英伟达营收结构的AI占比分析
英伟达营收结构的AI占比分析
Conclusion
“One-third of Nvidia’s business unrelated to AI” should not be read as evidence that one-third of its reported revenue is non-AI. I could not verify that exact wording in Nvidia’s latest earnings release or the CNBC interview with Jensen Huang around the report. The latest disclosed revenue mix points the other way: Data Center revenue was out of total revenue of — of the quarter’s sales. Nvidia does not report revenue divided into “AI” and “non-AI” end uses, so its reported segments cannot validate a one-third claim on that basis.
Data as of: the U.S. close on . The latest company quarter covered the period ended .
What the mix does—and doesn’t—say
Nvidia reported Q2 FY2027 revenue up ; Data Center revenue was up year over year. The remaining was non-Data-Center revenue, not necessarily non-AI revenue. The company’s reporting categories describe businesses and products, not a clean split by how customers use them: Data Center is heavily tied to AI infrastructure, while products outside that segment can also support AI workloads.
A potentially confusing distinction appeared in Huang’s interview with CNBC on earnings day: he described roughly half of the business as hyperscalers and the other half as “everything else,” naming neoclouds, sovereign clouds and enterprises. That is a distinction by customer type, not by AI versus non-AI use; those customers may also be buying for AI deployments. (CNBC transcript)
Why the AI growth story remains compelling—and what investors may question
The latest outlook supports the growth case: Nvidia forecast Q3 FY2027 revenue of , with a range, and said it was not assuming any Data Center compute revenue from China. In the CNBC interview, Huang also discussed accelerating AI demand; the transcript reports management’s expectation of about 70% revenue growth in fiscal 2028. These are forward-looking management views, not guaranteed outcomes.
The counterweight is how concentrated the reported results remain in Data Center—and how much infrastructure and financing the expansion requires. Nvidia’s Q2 Form 10-Q disclosed supply commitments rising to from in the prior quarter. It also disclosed conditional land, power and data-center lease guarantees capped at , tied to a customer project. These are not equivalent to revenue or cash already spent; the filing describes conditions and potential exposure. But they help explain why investors scrutinize customer economics, financing arrangements and the durability of AI infrastructure demand. In the CNBC interview, the discussion explicitly turned to whether Nvidia was becoming a financier or backstop for the AI boom.
What the share-price response shows
Nvidia’s dividend-adjusted close rose from August 26 to August 27, the session after the results. By October 26, it had given back from the August 27 close; the latest quote was . That is a strong initial response to the results followed by a partial retracement—not a clean signal that investors were reacting to the “one-third” comment itself. The price move reflects the broader earnings, outlook and changing market expectations, and cannot isolate the effect of any single remark.
Bottom line: the comment, if accurately paraphrased, could point to a wider set of uses or a source of business outside AI. But Nvidia’s latest filings do not establish that one-third of revenue is non-AI; they show a business whose reported sales are overwhelmingly concentrated in Data Center, alongside a very strong growth outlook and substantial infrastructure-related commitments.