Nvidia's $81.6bn Quarter: Networking Surges, China Goes to Zero
· TexaHub · Semiconductors, Geopolitics & Tech
Nvidia beat estimates and guided to $91bn, but the detail sits in networking growth, a new segment structure, CPU ambitions and a China business that has vanished.
The numbers
Nvidia reported revenue of $81.6 billion for the first quarter of fiscal 2027, which ended on 26 April 2026, up 85% from a year earlier and 20% from the previous quarter [1]. Data Center revenue was $75.2 billion, up 92% year on year [1]. GAAP and non-GAAP gross margins were 74.9% and 75.0%, and non-GAAP diluted earnings per share were $1.87 [1].
Against LSEG consensus figures cited by Reuters, revenue beat an estimate of $78.86 billion, Data Center revenue beat an estimate of $72.8 billion, and adjusted earnings per share beat an estimate of $1.76 [3]. The company guided second-quarter revenue to $91.0 billion, plus or minus 2%, against an LSEG estimate of $86.84 billion [1] [3]. Shares still fell 1.6% in extended trading [3].
Capital returns were also raised. The board approved an additional $80.0 billion of share repurchase authorization and lifted the quarterly dividend from $0.01 to $0.25 per share, after the company returned about $20.0 billion to shareholders during the quarter [1].
Networking is the faster engine
Inside Data Center, compute revenue was $60.4 billion, up 77% year on year and 18% sequentially, while networking revenue was $14.8 billion, up 199% year on year and 35% sequentially [2]. The CFO commentary attributes the growth to the ramp of Blackwell 300 products and to demand for InfiniBand, Spectrum-X Ethernet and NVLink [2].
The proportions matter more than the totals. Networking growing at more than twice the rate of compute suggests customers are buying complete rack-scale and cluster-scale systems, where the interconnect is a larger share of the bill, rather than accelerators alone. That should make Nvidia's revenue more dependent on large integrated deployments and less on individual chip sales, which raises the stakes on execution of its rack-scale roadmap.
Nvidia is also building inventory for that roadmap. Inventory rose to $25.8 billion from $21.4 billion in the prior quarter, and total supply-related commitments reached $119.0 billion, which the company said reflected inventory and capacity secured "to meet demand beyond the next several quarters" [2]. GAAP operating expenses were $7,621 million, up 52% year on year [2].
A new way of reporting
The quarter introduced a new segment structure with two market platforms, Data Center and Edge Computing [2]. Data Center is split into Hyperscale, covering public clouds and large consumer internet companies, and ACIE, covering AI clouds, industrial and enterprise customers [2]. Edge Computing covers PCs, game consoles, workstations, AI-RAN base stations, robotics and automotive, and generated $6.4 billion in the quarter, up 29% year on year [1] [2].
Nvidia said hyperscalers accounted for approximately 50% of Data Center revenue, with the remainder coming from AI clouds, industrial, enterprise and sovereign customers [2]. TrendForce put the split at $38 billion for hyperscalers and $37 billion for ACIE [5]. For investors, the disclosure offers the first official measure of how concentrated demand is among the largest cloud buyers, a long-running concern as those buyers develop their own accelerators.
Reuters reported that Huang told analysts Nvidia "should be growing faster than hyperscale capex" and that he expected to be "supply-constrained through the entire life of Vera Rubin" [3]. Those are company expectations rather than commitments, and their credibility will depend on how the Rubin ramp proceeds later this year.
The CPU pitch
The most notable new claim concerned Vera, Nvidia's Arm-based CPU. Huang said Vera "opens a brand new $200 billion TAM for Nvidia", a market the company has not previously addressed, and described it as "the world's first CPU, purpose-built for agentic AI" [4]. Reports differ on how to characterise the accompanying figure of about $20 billion: TechCrunch reported it as $20 billion of standalone Vera CPUs sold this year, while TrendForce reported that CPU revenue could reach $20 billion in 2026 [4] [5].
The argument rests on agentic workloads shifting more computation onto general-purpose processors. Nvidia is not alone in making it. In April, Meta agreed to bring "tens of millions of AWS Graviton cores" into its compute portfolio, saying agentic AI compute requirements "are evolving to demand more CPU" [6]. If that view holds, the server CPU market becomes a new front in which Nvidia competes not only with Intel and AMD but with its own hyperscale customers' in-house Arm designs [4].
China and what to watch
China has dropped out of the Data Center figures. Nvidia said no Data Center Hopper products were shipped to China in the quarter, compared with $4.6 billion in the same quarter a year earlier, and its outlook assumes no Data Center compute revenue from China [2]. TrendForce reported Huang as saying Nvidia has "really largely conceded that market" to Huawei, which he described as "very, very strong" [5]. Year-on-year margin comparisons are flattered by the prior year's $4.5 billion charge related to H20 excess inventory and purchase obligations [2].
The result leaves three questions for the coming quarters: whether networking growth continues to outpace compute as Rubin systems arrive, whether the 50% hyperscale share moves as those customers expand their own silicon, and whether the CPU revenue Nvidia has described appears in the new Data Center disclosures. The company's guidance implies continued sequential growth, but the after-hours share reaction indicates that investors are now weighing durability rather than the size of the beat [3]. As eMarketer analyst Jacob Bourne put it to Reuters, "The lingering question is whether it can convince investors the AI buildout has durability into 2027 and 2028" [3].
Sources
[1] NVIDIA Announces Financial Results for First Quarter Fiscal 2027 — NVIDIA, 20 May 2026
[2] CFO Commentary on First Quarter Fiscal 2027 Results — NVIDIA, 20 May 2026
[3] Nvidia forecasts revenue above estimates, announces $80 billion share buyback — Reuters, 20 May 2026
[4] Jensen Huang says he's found a 'brand new' $200B market for Nvidia — TechCrunch, 20 May 2026
[5] [News] NVIDIA Earnings Spotlight: $20B CPU Sales Target in 2026, China Concession to Huawei, and Reporting Overhaul — TrendForce, 21 May 2026
[6] Meta Partners With AWS on Graviton Chips to Power Agentic AI — Meta, 24 Apr 2026