STATS | AI Infrastructure. NVIDIA reported $107.6 billion in Data Center revenue for its fiscal third quarter ended October 25, 2026, up 110% from the corresponding quarter a year earlier. Total company revenue reached $115.844 billion, up 103% year over year, according to its November 17 earnings release. The figures reflect realized sales, not forecast infrastructure spending.
Data Center now represents approximately 92.9% of NVIDIA’s quarterly revenue. The segment spans accelerated computing and networking sold into AI training, inference, cloud and enterprise infrastructure. Its growth makes AI infrastructure the dominant driver of the company’s financial results, although revenue alone does not disclose how many GPUs were installed or how much electrical capacity customers brought online.
The financial comparison
NVIDIA reported total revenue of $57.006 billion in Q3 FY2026, $96.221 billion in Q2 FY2027 and $115.844 billion in Q3 FY2027. On the exact disclosed totals, year-over-year growth equals ($115.844 billion − $57.006 billion) ÷ $57.006 billion × 100 = 103.21%. The dollar increase was $58.838 billion. The quarter-over-quarter comparison is ($115.844 billion − $96.221 billion) ÷ $96.221 billion × 100 = 20.39%.
Data Center revenue increased 21% sequentially and 110% annually, based on NVIDIA’s reported growth rates. With the segment reported to one decimal place in billions, deriving an exact prior-year Data Center dollar figure by reversing the rounded percentage would introduce false precision. The approximately 92.9% revenue share likewise uses rounded $107.6 billion segment revenue divided by the more precise $115.844 billion consolidated figure.
What the chart should show
The appropriate comparison is a zero-baseline vertical column chart with Q3 FY2026 ($57.006B), Q2 FY2027 ($96.221B) and Q3 FY2027 ($115.844B) on the horizontal axis and consolidated revenue in US$ billions on the vertical axis. The observations are discrete fiscal quarters; Q3 FY2026 and Q2 FY2027 are not adjacent periods. All data originate from NVIDIA’s November 17, 2026 primary earnings release.
Vera Rubin shipments and the limits of revenue data
NVIDIA said it had ramped production and begun shipping Vera Rubin NVL72 systems, with racks operating at several named cloud partners. That statement describes a company-reported commercialization milestone. It does not quantify installed rack counts, total GPUs, uptime, achieved inference throughput or megawatts. Large-scale AI infrastructure also depends on optical networking, cooling, power distribution and high-bandwidth memory, each with separate supply constraints.
The acceleration intersects with other networking trends covered by BitcoinVersus.Tech, including Cisco’s AI networking outlook, Arista’s revenue growth, 1.6 Tb/s optical modules, and 800 VDC data center power architecture. Those reports cover different products and measurement periods; their revenue figures must not be added to NVIDIA’s results.
Forecasts are not recorded sales
NVIDIA projects approximately $130 billion in fourth-quarter FY2027 revenue, plus or minus 3%, implying a guidance interval of $126.1 billion to $133.9 billion. Management excludes Data Center compute revenue from China in that outlook. The next fiscal quarter also contains an extra week, making an unadjusted sequential growth comparison potentially misleading. The forecast is not yet earned revenue and may differ from actual results.
The reported GAAP gross margin of 74.8% and net income of approximately $65.3 billion are accounting measures, not indicators of AI system efficiency. Future margins, chip availability and infrastructure demand remain uncertain. Revenue recognition, rather than hardware commissioning, determines when sales appear in the financial statements.
Source: NVIDIA, third-quarter FY2027 earnings release, November 17, 2026. Featured illustration is conceptual, not a photograph of a named NVIDIA customer facility.

Leave a Reply