August and September and found that while companies were using more tokens through OpenAI and Anthropic's models than ever before, the dollar spending growth nearly flatlined at just 1.5% month-over-month. This represented a dramatic slowdown compared to July's 42% month-over-month growth, indicating that increased usage was not translating into proportional revenue increases.
How might Ramp's AI spending data impact Anthropic's upcoming IPO?
The timing of Ramp's report is significant as Anthropic approaches its IPO, which is built on assumptions about enterprise AI spending growth. Public-market investors will likely scrutinize the gap between token usage and actual spending, as a 1.5% rise in token spend over two months is substantially lower than the 42% growth seen in July alone, potentially raising concerns about the company's revenue trajectory.
What factors are causing the slowdown in AI spending despite higher token volumes?
According to the report, price competition and procurement discipline from increasingly savvy customers appear to be capping revenue growth for AI providers like OpenAI and Anthropic. This suggests that as companies become more experienced with AI tools, they are negotiating better pricing and managing their spending more carefully, even as they increase their actual usage of AI models.
What is the significance of the 42% growth drop from July to the 1.5% growth in August-September?
The sharp decline from 42% month-over-month growth in July to just 1.5% growth over August and September is not a minor fluctuation but rather a significant indicator that the AI spending boom may be plateauing. This dramatic slowdown suggests that the initial hype-driven spending surge is being replaced by more measured, price-conscious purchasing decisions from enterprises.
Further Reading
- Ramp data shows AI spending stalling just as Anthropic nears its IPO - TechCrunch
- October 2026 Ramp AI Index: Cracks in the AI thesis, part 3 - Ramp
- Anthropic's IPO: A close look at AI frontrunner's rapid growth, spiraling costs - Reuters
- Anthropic's path to the Nasdaq just got a lot bumpier - CNBC
- AI Spend per Token Fell 21% as Google Commitments Hit $924B - FourWeekMBA
