OpenAI’s share of AI token spending among businesses on Ramp rose to 46.1% in the week ending 1 November, from 42.3% in the week ending 4 October, while the blended price those businesses paid OpenAI per million tokens fell 17.7% to $0.50, according to the Ramp AI Index page updated on 11 November. Anthropic kept the largest share, at 49.3%, down from 53.2%. By our arithmetic, Anthropic’s lead narrowed to 3.2 points from 10.9.
Across all providers, total token spending rose 10.5% against the same week a month earlier and total token volume rose 29.0%, the page shows. The token figures come from Ramp’s Token Spend Management product, which Ramp describes as a sample of its businesses that spend on AI tokens and use that product.
Business Pill · Tokens Explained
An evergreen Business Engineer explainer on tokens, the unit AI reads, writes and bills in. It teaches the general idea only and does not discuss Ramp’s data.
The key insight: Ramp’s latest token table puts OpenAI at 46.1% of business token spend in the week ending 1 November, up from 42.3% in the week ending 4 October, as its blended price per million tokens fell 17.7% to $0.50. Anthropic’s lead narrowed to 3.2 points from 10.9, by our arithmetic, and total token volume rose 29.0% while spend rose 10.5%, according to the page.
Update, 12 November 2026, 17:00 UTC: AI spend by department. Ramp has started classifying business AI spend by department, and 72% of it comes from engineering teams, according to the latest Ramp AI Index, Ramp economist Ara Kharazian wrote on X at 16:33 UTC. That spend “is starting to slow”, he wrote, with “customer support, and even management / leadership driving outsized spend”. A Ramp chart attached to the post says customer support teams accounted for 13% of classified AI token spend in October; on the two shares Ramp published, engineering spends about 5.5 times as much as customer support, by our arithmetic. On that chart (indexed, four-week moving average, department names normalised from API keys connected to Ramp), engineering spend climbed steeply from July 2026 and has since moved sideways below its late-summer peak. The department split is not yet on the AI Index page itself, and Ramp did not publish shares for the other departments. Kharazian framed the challenge for OpenAI and Anthropic as figuring out new revenue streams, citing lately Microsoft Office competitors and consumer agents.
OpenAI and Anthropic together took 95.4% of token spend and 96.5% of token volume in the week ending 1 November, by our arithmetic. In Ramp’s table, xAI held 2.4% of spend, Google 1.7% and DeepSeek 0.2%, up from 0.0%.
On volume, Anthropic’s share fell to 55.1% from 56.8% and OpenAI’s rose to 41.4% from 39.9%, according to the table. Ramp’s monthly figures, published on the same page, are in our piece on the top 1% of Ramp firms by AI spend.
Share of AI token spend among businesses on Ramp, Anthropic vs OpenAI, by week ending: 25 Sep and 23 Oct as the Ramp AI Index page showed them on 31 Oct 2026; 4 Oct and 1 Nov as read on 12 Nov 2026. The point gaps are our arithmetic.
The Price Gap Behind the Shift
The blended price per million tokens for OpenAI and Anthropic together fell 14.8% to $0.53 from $0.62, Ramp’s table shows. OpenAI’s fell 17.7% to $0.50 from $0.61, and Anthropic’s fell 12.5% to $0.55 from $0.63.
By our arithmetic, Anthropic’s blended price is now about 10% above OpenAI’s, against about 3% a month earlier. Token volume growth of 29.0% was about 2.8 times the 10.5% rise in spend, also by our arithmetic. On OpenRouter, a different dataset, we found Anthropic drawing 2.1 times OpenAI’s spend per token in a late-October week.
Four Weekly Readings
Ramp’s weekly shares move a lot. When we read the page on 31 October, it showed Anthropic at 60.5% of token spend and OpenAI at 35.1% in the week ending 23 October, and 51.3% and 44.2% in the week ending 25 September. By our arithmetic, Anthropic’s lead across the four readings ranged from 3.2 points to 25.4 points.
As we read the table, cheaper tokens did not shrink the bill: the blended price for the two labs fell 14.8% while volume rose 29.0% and spend still rose 10.5%, the page shows. The share moves are large for single weeks, though. Ramp’s readings put Anthropic’s lead at 25.4 points in the week ending 23 October, by our arithmetic, so one reading is a data point rather than an established trend.
Ramp AI Index methodology, on its token metrics
“The product imports usage and billing data from connected AI providers and records costs using provider and model identifiers”
Three Implications
FOR OPENAI Its share of business token spend on Ramp rose 3.8 points to 46.1% in a week in which its blended price per million tokens fell 17.7% to $0.50, according to the table.
FOR ANTHROPIC It kept the largest share of spend, 49.3%, and of volume, 55.1%, and its blended price of $0.55 is about 10% above OpenAI’s, by our arithmetic.
FOR FINANCE TEAMS The blended price for the two labs fell 14.8% while token volume rose 29.0% and spend rose 10.5%, according to the page; volume grew about 2.8 times as fast as spend, by our arithmetic.
The framework says, citing a Zenskar pricing guide, that with one action triggering dozens of model calls, “lower cost per token multiplied by exploding token volume increases total spend.”
As we read it, Ramp’s token table shows that arithmetic over one month of weekly data: the blended price fell 14.8%, volume rose 29.0% and spend rose 10.5%. The framework predates the table and does not discuss it.
Blended price per million tokens paid by businesses on Ramp, week ending 4 October vs week ending 1 November 2026, from the Ramp AI Index Tokenomics table.
What Is Not Established
Ramp says the token sample “is not representative of Ramp’s total AI spend volume, though results are directionally similar”. The page does not say how the blended price is weighted. Its shares and prices do not reconcile by simple division: Anthropic’s share of volume is above its share of spend although its blended price is higher, by our arithmetic, so we do not derive growth rates for each lab.
Each comparison covers single weeks, not months. Ramp has revised published weekly token figures before: the page showed different blended prices for the week ending 23 October when we read it on 27 and on 31 October. The figures describe businesses in Ramp’s sample, not either company’s total revenue. Ramp, OpenAI and Anthropic were not contacted for this piece.
This Week In The AI Supercycle
The Great Convergence of Capital, Power, and Geopolitics
This week’s issue: for two years the AI build-out was priced as a supply story, and last week as a credit story. This week it was priced as a permission story, and the answer arrived as a cost, not a ban.
In Ramp’s latest token data, OpenAI’s share of business token spend rose to 46.1% as its blended price per million tokens fell 17.7% to $0.50, and Anthropic’s lead narrowed to 3.2 points, by our arithmetic. Total token volume rose 29.0% while spend rose 10.5%.
A note on sourcing. All Ramp figures for the weeks ending 4 October and 1 November and the quotation come from the Ramp AI Index page, updated on 11 November 2026, which we read in full on 12 November. The figures for the weeks ending 25 September and 23 October come from the same page as we captured it on 31 October. The point gaps, sums and ratios are our arithmetic. Ramp’s token figures cover businesses in its Token Spend Management sample. Ramp, OpenAI and Anthropic were not contacted for this piece.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
