A new quarterly scale for Nvidia
Nvidia’s latest earnings report pushes the company into territory that even the AI boom had not yet delivered: quarterly revenue above $100 billion. For the fiscal third quarter ended October 25, 2026, Nvidia reported revenue of $115.8 billion, up 20% from the prior quarter and 103% from the same period a year earlier . The company’s own revenue-by-market trend data shows that the previous quarter, fiscal Q2 2027, reached $96.221 billion, making the new $115.844 billion total the first reported quarter above the $100 billion line .
The scale of the result matters as much as the growth rate. A company that was already treated as the defining infrastructure supplier of the AI era has now produced a single-quarter sales figure that resembles the annual revenue of large global technology groups. The headline number was not just a symbolic milestone: it came with GAAP net income of $65.3 billion, GAAP gross margin of 74.8%, and diluted GAAP earnings per share of $2.70 . On a non-GAAP basis, Nvidia reported gross margin of 74.9% and diluted earnings per share of $2.68 .
The quarter also exceeded expectations. Reuters reported that Nvidia’s fourth-quarter revenue forecast of $130.0 billion, plus or minus 3%, compared with an LSEG-compiled analyst average estimate of $125.25 billion . Investing.com’s transcript coverage said the reported $115.8 billion in revenue was above an expected $108.85 billion, with adjusted EPS of $2.68 versus a $2.47 forecast . In other words, Nvidia beat on the completed quarter and set a higher-than-expected bar for the current one.
Data center remains the whole story
The earnings release leaves little doubt about what is driving Nvidia’s financial profile. Data Center revenue was $107.6 billion in the third quarter, up 21% sequentially and 110% year over year . That means the Data Center platform accounted for almost 93% of Nvidia’s total quarterly revenue, reinforcing how concentrated the company’s current growth is in AI infrastructure, accelerated computing, networking, and large-scale cloud deployments.
Nvidia’s supplemental trend table breaks the Data Center figure into two major channels. Hyperscale revenue reached $60.136 billion in fiscal Q3 2027, while AI Clouds, Industrial, & Enterprise revenue reached $47.474 billion . Both were higher than in the previous quarter, when Hyperscale revenue was $48.710 billion and AI Clouds, Industrial, & Enterprise revenue was $40.313 billion . That mix suggests the boom is not limited to one category of buyer: the largest cloud platforms are still expanding, while specialized AI-cloud providers, industrial users, and enterprise customers are also absorbing capacity.
The company attributed Data Center growth to Rubin infrastructure, standalone Vera CPUs, continued Blackwell Ultra growth, and what it described as its fastest architecture ramp ever . Nvidia also said in its release that Vera Rubin NVL72 had shipped, and that production had begun with racks running at partners including CoreWeave, Google Cloud, Microsoft Azure, Oracle Cloud Infrastructure, and Nebius . For investors, that product transition is crucial: a clean ramp supports the argument that demand is not simply being pulled forward into one generation, but is continuing across successive AI platforms.
Edge Computing, the other market platform in Nvidia’s recast presentation, produced $8.234 billion in fiscal Q3 revenue, up from $7.198 billion in fiscal Q2 and $5.791 billion in the year-earlier quarter . That growth is real, but it is no longer the center of gravity. The financial story is overwhelmingly about data centers.
The $130 billion guide, with China excluded
The most consequential forward-looking figure was the fourth-quarter outlook. Nvidia guided for $130.0 billion in revenue, plus or minus 3%, and stated that the outlook “is not assuming any Data Center compute revenue from China” . The company also said fiscal Q4 2027 includes an additional week, a detail that should be considered when comparing the guide with a standard 13-week quarter .
That China exclusion is not a footnote. Reuters likewise reported that Nvidia did not assume any data-center chip sales from China in its outlook . The phrasing means the $130 billion midpoint is not dependent on China data-center compute returning to the model. If China contributes nothing in that category, Nvidia still expects another sequential revenue increase from $115.8 billion to about $130.0 billion.
The guide implies a roughly 12% sequential rise at the midpoint, though the extra week makes the operating comparison less clean. It also places the company’s next-quarter revenue range at approximately $126.1 billion to $133.9 billion, based on the stated plus-or-minus 3% band. Nvidia expects GAAP and non-GAAP gross margins between 71.0% and 72.0% in the fourth quarter, below the 74.8% to 74.9% reported for the third quarter . That margin step-down gives the outlook a more nuanced shape: revenue is projected to keep rising, but product mix, ramp costs, supply commitments, and platform transition dynamics may weigh on near-term profitability.
