Alphabet said on 28 October 2026 that its first-party model APIs are now processing nearly 32 billion tokens per minute, more than four times what it processed a year ago. Its earnings slides put the quarter-on-quarter increase at 40%.
By our arithmetic, nearly 32 billion tokens a minute is nearly 46 trillion tokens a day, or about 1.4 quadrillion over a 30-day month.
A vertical news Short narrated from the published headlines and opening lines of this piece and two related FourWeekMBA pieces.
Business Pill · Tokens per Watt
An evergreen Business Engineer explainer on what an AI data centre produces, measured in tokens per watt. It explains the general idea only and does not discuss Google’s figures.
The key insight: By our arithmetic, Google’s stated API rate of nearly 32 billion tokens a minute is nearly 46 trillion tokens a day, through direct API use alone.
Google has given the API figure at each of its last five earnings releases. In October 2025 its CEO remarks said its first-party models “now process 7 billion tokens per minute, via direct API use by our customers.”
In February 2026 the figure was over 10 billion, and in April 2026 more than 16 billion. In July 2026 Google said its model APIs were processing approximately 22 billion tokens per minute. In October 2026 it was nearly 32 billion.
By our arithmetic from the stated figures, the rate rose about 4.6 times between the October 2025 and October 2026 releases. Google’s own wording is “more than four times.”
Tokens per minute processed by Google’s first-party models via direct API use, as stated at each earnings release: 7 billion (Q3 2025), over 10 billion (Q4 2025), more than 16 billion (Q1 2026), approximately 22 billion (Q2 2026), nearly 32 billion (Q3 2026). Sources: Alphabet CEO earnings remarks on blog.google; Alphabet Q3 2026 release, 28 October 2026.
What Was Posted on X
Q3 is in the books! Weโre now on a half-trillion annual revenue run rate, with Alphabet revenues growing at 25%.
– Our AI investments are driving overall Search query growth, including commercial queries – our YoY query growth rate is the highest itโs been in the last five years. – Google Cloud revenue grew 103%, powered by extraordinary demand for our AI infrastructure and AI solutions. Our 1P model APIs now process nearly 32B tokens/min (up 40% QoQ). – Strong momentum in model progress with Argon and Carbon leading the frontier, and Genie 4 our most advanced new world model yet.
Thanks to our partners and employees around the world. More to share on the call soon!
Sundar Pichai’s post says Google’s first-party model APIs now process nearly 32 billion tokens a minute, up 40% from the previous quarter.
A Year Ago, Across Every Surface
The API figure covers only direct API use by customers. A year ago, Google gave a wider measure: in its Q3 2025 remarks it said it was processing over 1.3 quadrillion monthly tokens across all its surfaces, up from 980 trillion announced in July 2025.
By our arithmetic, the direct API now runs at about 1.4 quadrillion tokens over a 30-day month, roughly the volume Google reported across all surfaces in October 2025. The two measures have different scopes, and Google did not give an all-surfaces figure in its Q3 2026 release.
The Gemini App and Search
Alphabet’s release quotes CEO Sundar Pichai: “the Gemini app has reached one billion usersโmaking it the fastest-growing product in Google’s history.” Its slides describe the figure as 1B+ Gemini app monthly active users.
Google’s earlier remarks gave more than 450 million monthly active users in July 2025, over 650 million in October 2025, over 750 million in February 2026 and 950 million in July 2026. In July 2026 Google also said AI Mode in Search had surpassed 1 billion monthly active users.
On Search, the release says Search & other revenue grew 12%, “with the year-over-year Search query growth rate the highest it has been in the last five years.” It also says Google surpassed 150 million YouTube Music and Premium subscribers globally, including trials.
The Enterprise Side
Google’s remarks have also tracked heavy enterprise users. In October 2025, it said nearly 150 Google Cloud customers had each processed approximately 1 trillion tokens over the past 12 months. In April 2026 the count was 330 customers, and in July 2026 nearly 500.
In its April 2026 remarks, Google said revenue from products built on its generative AI models grew nearly 800% year-over-year in the first quarter.
How OpenAI Reports
OpenAI reports a different measure. On 21 October 2026 it said: “More than 1.2 billion people use ChatGPT each week.” That is a weekly count of people, while Google’s Gemini figure is monthly active users, so the two are not directly comparable.
As we read the five releases, each quarter’s API figure has been higher than the last: 7 billion, over 10 billion, more than 16 billion, approximately 22 billion and nearly 32 billion tokens a minute.
As we read the scopes, the API number counts only customers calling Google’s models directly; the Gemini app and AI Mode figures count people, and Google gives them in monthly active users.
By our arithmetic, the direct API alone now runs at about 1.4 quadrillion tokens over a 30-day month, against the 1.3 quadrillion monthly tokens Google gave for all its surfaces in October 2025.
Sundar Pichai, CEO of Google and Alphabet, Q3 2026 release, 28 October 2026
“Overall, our first-party model APIs are now processing nearly 32 billion tokens per minute, more than four times what we processed a year ago”
Three Implications
FOR DEVELOPERSGoogle said in July 2026 that nearly 500 Cloud customers had each processed more than 1 trillion tokens in the last year; in October 2025 it said nearly 150 had each processed approximately 1 trillion.
FOR CONSUMERS Alphabet’s release says the Gemini app has reached one billion users, and its July 2026 remarks put AI Mode above 1 billion monthly active users.
FOR COMPARISONS OpenAI counts weekly people, more than 1.2 billion for ChatGPT on 21 October 2026, while Google counts monthly active users, so the two figures measure different things.
The framework describes the token this way: “It is simultaneously the unit of cognition the model produces, the unit of compute the data center serves, the unit of price the lab charges, and the unit of value the enterprise extracts.”
Alphabet’s release measures one of those roles directly: nearly 32 billion tokens a minute served through its APIs. The framework does not discuss these figures.
What Is Not Established
Google does not say how it counts tokens across models or whether input and output tokens are weighted the same. The tokens-per-day and per-month figures are our arithmetic from a per-minute rate and assume that rate held around the clock.
The quarterly figures are rounded and hedged in Google’s own words: over, more than, approximately and nearly. The growth multiple is our arithmetic from those rounded figures.
This Week In The AI Supercycle
Execution Fault Lines, Concentrated Demand, and Geopolitical Firewalls
This week’s issue: the AI build is not short of demand for its parts. It is short of cheap money, and its end revenue rests on a narrow base of large, often venture-funded buyers.
Google’s stated API token rate went from 7 billion a minute in October 2025 to nearly 32 billion in October 2026. By our arithmetic, that is nearly 46 trillion tokens a day through the direct API alone.
Business Engineer Premium
Premium: Alphabet & Microsoft on the Map of AI
Cash conversion, the lengthening backlog, the arms-dealer turn at Layer 3: the full Business Engineer read of both prints.
95,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. Figures come from Alphabet’s third-quarter 2026 results release and slides of 28 October 2026, Sundar Pichai’s post on X the same evening, Alphabet’s CEO earnings remarks published on blog.google for the previous four quarters and July 2025, and an OpenAI page of 21 October 2026. Tokens per day and per month, and the growth multiple, are our arithmetic; linked earlier pieces carry their own sources.
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.
