OpenRouter, the AI modelmarketplace that Stripe agreed to acquire in August, crossed a 10 quadrillion token a year run rate, according to a chart labelled with OpenRouter’s rankings data that Menlo Ventures principal Deedy Das posted on X on 2 November. The chart says the week of 26 October closed at 194 trillion tokens. Menlo Ventures, where Das works, invested in OpenRouter’s seed round in March 2025, the firm has said.
That run rate is about 6.7 times the ~1.5 quadrillion tokens a year that Menlo Ventures reported on 26 May 2026, by our arithmetic, and 194 trillion tokens a week works out to about 27.7 trillion tokens a day.
The key insight: OpenRouter crossed a 10 quadrillion token a year run rate in the week of 26 October, when it processed 194 trillion tokens, according to a chart of OpenRouter data posted by Menlo Ventures principal Deedy Das; that is about 6.7 times the ~1.5 quadrillion run rate Menlo reported on 26 May, by our arithmetic.
The chart says OpenRouter’s volume has doubled every 10ยฝ weeks, about 33% growth month over month, sustained for over three years, and that the last nine weeks processed more tokens than the previous 3ยผ years combined. Das wrote that “Token volumes continue to 2x every 10.5 weeks.”
The milestones labelled on the chart are 10 billion tokens a week in January 2024, 100 billion in September 2024, 1 trillion in March 2025 and 10 trillion in February 2026. From 10 billion a week to 194 trillion a week is about 19,400 times, by our arithmetic.
Menlo Ventures’ own posts give the run rate along the way. On 26 May 2026 it said OpenRouter would be at a ~1.5 quadrillion tokens a year run rate that week, up from ~100 trillion tokens a year about a year earlier. On 19 August 2026, the day OpenRouter announced the Stripe agreement, Menlo wrote that token usage had grown ~30,000x since launch to a 4.5+ quadrillion-token annual run rate, doubling every 11 weeks. From 4.5 to 10 quadrillion is about 2.2 times in roughly ten weeks, by our arithmetic.
Business Pill · The Jevons Paradox Explained
An evergreen Business Engineer explainer on the Jevons paradox. It teaches the general idea only and does not discuss OpenRouter’s figures.
What Was Posted on X
Congratulations to OpenRouter for crossing 10 QUADRILLION annual token run rate today!
Token volumes continue to 2x every 10.5 weeks.
In the last 9 weeks, OpenRouter processed more tokens than all of time before that.
Deedy Das, a principal at Menlo Ventures, which invested in OpenRouter, posted that OpenRouter crossed a 10 quadrillion annual token run rate and that token volumes double every 10.5 weeks. His chart, which cites OpenRouter’s rankings data, is the source of the 194 trillion weekly figure in this piece.
OpenAI CEO Sam Altman on AI:
The energy per query is falling incredibly dramatically. We’re just delivering huge efficiency gains.
On the other hand, as we make it more efficient, people just want to use it much more.
And I think we, the AI industry and the energy industry, have an obligation to society to deliver that.
A clip of Sam Altman at ADIPEC, posted by Clash Report, in which he says the energy per query is falling and that people use AI much more as it gets more efficient. The piece quotes the same remarks from YouTube’s auto-captions of the session.
OpenRouter tokens processed per week, log scale, from the milestones labelled on a chart of OpenRouter rankings data posted by Menlo Ventures’ Deedy Das on 2 November 2026: 10 billion in January 2024, 100 billion in September 2024, 1 trillion in March 2025, 10 trillion in February 2026 and 194 trillion in the week of 26 October 2026. Values are read from the chart’s labels.
Cheaper per Query, More Queries
On the same day, OpenAI chief executive Sam Altman described the same pattern from OpenAI’s side at ADIPEC in Abu Dhabi. “The energy per query is falling incredibly dramatically,” he said, according to YouTube’s auto-captions of the session. “As we make it more efficient, people just want to use it much more.”
In the same answer Altman said the global average per capita is about 100,000 tokens a month and the token leader uses about 500 billion tokens a month, as we reported from the session. As we read the two together, OpenRouter’s weekly totals are one public measure of that demand side, on a single routing platform.
As we read the chart and Menlo’s posts together, the growth has not slowed as the base got larger: Menlo put the run rate at 4.5+ quadrillion on 19 August, and the 10 quadrillion mark came about ten weeks later, a 2.2 times step by our arithmetic.
As we read Sam Altman’s ADIPEC remarks alongside the chart, he described lower energy per query and higher usage at the same time; OpenRouter’s weekly totals show the usage half of that on one routing platform.
Sam Altman, OpenAI chief executive, at ADIPEC, 2 November 2026 (YouTube auto-captions)
“as we make it more efficient, people just want to use it much more”
Three Implications
FOR MODEL PROVIDERS OpenRouter processed 194 trillion tokens in the week of 26 October, according to the chart, and Menlo Ventures wrote on 19 August that token usage had grown ~30,000x since launch.
FOR STRIPE Menlo Ventures wrote on 19 August that OpenRouter had entered into an agreement to be acquired by Stripe; the run rate it gave then, 4.5+ quadrillion tokens a year, is less than half the 10 quadrillion on Monday’s chart.
FOR AI BUYERS Altman said the token leader uses about 500 billion tokens a month and the global average per capita is about 100,000, according to YouTube’s auto-captions of the session.
The analysis opens: “AI has crossed from the training era to the inference era.”
As we read it, a platform that routes inference requests to many models processing 194 trillion tokens in a week is one measure of that shift. The framework is the owner’s argument and does not discuss OpenRouter’s October figures.
The 194 trillion figure and the milestones come from a chart that cites OpenRouter’s rankings data and was posted by an investor in OpenRouter; we did not find a separate OpenRouter statement of the figure, and the weekly totals are not independently audited. The milestone values in the chart above are our reading of the labels on that chart.
Token counts are not revenue, and OpenRouter is one platform among many ways developers reach AI models, so its volume does not measure total industry usage. Altman’s efficiency remarks were about OpenAI and did not mention OpenRouter.
This Week In The AI Supercycle
Anthropic’s Reality Check, The Capital Squeeze, & The New World Order
This week’s issue: for two years the AI build-out was priced as a growthstory. This week it started being priced as a credit story.
OpenRouter crossed a 10 quadrillion token a year run rate in the week of 26 October, when it processed 194 trillion tokens, according to a chart of OpenRouter data posted by Menlo Ventures’ Deedy Das. That is about 6.7 times the run rate Menlo reported in May, by our arithmetic.
A note on sourcing. The run rate, the 194 trillion weekly total and the milestones come from a chart citing OpenRouter’s rankings data that Menlo Ventures principal Deedy Das posted on X; Menlo Ventures is an OpenRouter investor. The May and August run rates come from Menlo Ventures’ posts, and Sam Altman’s words from YouTube’s auto-captions of the ADIPEC session. The 6.7x, 2.2x, 19,400x and per-day figures are our arithmetic.
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.
