Dear Colleagues: Today’s letter includes my monthly update of Ramp AI Index, our flagship research using spend data from Ramp to track how American businesses are using AI. Last month, I wrote a post called Cracks in the AI Thesis, in which I presented my view that rising adoption of cheaper models and mediocre adoption of Fable 5, posed a threat to the model companies, which rely on rising business spend on frontier models. This post is an update with my latest thinking.
Extras from Ramp data:
I’m hiring! Looking for an economist / economics-minded data scientist who has demonstrated experience producing economic research for across audiences.
Moody’s: I joined Mark Zandi, chief economist at Moody’s Analytics on his podcast Inside Economics to talk about the latest trends in AI adoption. Apple Podcasts, Spotify, YouTube).
Model market share: Anthropic widens its lead, but overall business AI adoption slows
In August, Anthropic extended its lead in business AI adoption. 43.8% of U.S. businesses paid for subscriptions or tokens from Anthropic, up 0.34 percentage points month over month. OpenAI underperformed overall AI adoption, rising only 0.09 percentage points to 39.8% of businesses. In general, new AI adoption continues to grow but is decelerating. Technical sectors (information, finance, and professional services) lead adoption rates.
Per employee AI spend declines
This chart shows median per employee per month AI spend for the top 1%, top 10% and median firm. This dataset is closely watched for a benchmark of where the typical firm’s AI spend is going to land. Will the median firm today start to look more like a top 1% firm tomorrow?
AI bulls expect spending to increase across segments, but growth is especially important among the top 1%, which drives the vast majority of enterprise revenues for the model companies. Last month, that number fell 9.7% from $7,976 to $7,205.
Methodological note: Because the top 1% represents a small group of firms, its spending estimate is more volatile than our median and top 10% estimates. We also revised up July spending for the top 1% from approximately $7.4K to $8K per employee per month after additional July transactions entered our dataset. These results are also subject to change.
There’s an innocuous explanation here. It’s summer, and a lot of engineers take off in August. We’ve similarly observed declines in AI spend around November and December. I wouldn’t underrate this, but there are a few other factors that are driving AI spend down, and will affect all firms, even those outside of the top 1%.
First, the price of AI is coming down. OpenAI and Anthropic have both announced a series of price cuts over the last month. Our latest index tracking the effective price per million tokens shows that prices have declined 41% to $0.68 as of this week, down from the 2026 peak of $1.15 in March.
That would be fine, if AI companies are able to drive growth in the form of volume. So far, they have been. Unfortunately, that volume may not be large enough to counteract the effective decline in prices.
The models driving volume increases are relatively cheap, standard models like GPT-5.6 Terra and Claude’s Sonnet series. Frontier models like Opus, Fable, and Sol drove 45% of token share, up from the start of July, but down from a 53% peak in August. We’ve heard from businesses who are imposing company-wide defaults that reduce usage of frontier models, saying standard models are still highly performant and also more cost effective.
Finally, it seems these trends are not driven by adoption of open source models or Chinese models. While rising, open source adoption remains limited to a small segment of businesses on our platform. Only 6.4% of AI spending businesses use them and 3.6% of businesses overall (note, we measure using adoption of routing platforms, which also provide access to closed-models, so actual open source adoption is likely lower).
Something from my camera roll
The McDonald’s in Altadena has fast wifi and is a great place to work. January 2026 around 7pm.
The thing that strikes me is the sheer difference between AI spend in big companies (avg $8, 000) vs. Median spend $14. That is a crazy difference
It would be interesting to compare these adoption figures with measurable workflow outcomes. Paying for an AI product shows interest, but it doesn’t necessarily mean the business has changed how work is performed. The next important measurement may be how many organizations have moved from individual tool usage to redesigned, repeatable processes.