Microsoft finance chief Amy Hood said on the company’s 28 October 2026 earnings call that it “had higher token pricing than we expected in the quarter”, according to a machine transcript of Microsoft’s webcast. Chief executive Satya Nadella said Microsoft now has “over 120,000 Foundry customers with revenue more than tripling year over year.”
Microsoft’s 10-Q for the quarter puts Microsoft Cloud gross margin at 66% and explains the decline with “sales mix shift to Azure”. In the four earlier filings we reviewed, Microsoft’s explanation cited AI infrastructure; this one does not, by our reading.
Business Pill · Usage-Based Pricing
An evergreen Business Engineer explainer on usage-based pricing. It explains the general idea only and does not discuss Microsoft’s results or token pricing.
The key insight: Microsoft’s finance chief said token pricing came in higher than expected, its chief executive said Microsoft’s own chips are raising tokens per watt, and its 10-Q now explains the 66% Microsoft Cloud gross margin with the shift to Azure rather than AI infrastructure, by our reading.
In her prepared remarks, Hood said Foundry revenue more than tripled “with upside driven by efficiency gains across our GPU fleet that allowed us to allocate additional capacity to customers, as well as higher revenue per token that we’d benefited from in the quarter.”
Answering an analyst, she said “token pricing will create a little bit of variability”, and later that “there’s also some unknown on some token pricing through the quarter that stuff can change.”
Nadella said “Azure OpenAI directly benefited in the last quarter with big advances in OpenAI models”, and called Copilot “the new OS for work.”
Microsoft Cloud gross margin percentage by quarter, from Microsoft’s 10-Qs; the June 2026 quarter is not stated separately. The December 2026 figure is Amy Hood’s guidance on the 28 October call (machine transcript of the webcast).
What Microsoft Said About the Cost of a Token
Nadella said Microsoft’s own Maia 200 chips are already delivering “up to 1.7 times more tokens per watt across the AI experiences in Teams and Copilot”. The machine transcript renders the chip name as “Maya”.
He said the next generation, Maia 300, “will drive further gains delivering 40% more tokens per watt compared to the leading GPU-based systems.” That is Microsoft’s claim; it gave no benchmark details on the call.
Nadella also said “Accelerating demand led to us adding 24 new data centers across five continents this quarter.” Hood said efficiency work let Microsoft “put more allocation into Foundry in the quarter.”
How the Margin Explanation Changed
Microsoft’s 10-Q for the quarter to 30 September 2025 said Microsoft Cloud gross margin “decreased to 68% driven by scaling our AI infrastructure and growing usage of AI product features”.
The 10-Qs for the December 2025 and March 2026 quarters gave 67% and 66%, each “driven by continued investments in AI infrastructure and growing AI product usage”. The fiscal 2026 10-K gave 66% for the year with the same wording.
The 10-Q filed on 28 October 2026 gives 66% for the September 2026 quarter, “driven by sales mix shift to Azure, offset in part by product-level margin improvements, primarily in Azure.” By our arithmetic, that is two points below the year-earlier quarter.
On the call, Hood said Microsoft Cloud gross margin percentage “should be roughly 65%, down year over year, as continued improvement in product-level margins is more than offset by sales mix shift to Azure.”
Google’s Price Card the Same Day
Google’s Gemini 4 Carbon announcement on 28 October said the model will launch at the same introductory price as Argon, “at $2 per million input tokens and $10 per million output tokens”, and that after the introductory period “the price of $4 per 1M input tokens and $20 per 1M output tokens will apply.”
Microsoft did not give a per-token price or a revenue-per-token figure on the call, so the two companies’ statements cannot be set side by side on one number.
As we read the call, Microsoft described both sides of the token: Hood on the revenue it earns per token and Nadella on the energy each token costs on its own Maia chips.
As we read the filings, the four earlier explanations of the Microsoft Cloud margin cited AI infrastructure, while the September 2026 10-Q cites the sales mix shift to Azure and product-level margin improvements.
Google listed Gemini 4 Carbon at $2 per million input tokens and $10 per million output tokens on the same day; Microsoft gave no per-token figure, so the two cannot be compared on one number.
Microsoft CFO Amy Hood, on the 28 October 2026 call (machine transcript of the webcast)
“We had higher token pricing than we expected in the quarter”
Three Implications
FOR FOUNDRY Nadella said Microsoft has over 120,000 Foundry customers, with revenue more than tripling year over year.
FOR MARGINS Hood guided Microsoft Cloud gross margin percentage to roughly 65% for the December quarter, citing the sales mix shift to Azure.
FOR WHAT IS NOT DISCLOSED Microsoft gave no token price or revenue-per-token figure, and the 10-Q does not state the June 2026 quarter’s margin separately.
The framework argues: “The inference economy is no longer a prediction.”
On the call, Microsoft talked about inference in token terms: revenue per token and tokens per watt. The framework does not discuss Microsoft’s quarter.
What Is Not Established
The call quotations come from a machine transcript of Microsoft’s webcast, not Microsoft’s official transcript, which was not available when this was written. Microsoft gave no figure for token prices or revenue per token, and the 10-Q does not state the June 2026 quarter’s Microsoft Cloud gross margin separately.
Microsoft Cloud gross margin and Google Cloud operating margin measure different things, so we do not compare them.
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.
Hood said token pricing came in higher than Microsoft expected and helped Foundry, whose revenue more than tripled, while Nadella said Microsoft’s own chips are cutting energy per token. The 10-Q now explains Microsoft Cloud’s 66% gross margin with the shift to Azure rather than AI infrastructure, and Hood guided to roughly 65% for the December quarter.
95,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. Quotations from Microsoft’s 28 October 2026 earnings call come from a machine transcript of Microsoft’s own webcast, not the official transcript. Microsoft Cloud gross margin figures and wording come from Microsoft’s 10-Q and 10-K filings; Google’s prices come from its Gemini 4 Carbon announcement. Point differences 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.
