Skip to content
AI

AI token spending slowed to 1.5% as model builders trim valuations

Corporate card data from Ramp reveals a sharp slowdown in model usage growth, coinciding with trimmed valuations for Anthropic and OpenAI.

OSH Desk1 min read

Corporate spending growth on artificial intelligence models decelerated sharply at the end of the summer, according to corporate card data published by Ramp [0]. Outlays on AI tokens expanded by 1.5% across August and September combined, marking a steep decline from 42% growth recorded in July [1].

The moderation in enterprise consumption coincides with price cuts and lower valuation expectations across the top foundation model creators [1]. Anthropic marked down pricing for its Fable 5.1 and Opus 5.5 models, disclosing in a filing that the price reductions trimmed its revenue growth by roughly $600 million [1]. Anthropic's expected valuation has concurrently slipped to between $1.2 trillion and $1.3 trillion [1]. Rival OpenAI likewise reduced its proposed valuation target from $1.4 trillion to $1.25 trillion [1].

The figures suggest that demand is flattening after months of aggressive experimentation, even as model developers compete on unit costs [1]. "Almost the entire spend on AI is from venture-backed, high-growth tech companies," noted Ramp Lead Economist Ara Kharazian [1].

Questions and answers

How fast did corporate AI token spending expand recently?

Data from corporate spend management platform Ramp shows that spending grew by 1.5% during August and September combined, down from 42% in July.

What valuation changes have Anthropic and OpenAI faced?

Anthropic's prospective valuation has adjusted to between $1.2 trillion and $1.3 trillion, while OpenAI lowered its proposed valuation from $1.4 trillion to $1.25 trillion.

TopicsFundraising playbookCapital efficiencyFoundation models

Companies & investors in this story

Sources

  1. 1.Ramp data shows AI spending stalling just as Anthropic nears its IPO — TechCrunch AI, Nov 6, 2026
  2. 2.AI may be helping Anthropic make two years of progress in one, METR estimates — EdTech Innovation Hub — edtechinnovationhub.com

Researched and drafted with AI tools under OSH editorial rules. Every fact links to its source. Editorial policy · Suggest a correction