Why the China line matters
China has become a structural variable in Nvidia’s outlook rather than a normal geographic detail. Nvidia disclosed that shipments of Data Center Hopper products to China during the third quarter were $2.4 billion . At the same time, the fourth-quarter guidance excludes any Data Center compute revenue from China . The contrast makes clear that the company is not modeling a continuation of that specific China-related contribution into the current quarter.
For readers trying to interpret the guidance, the key point is not whether Nvidia sells anything at all in China. The specific exclusion concerns Data Center compute revenue from China. That is narrower than total company sales and narrower than all possible revenue involving China-based customers or entities. Still, because Data Center is the company’s dominant platform, the exclusion is material.
The statement also reduces one ambiguity in the quarter’s market reaction. If investors were worried that Nvidia needed China data-center sales to reach a strong fourth quarter, the company’s guidance says otherwise. Conversely, if export-control restrictions ease or compliant products become available later, China could become upside rather than a requirement. That is why the wording of the outlook is strategically important: the $130 billion guide is framed as achievable without that revenue stream.
Profitability remains extraordinary, but the bar keeps rising
Nvidia’s profitability remains unusual even by the standards of high-end semiconductors. GAAP operating income was $77.5 billion in the third quarter, up 22% sequentially and 115% year over year . GAAP net income of $65.3 billion was up 105% year over year . Non-GAAP net income was $64.9 billion, and non-GAAP diluted EPS was $2.68 .
Those numbers show that the company is not merely growing revenue through expensive capacity expansion. It is converting an extraordinary share of sales into operating profit. The challenge is that investors are no longer evaluating Nvidia as a fast-growing chipmaker alone. At this size, each new quarter must answer a larger question: can the AI infrastructure buildout keep expanding fast enough to justify the supply chain, customer financing, data-center power requirements, and capital commitments around it?
Nvidia’s own commentary indicates that demand remains above supply. Transcript coverage quoted management saying that “demand far exceeds supply” . If that remains true, the company may continue to grow into its massive scale. But the financial hurdle is rising: after a $115.8 billion quarter, even a large beat can be treated as only the latest proof point rather than a final answer.
Capital returns and confidence signals
Nvidia also used the quarter to underline its balance-sheet confidence. In September 2026, the board approved a $150 billion increase to the company’s share repurchase authorization, and Nvidia returned $28.1 billion to shareholders through repurchases and dividends during the third quarter . At quarter-end, the company had about $227.3 billion remaining under its repurchase authorization . It also declared a $0.25 quarterly cash dividend payable on December 23, 2026, to shareholders of record on December 2, 2026 .
Large buybacks can be read in two ways. They signal confidence in cash generation, but they also invite scrutiny when a company is simultaneously central to a capital-intensive AI infrastructure cycle. For Nvidia, the current numbers make both interpretations plausible. The business is generating enormous cash and profit, while its ecosystem also depends on expanding supply, securing memory, funding infrastructure, and helping customers deploy systems at speed.
The supplemental commentary said Nvidia had entered into strategic commitments across its supply, infrastructure, and partner ecosystems, and that supply commitments increased from $279 billion in the previous quarter to $390 billion, primarily tied to multi-year memory procurement . That figure highlights the scale of the machine now required to sustain the revenue curve.
The takeaway
The current state of the story is straightforward but historic. Nvidia has crossed $100 billion in quarterly revenue for the first time, reaching $115.8 billion in fiscal Q3 2027 . Its Data Center business alone produced $107.6 billion, more than doubling from a year earlier . The company then guided fiscal Q4 revenue to $130.0 billion, plus or minus 3%, while explicitly excluding any China Data Center compute revenue from that outlook .
That combination is the core news. The company is growing at a pace rarely seen at this scale, and it is telling the market that near-term demand outside China data-center compute is strong enough to support another record quarter. The risks are not absent: gross margin is expected to step down, the fourth quarter has an extra week, China remains politically and commercially complex, and the AI infrastructure cycle must keep absorbing enormous supply. But the numbers show that, as of the latest report, Nvidia’s revenue engine has not slowed. It has moved into a new category.
