NVIDIA earnings highlight infra breadth
- Nvidia reported first-quarter fiscal 2027 revenue of $81.6 billion on May 20, beating estimates as data-center sales rose 92% year over year. (investor.nvidia.com) - The clearest number was $14.8 billion in quarterly data-center networking revenue, up 199% year over year, as InfiniBand deployments and Ethernet expanded. (sec.gov) - Nvidia’s next formal milestone is its June 26 dividend payment, after the company raised the quarterly payout to $0.25 a share. (sec.gov)
Nvidia’s latest quarter did more than extend the company’s revenue streak. The May 20 results showed that the AI buildout is no longer a GPU-only story: networking, memory, CPUs and power-hungry cluster design are all showing up more clearly in the numbers. Nvidia reported first-quarter fiscal 2027 revenue of $81.6 billion, up 85% from a year earlier, with data-center revenue of $75.2 billion, up 92%. (investor.nvidia.com) (sec.gov) The market’s muted reaction after the beat helps explain why investors were split. Expectations had already run high into the print, but the underlying disclosures still pointed to broad infrastructure demand rather than a narrow spike in accelerator sales. (sec.gov) Nvidia’s own breakdown showed data-center compute revenue of $60.4 billion and networking revenue of $14.8 billion in the quarter. ### Why did a big Nvidia beat not produce a bigger stock reaction? Nvidia’s quarterly report arrived after a long run of outsized expectations. The company beat on revenue and posted record data-center sales, yet commentary around the release suggested that investors had already priced in a large part of the upside. (investor.nvidia.com) Motley Fool said the stock reaction was muted even after Nvidia “crushed earnings estimates,” citing elevated expectations going into the report. Jensen Huang used the release to frame the demand picture in broader terms. “The buildout of AI factories … is accelerating at extraordinary speed,” Huang said in the earnings statement, adding that agentic AI was “doing productive work” and scaling across industries. (sec.gov) ### What in the report showed the AI spend is spreading beyond GPUs? Nvidia’s own segment detail gave the clearest answer. Data-center networking revenue reached a record $14.8 billion in the quarter, up 199% from a year earlier and 35% sequentially, while data-center compute revenue rose 77% year over year to $60.4 billion. The earnings-call transcript added more color on what drove that networking line. (investor.nvidia.com) Nvidia executives said InfiniBand revenue more than quadrupled year over year, helped by next-generation XDR deployments, while the company’s broader data-center networking business was nearing a $60 billion annualized run rate. (sec.gov) ### Why does networking matter so much in this cycle? Large AI clusters need to move data between accelerators fast enough to keep expensive GPUs from sitting idle. Nvidia’s results showed customers are spending on that layer at scale, not treating it as a secondary add-on. 24/7 Wall St. said the company’s networking business reflects how modern AI data centers require ultra-fast networking alongside compute. (sec.gov) Nvidia also changed how it reports the business. The company said it is moving to a framework centered on Data Center and Edge Computing, with Data Center split into Hyperscale and ACIE — AI Clouds, Industrial and Enterprise — a structure meant to reflect where future growth is coming from. (fool.com) ### Which other chip suppliers does this help illuminate? SK Hynix offered one of the clearest read-throughs on memory demand. Reuters reported on April 23 that the Nvidia supplier posted record quarterly profit and said AI memory demand would exceed manufacturing capacity, easing concerns about slowing spending by big tech customers. (fool.com) That matters because Nvidia’s systems pull through more than GPUs. High-bandwidth memory, interconnects, switches and CPUs all sit inside the same buildout. Nvidia executives said on the call that they anticipate $20 billion in standalone CPU revenue for the year from the Vera platform, underscoring that compute architecture choices are broadening inside AI infrastructure. (sec.gov) ### What is the cleanest takeaway from this quarter? The quarter showed that AI infrastructure spending is widening across the stack. Nvidia generated $81.6 billion in revenue, returned about $20 billion to shareholders during the quarter, approved an additional $80 billion in buybacks and raised its quarterly dividend from $0.01 to $0.25 a share. (whbl.com) The next dated item in the company’s capital-return plan is June 26, when Nvidia said it will pay the higher dividend to shareholders of record as of June 4. Investors will likely look to the next quarterly report for updated figures on compute, networking and the new Hyperscale and ACIE reporting lines. (sec.gov) (fool.com)