Among US businesses paying for AI on Ramp, the top 1% spent a median $10,199.20 per employee per month on AI in October, up from $3,230.04 in January, according to the monthly series behind the charts on the Ramp AI Index. That is 3.16 times the January level, by our arithmetic. The median firm paying for AI spent $16.32, up from $6.19, and the top 10% spent $800.84, up from $285.68.
Our number: the top 1% now spends about 625 times the median firm per employee, up from about 522 times in January, by our arithmetic. Over twelve months the picture is different: from October 2025 the top 1% grew 3.19 times and the median firm 3.07 times. Ramp puts AI adoption at 56.8% of US businesses on its platform in October, up from 45.4% in January.
Business Pill · Percentiles Explained
An evergreen Business Engineer explainer on percentiles and medians. It teaches the general idea only and does not discuss Ramp’s data.
The key insight: As we read Ramp’s series, the top 1% of AI-paying firms tripled spend per employee in 2026 while the median firm grew 2.64 times, so the gap widened to about 625 times, by our arithmetic; over twelve months, though, both groups grew about threefold.
What Ramp’s Series Shows, January to October
Ramp says the index uses transaction data from more than 85,000 firms on its platform. “Spend per employee covers only firms paying for AI; for the top 10% and top 1% of those firms, it is the median within each group”, the page says. Ramp’s summary line reads: “Among firms paying for AI, the top 1% spent a median of $10,199.20 per employee per month on AI, compared with $16.32 for the median firm”.
From January to October, by our arithmetic, the top 1% rose 216%, the top 10% rose 180% and the median firm rose 164%. In dollars, the top 1% added $6,969 per employee per month. In October alone it added $852.70, while the top 10% added $7.59 and the median firm $0.93, according to Ramp’s table.
The figures were relayed on X in a post by @KobeissiLetter, a markets commentary account, which wrote that the top 1% “allocated a record $10,199 per employee to AI in October”. Its percentages match Ramp’s series. Ramp’s figures are per employee per month, and Ramp’s January adoption reading is 45.44%, which the post gave as 45.5%.

What Was Posted on X
The Kobeissi Letter, a markets commentary account, posted this at 19:21 UTC on 16 November, relaying Ramp’s October figures for the top 1%, top 10% and median firm and Ramp’s adoption rate. The piece takes every figure from Ramp’s own AI Index page and its monthly series, not from the post; Ramp’s figures are per employee per month.
The Gap Widened Since January
The top-1%-to-median multiple was about 184 times in July 2023, the first month in Ramp’s series, about 601 times in October 2025, about 522 times in January and about 687 times in July 2026, before easing to about 625 times in October, by our arithmetic. Over twelve months the multiple rose from about 601 times to about 625 times. The top 1% now spends about 12.7 times the top 10%.
Year on year, all three groups grew at similar rates: 3.19 times for the top 1%, 3.61 times for the top 10% and 3.07 times for the median firm, by our arithmetic. October’s monthly figures times twelve come to $122,390 per employee for the top 1%, $9,610 for the top 10% and $196 for the median firm.
Ramp economist Ara Kharazian wrote in the November letter that “AI spend is highly concentrated. 1% of businesses drive 80% of the revenues to OpenAI and Anthropic”, and that growth in the median and top 10% groups “is relatively sluggish”. Our 11 November piece on the letter covered October’s level.
Anthropic, OpenAI and Firms That Pay Both
In October, 44.8% of Ramp businesses paid Anthropic and 40.6% paid OpenAI, according to the index. Because only 56.8% paid for any AI, at least 28.6% of all Ramp businesses must have paid both labs, about half of AI-paying firms, by our arithmetic. That is a floor from the two shares, not a count of shared customers.
On tokens, Anthropic took 49.3% of token spending in the week ending 1 November, down from 53.2% in the week ending 4 October, and OpenAI took 46.1%, up from 42.3%, according to Ramp’s Tokenomics table. The gap narrowed from 10.9 points to 3.2 points, by our arithmetic. Ramp says the blended price per million tokens for the two labs fell 14.8% over the same comparison, to $0.53.
