In the third quarter of 2026, the highest-spending 1% of AI users in US businesses curbed their per-employee AI spending after a summer peak, coinciding with a deceleration in Anthropic’s quarterly revenue growth as revealed in their amended S-1 filing.
- Top 1% of AI spenders reduced per-employee spend by 3.1% in September from July’s peak.
- Anthropic’s Q3 revenue growth slowed, adding 21% less than Q2.
- AI adoption among US businesses rose slightly, but new adopters spend far less than top-tier firms.
What happened
Ramp’s latest AI Index reveals that the top 1% of US businesses on its platform spent a median of $9,568 per employee per month on AI in September 2026, down 3.1% from their July peak of $9,873. August had seen a sharper 13% decline from July before a partial recovery in September. Meanwhile, the broader top 10% and the median firms increased AI spending during this period, signaling growth outside the heaviest spenders.
Anthropic’s amended S-1 filing corroborated this trend with quarterly revenue additions slowing in Q3. The company’s average monthly revenue growth dropped from approximately $760 million in Q2 to around $580 million in Q3, resulting in a 21% decrease in the dollar amount added. Despite a $16.96 billion revenue figure in Q3, growth momentum softened after July’s run rate peak of about $65 billion.
Why it matters
The stalling of heavy AI spenders and moderation in Anthropic’s growth highlights the concentration risks within AI enterprise spending. Ramp data shows that 80% of spending on Anthropic and OpenAI is driven by just 1% of customers, demonstrating a high dependency on a small group of extremely heavy users. A slowdown among these customers can significantly impact vendor revenues regardless of broader adoption trends.
Price reductions for top AI model services also played a large role, with Anthropic’s Claude Opus 5.5 model lowering costs per task by approximately 40%. This makes AI usage more affordable but reduces the dollar value of spend for heavy users unless they substantially increase workload. The combination of price cuts and less expansion of AI workloads among top spenders resulted in subdued dollar growth in the third quarter.
What to watch next
Looking ahead, it will be important to track if top-tier firms increase their AI workloads enough to offset price declines or if the spending plateau persists. Meanwhile, continued adoption growth among the wider base of smaller spenders could drive more steady aggregate growth despite less volatility in the top 1%. Ramp’s data shows AI adoption by US businesses rising to 55.9% in September from 55.3% in August.
Anthropic’s revenue trajectory in Q4 and beyond will also be closely watched for signs of renewed acceleration or further slowdown, given their revenue concentration and pricing dynamics. Broader market shifts in AI pricing, enterprise use cases, and competition among providers will shape how growth patterns unfold in the coming quarters.