Anthropic IPO Meets Stalling AI Spend: The Risk Is Who Pays

Usage Is Rising. Spend Is Not.
Business AI spending stalled in late summer 2026 even as usage kept climbing: according to TechCrunch's November 6 report on Ramp's data, token spend only grew 1.5% across August and September, after growing 42% from the previous month in July. The timing is awkward. Anthropic, which announced that it had confidentially submitted a draft Form S-1 on June 1, 2026, launched the roadshow for its proposed initial public offering on November 10, and CNBC reported that pricing was expected the evening of November 18.
But here is the detail that matters more than the growth rate. Ramp's November 2026 AI Index reports that 1% of businesses drive 80% of the revenues to OpenAI and Anthropic, and that this top 1% is "largely made up of VC-backed companies in tech, finance, and technical services." The question about model-company revenue is also a question about the venture cycle that funds their largest customers.
Ramp's lead economist Ara Kharazian stated the consequence directly to TechCrunch: "Almost the entire spend on AI is from venture-backed, high-growth tech companies." If that spend slows, he said, "they will all start to slow down simultaneously, and that could lead to other cascading effects."
The Distinction That Matters: Token Volume Is Not Revenue
Usage and spend are two different numbers, and Ramp's data shows them moving apart. The November index states that "token volume growth continues to outpace token spend as AI gets cheaper and cheaper," and that the average price per million tokens fell 21% over the last month.
Prices are falling. Ramp's September 2026 AI Index found that the effective price per million tokens declined 41% to $0.68, down from the 2026 peak of $1.15 in March. TechCrunch reports that in September OpenAI cut API prices in half for GPT-6 Sol and GPT-6 Luna, and that Anthropic made significant reductions to pricing for Fable 5.1 and Opus 5.5.
Buyers are getting disciplined. Kharazian told TechCrunch that advanced users are more likely to set defaults that keep expensive models away from long-running tasks, impose budgets on employees, and route work to cheaper light models. His summary: "You can use more AI, but use it for less."
Adoption leadership is not spend growth. In October, Anthropic adoption rose to 44.8% of Ramp businesses against 40.6% for OpenAI. Yet Ramp reports that October token spend grew 2.1% month over month for OpenAI and only 0.5% for Anthropic. A company can win customers and see slower spend growth in the same month.
Capital Formation: A Revenue Base Underwritten by Venture Money
Overall AI spending rose 1.65% in October, according to Ramp, and TechCrunch reports token spend was up 8.9% from September 23 to October 23. Ramp's October figure shows modest growth, not a recovery to July's pace. Ramp also shows where the money comes from: the top 1% of firms spend a median of about $10,200 per employee per month, the top 10% spend $801, and the median firm spends $16.32.
That gap matters for capital formation: the spend driving model-company revenue sits with the top 1% of firms, which Ramp describes as largely VC-backed, so model-company revenue moves with the venture funding available to those buyers. Engineering teams drove 72% of classified AI spend in October, per Ramp, while customer success and support rose to 13.3% from 0.8% in June. Kharazian questioned whether "a doctor's office or a car dealership will need to spend that much on AI." Ramp reports that AI spend is failing to proliferate outside of tech, finance, and technical services.
Reported valuations have also fallen. TechCrunch, citing Bloomberg, reported that discussed valuations for Anthropic fell from $2 trillion in early October to $1.2 trillion to $1.3 trillion, and that OpenAI lowered its proposed valuation to $1.25 trillion from $1.4 trillion.
Ramp's own data carries caveats worth stating. Its September index notes that the top 1% estimate "is more volatile" because the group is small, that results "are also subject to change," and that August declines may partly reflect engineers taking summer vacation. The trend is a signal, not a settled fact.
What Founders and Fund Managers Should Decide This Quarter
The read-through is not limited to Anthropic and OpenAI. Any company whose revenue depends on venture-backed tech buyers, or whose margins depend on model pricing, carries a version of the same exposure.
For founders building on model APIs
- Rebuild unit economics on current prices. A 41% decline in effective token prices between March and September changes cost of goods sold and should change the projections your board approves.
- Separate usage from revenue in investor materials. Fundraising decks that present token growth as a revenue proxy describe two numbers Ramp shows diverging.
- Map customer concentration. If most of your revenue comes from venture-backed tech, finance, or technical services companies, your pipeline shares the correlation Kharazian describes.
For GPs and investment committees
- Test portfolio correlation. Holdings that sell AI tools to other portfolio-style companies may slow at the same time.
- Treat IPO valuation figures as reported, not settled. CNBC reported that Anthropic's offering was expected to price the evening of November 18 and that the amended S-1 set a range of $70 to $78 per share.
- Document comparables used in marks. Valuation committees referencing AI public comparables should record which figures were reports and which were transactions.
Key Takeaways
These points stand on their own and can be shared with a board or investment committee.
- AI usage growth is no longer a reliable proxy for AI revenue growth. Ramp's data shows token spend grew only 1.5% across August and September 2026 while volumes kept rising, according to TechCrunch.
- In Ramp's data, OpenAI and Anthropic revenue is concentrated in venture-backed buyers. Ramp reports 1% of businesses drive 80% of the revenues to OpenAI and Anthropic, largely VC-backed firms in tech, finance, and technical services.
- Falling model prices change startup unit economics. Ramp's effective price per million tokens fell 41% to $0.68 by September from a March 2026 peak of $1.15.
- Anthropic's proposed IPO was expected to price on November 18. CNBC reported pricing was expected that evening, so valuation figures in circulation remain reported estimates.
What This Means for Your Next Raise
Ramp's data shows AI usage and AI revenue moving apart, with the revenue resting on a narrow, venture-backed customer base. Founders raising on AI-driven projections generally need their offering materials and board reporting reviewed before the next round closes, so that usage metrics, pricing assumptions, and customer concentration are described no more broadly than the data supports. To discuss that review with FinTech Law's startup counsel, you can start that conversation with our team here.
This blog post is for informational purposes only and does not constitute legal advice. No attorney-client relationship is formed by reading this content. If you need legal advice, please contact a qualified attorney.
FinTech Law fixed engagements: FinTech Startup Regulatory Path ($1,500) · Emerging Manager Launch Kit ($7,500) · Regulatory Path Outline ($7,500) · RIA/ERA Registration ($8,500) · Private Fund Formation (from ~$18,000) See the catalog.
Fractional General Counsel · from $7,500/qtr