Headlines about the AI economy lean heavily on one figure: NVIDIA's 'Data Center' revenue. It is worth knowing precisely what that number is in the company's own filings, because it is not, technically, a reportable segment. Segment reporting is the accounting discipline that governs how a company carves its business into pieces for financial disclosure, and the rules draw a sharp line between a reportable segment and a product category. Confusing the two leads readers to look for numbers in the wrong place in a filing.

Under U.S. generally accepted accounting principles, an operating segment is a component of a business whose discrete financial information is regularly reviewed by the entity's chief operating decision-maker to allocate resources and assess performance. A reportable segment is an operating segment that either crosses defined quantitative thresholds — such as 10 percent of combined revenue or profit — or is otherwise disclosed separately. The required disclosures live in a specific place: a 'Segment Information' note within the audited consolidated financial statements, where a company reports revenue, a measure of profit or loss, and certain other line items for each reportable segment. That note is the authoritative source; everything in the marketing materials traces back to it.

We report our business results in two segments. The Compute & Networking segment includes our Data Center accelerated computing platforms and AI solutions and software; networking; automotive platforms and autonomous and electric vehicle solutions; Jetson for robotics and other embedded platforms; and DGX Cloud computing services.— NVIDIA Corporation, Form 10-K (fiscal year ended January 26, 2025), source

NVIDIA's 10-K is explicit: the company reports two segments, Compute & Networking and Graphics. Data Center is named as a product platform inside the first of those, not as a segment of its own. In the segment note for the fiscal year ended January 26, 2025, the Compute & Networking segment reported revenue of $116,193 million and the Graphics segment $14,304 million, for consolidated revenue of $130,497 million. The much-quoted 'Data Center revenue' is a finer-grained, product-level breakout the company also provides — but it sits one level below the reportable-segment structure.

Why the distinction changes how you read the filing

The practical consequence is about where each figure is reported and what travels with it. Reportable-segment data carries an operating-profit measure: NVIDIA's segment note discloses operating income of $82,875 million for Compute & Networking and $5,085 million for Graphics, with a reconciling 'All Other' category that absorbed an operating loss of $6,507 million on the way to consolidated operating income of $81,453 million. Product-line revenue such as Data Center, by contrast, is generally disclosed as a revenue disaggregation without its own segment-level profit measure. So when a reader asks 'how profitable is NVIDIA's data-center business,' the filing answers at the segment level (Compute & Networking) rather than at the product level, because that is the level at which the company discloses profit.

The structure also explains why year-over-year comparisons can mislead if read at the wrong altitude. Compute & Networking segment revenue rose from $47,405 million for the year ended January 28, 2024 to $116,193 million for the year ended January 26, 2025 — the segment-level expression of the AI demand surge. That figure bundles networking, automotive, robotics, and DGX Cloud alongside the Data Center accelerators, so the segment line and the Data Center product line move together but are not identical. A precise claim cites whichever level the company actually reports the relevant figure at.

How to source a segment claim correctly

For any AI company, the method is the same. To find reportable segments and their profit, go to the Segment Information note in the audited financial statements — not the MD&A summary, and not an investor slide. To find a product-line figure such as Data Center, look for the revenue-disaggregation disclosure, which may sit in the same note or in the revenue note. Confirm how many segments the company reports and what each contains, because two companies in the same sector may draw their segment lines differently: a competitor might place its data-center accelerators in a segment defined around a different organizing principle, which makes cross-company 'segment' comparisons treacherous unless the definitions are matched first.

The segment structure is anchored in the audited financial statements for a reason worth understanding. Regulation S-X (17 CFR Part 210) governs the form and content of the financial statements filed with the SEC, and the segment note sits inside those statements rather than in the narrative chrome of the filing — which means it falls within the scope of the independent auditor's work. That is the practical distinction between a segment figure and a product-line statistic quoted on an earnings call: the former lives in a note the auditor has examined, the latter may be management commentary. When precision matters, the segment note carries more assurance than a slide, and a reader who wants the most defensible version of a revenue or profit figure goes to the note, not the deck.

None of this is a quirk of NVIDIA. It is the general architecture of segment reporting: the company defines its segments around how management actually runs and reviews the business, the auditor signs off on the segment note, and product-level breakouts ride below that structure. The headline 'Data Center revenue' is real and disclosed — but it is a product line inside a two-segment company, and a reader who knows that will find the number, and its profit context, exactly where the filing puts it.