Nvidia's latest quarterly report is mostly a data center report. The company said on Nov. 17, 2026 that revenue for the third quarter of fiscal 2027, which ended Oct. 25, was $115.8 billion. Its Data Center segment accounted for $107.6 billion of that, according to the company's earnings release.

For utilities, developers, colocation operators and the communities where new campuses are being proposed, the figures matter less as a scorecard and more as a measure of how much accelerated computing hardware is being bought, and therefore how much of it will need buildings, power and cooling. The quarter also marked the start of shipments of Nvidia's Vera Rubin NVL72 rack-scale systems, which the release says are running at several cloud partners.

The company's leadership situation, including the board's decision on former CEO Jensen Huang, has been covered separately. This article covers what the results say about infrastructure demand.

The Numbers

Nvidia reported that quarterly revenue rose 20% from the prior quarter and 103% from the same quarter a year earlier. Data Center revenue grew 21% sequentially and 110% year over year, according to the release. The company also reported $8.2 billion in Edge Computing revenue.

The release lists GAAP gross margin at 74.8% and GAAP diluted earnings per share at $2.70. 10News/Scripps reported net income of $65.3 billion and said the company's guidance came in above analyst forecasts. The BBC, in a report carried by Yahoo Finance, noted that this was the first time Nvidia's quarterly sales topped $100 billion, and reported that the shares rose 4.8% in after-hours trading.

The BBC also quoted analyst Daniela Hathorn, who said the results answer concerns that demand for AI infrastructure is cooling.

Vera Rubin Reaches Customer Data Centers

The release says Nvidia ramped production and began shipping Vera Rubin NVL72, with racks running at partners including CoreWeave, Google Cloud, Microsoft Azure, Oracle Cloud Infrastructure and Nebius, and coming online soon at SpaceXSI. The NVL72 is a rack-scale product, so each installation is a full rack system that a host facility has to be ready to power and cool.

That list mixes large general-purpose cloud platforms with GPU-focused cloud providers. For site selectors and grid planners, the moment Vera Rubin moves from announcement to installation is when the generation's facility requirements stop being projections and start showing up in actual data halls.

Nvidia also pointed to its DSX POD in the release, describing it as a reference design for complete Vera Rubin AI factories built by partners inside modular containers, integrated and validated before delivery. The release did not attach a specific megawatt figure to it.

Outside the United States, the company described a collaboration with Australia's data center ecosystem on infrastructure supporting a planned buildout of up to 2 gigawatts by 2027 for DSX AI factories, according to the release. That 2 GW figure is a plan stated by Nvidia, not a measure of capacity already built or energized.

A $130 Billion Outlook, Without China Data Center Compute

For the fourth quarter of fiscal 2027, Nvidia projected revenue of $130.0 billion, plus or minus 3%. The release notes two qualifications. The quarter includes an additional week, and the outlook does not assume any Data Center compute revenue from China.

On the earnings call, interim CEO Colette Kress said Nvidia is no longer shipping Hopper data center chips to China, the BBC reported. The BBC noted that the H200, a Hopper-generation chip, had been approved for sale there, and that Kress did not explain why shipments had stopped.

The export-control backdrop moved this week too. 10News/Scripps reported that the Commerce Department, a day before the earnings report, moved to close loopholes that let Chinese companies obtain advanced chips by routing them through third countries or renting computing power from data centers outside China. The same report said Nvidia's board announced a new National Security and Export Control Oversight Committee.

As a result, Nvidia's guidance for the coming quarter includes no Data Center compute revenue from China, as the release states.

Capital Returns

Nvidia said it had approximately $227.3 billion remaining under its share repurchase authorization and will pay a quarterly dividend of $0.25 per share, according to the release.

On the call, Kress described the company's operations as "business as usual," 10News/Scripps reported. Board chair Harvey Jones said "the board does not intend to reinstate" Huang, according to the same report. The BBC reported that Nvidia is working with an executive search firm to find a permanent CEO.

What It Means for the Buildout

A single supplier's quarter can't show where individual campuses will go or how quickly interconnection queues will clear. The report does establish three on-the-record points. Nvidia's Data Center revenue rose 110% year over year, to $107.6 billion. Vera Rubin NVL72 racks are running at named cloud partners including CoreWeave, Google Cloud, Microsoft Azure, Oracle Cloud Infrastructure and Nebius. And the company's outlook for the next quarter assumes no Data Center compute revenue from China.

For local officials and residents assessing proposed facilities, the Vera Rubin shipments are one link between these results and physical construction, though the release does not say where individual racks are installed.

What Happens Next

Nvidia's $0.25 quarterly dividend is payable Dec. 23, according to the company's release. Outside the company's financial calendar, 10News/Scripps reported that Huang is scheduled to meet with three House committees on Dec. 1, and that Nvidia's general counsel is scheduled to testify on Dec. 2.

Sources