Ramp says the token figures cover businesses that use its AI Spend Management product, that “our historical estimates are subject to revision” and that the data “is not representative of Ramp’s total AI spend volume, though results are directionally similar”.
Ramp Next to Databricks: Two Different Gaps
Databricks published its own index on 16 November. In our Databricks AI Index piece, a top-1% user of models through its Unity Gateway spent about 281 times the median user, down from about 1,279 times in July, by our arithmetic. On Ramp, the top-1%-to-median multiple also narrowed after July, from about 687 times to about 625 times.
The units differ: Databricks measures weekly spend per user of its gateway, Ramp monthly spend per employee across firms. In our piece on multi-provider use, 71.1% of companies active on the Databricks gateway had used two or more model providers. Ramp’s floor of 28.6% of all businesses paying both Anthropic and OpenAI points the same way, by our comparison, from a different population.
The Structural Read
As we read the series, the January-to-October widening comes from the top: the top 1% added $6,969 per employee per month, while the median firm added $10.13, by our arithmetic.
As we read the adoption table, Anthropic and OpenAI share most of the same customers: with 56.8% of businesses paying for any AI, at least 28.6% must pay both labs, by our arithmetic.
Ara Kharazian, lead economist, Ramp, 11 November 2026
“AI spend is highly concentrated. 1% of businesses drive 80% of the revenues to OpenAI and Anthropic”
Three Implications
FOR AI VENDORS Ramp’s top 1% of AI-paying firms spend a median $10,199.20 per employee per month, about 625 times the median firm, by our arithmetic.
FOR THE TWO LABS Anthropic’s token-spend share fell to 49.3% and OpenAI’s rose to 46.1% in the week ending 1 November, according to Ramp.
FOR THE MEDIAN FIRM The median AI-paying firm spends $16.32 per employee per month, up 164% since January, by our arithmetic.
The Business Engineer Lens
This story maps onto the Business Engineer framework The Product Overhang Doctrine.
The framework says: “In any product cycle, there are two curves to follow. The first is what your underlying technology can do. The second is what your users can actually do with it.”
As we read Ramp’s series, the top 1% at $10,199.20 and the median firm at $16.32 per employee per month show how far apart the second curve sits across firms; the framework predates the data and does not discuss it.
What Is Not Established
Ramp’s data covers businesses on Ramp, not every US firm, and only spending seen in corporate card, invoice and ACH transactions processed by Ramp. The page says “It’s likely our results underestimate actual AI adoption and spend due to the prevalence of businesses using free AI tools”. The tier figures are medians within groups, not totals, so the multiples compare typical firms in each group, not total spending.
The twelve-month figures are October’s level times twelve, not a forecast. The overlap of Anthropic and OpenAI customers is a lower bound from Ramp’s adoption shares, not a figure Ramp reports. The comparison with Databricks is ours; the two companies use different samples and measures.
The Bottom Line
Ramp’s series shows the top 1% of AI-paying firms at a median $10,199.20 per employee per month in October, 3.16 times January, while the median firm reached $16.32. The gap rose from about 522 times to about 625 times since January, by our arithmetic, though over twelve months the top 1% and the median firm grew at similar rates.
95,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. The figures come from the Ramp AI Index page, including the monthly series behind its charts, and Ramp Economics Lab’s 11 November letter, both on ramp.com. The @KobeissiLetter post is a relay of Ramp’s data and is attributed as such. Growth multiples, ratios, the twelve-month figures, the overlap floor and the Databricks comparison are our arithmetic. We did not contact Ramp for this piece.
Sources: Ramp Economics Lab: Ramp AI Index (data through Oct 2026) · Ramp Economics Lab: Ara Kharazian, New Fronts in the AI Trade (11 Nov 2026) · X: @KobeissiLetter post (16 Nov 2026, 19:21 UTC)









