Nvidia’s Quarterly Revenue Tops $100 Billion, Chinese Computing-Power Stocks Rally
Nvidia reported third-quarter revenue of approximately $115.8 billion, up 103% year over year, and net income of $65.287 billion, up 105%. Data center revenue reached $107.6 billion, accounting for approximately 93% of total revenue, and the company expects fourth-quarter revenue of about $130 billion. The report also said Nvidia tightened financing guarantees while continuing to expand capacity, as stocks related to Chinese-made chips, semiconductor foundries, and advanced packaging rallied.
Why read this
SelectedAI score 76
The earnings report places Nvidia’s data center revenue, capacity-expansion commitments, and tightened financing guarantees within the same framework, highlighting simultaneous shifts in the computing-power supply chain’s strength and risk controls.
Topics
Body · AI translation
Industry observer Liu Chao said Chinese model companies need to shift to domestically produced chips, making demand for domestic chips and wafer foundry services more certain, and recommended watching segments constrained by capacity bottlenecks.
On November 18, Nvidia released its first earnings report since Jensen Huang’s departure. Quarterly revenue exceeded $100 billion for the first time, while revenue from its core data center business reached $107.6 billion, up 110% year on year.
Third-quarter revenue reaches $115.8 billion
In the third quarter, Nvidia generated revenue of approximately $115.8 billion, up 103% year on year and 20% quarter on quarter. On a GAAP basis, net income was $65.287 billion, up 105% year on year, while diluted earnings per share were $2.70. Non-GAAP gross margin was 74.9%, above the market’s previous expectation of approximately 74%. Among its main businesses, data center revenue totaled $107.6 billion, up 110% year on year and 21% quarter on quarter, accounting for approximately 93% of total revenue.
Looking ahead to the fourth quarter, Nvidia expects revenue of approximately $130 billion, plus or minus 3%, and GAAP and non-GAAP gross margins of 71%–72%. It raised its revenue growth forecast for fiscal 2028 to approximately 75% year on year.
As of the end of the third quarter, Nvidia’s commitments related to agreements with emerging cloud providers totaled $32 billion, while total guarantees stood at $118 billion at quarter-end. However, management denied that its guarantees had changed, saying it had merely further clarified how they operate. “We will examine very carefully which companies merit financing support,” said interim CEO and CFO Colette Kress.
Even as it tightens financing guarantees, Nvidia continues to treat capacity expansion as its top priority, with purchase commitments rising to $390 billion in the third quarter. The company’s next-generation Vera Rubin NVL72 and Vera CPU entered mass production and began shipping during the third quarter. It also announced new agreements with substrate suppliers Kinsus Interconnect Technology and Samsung Electro-Mechanics, as well as advanced-packaging company SPIL.
By business segment, Nvidia’s networking revenue exceeded $21 billion in the third quarter. Its co-packaged optics roadmap is progressing as planned, and CPO scale-out products have entered mass production. Under its technology roadmap, Nvidia plans to launch Vera Rubin Ultra next year and Feynman in 2028.
Company adjusts the weighting of strategic decision-making
Although Jensen Huang expressed a desire to return as Nvidia’s CEO, the company’s new management firmly rejected the proposal, stating that it had no intention of reinstating him as CEO, and began the selection process for a new chief executive. The board also established a National Security and Export Control Oversight Committee comprising several independent directors, senior external advisers, and legal counsel. The committee is responsible for strengthening communication between the company, the U.S. government, and all relevant agencies.
In addition to the personnel changes, Nvidia clarified its approach to supply-chain financing and guarantees during its earnings call. It will no longer provide guarantees for data centers during construction, activating them only once facilities are ready; guarantees will automatically lapse if construction is delayed. Financing support for new cloud providers will be “selective and more prudent,” with every transaction subject to rigorous review.
Gu Zhengshu, chief semiconductor industry analyst at Moyuan Think Tank, said Nvidia’s newly established National Security and Export Control Oversight Committee indicates that the weighting of the company’s strategic decisions is shifting from “prioritizing the technology roadmap” toward “prioritizing compliance and risk control.” He added that Nvidia’s tighter financing policy does not mean it is “stopping support for customers,” but rather that supplier commitments and repurchases will take precedence over customer financing guarantees. As financing conditions tighten for marginal buyers, however, their pace of expansion risks slowing.
Domestic chips gain market opportunities
Against the backdrop of tighter overseas AI regulation and sales restrictions, Chinese computing-power stocks have recently strengthened. On November 18, the A-share GPU index rose more than 2%. Moore Threads gained more than 7%, while Cambricon rose 4.89%, with turnover reaching RMB 11.3 billion, the highest on the STAR Market. Leading advanced-packaging and domestic foundry stocks also rallied: SJ Semiconductor rose approximately 11%, while Hua Hong Semiconductor surged about 7%. Nvidia supply-chain stocks also strengthened that day, with Victory Giant Technology climbing approximately 6%, Foxconn Industrial Internet and Hongbo Co. each gaining more than 5%, and TFC Optical Communication rising 4.6%.
Industry observer Liu Chao said Chinese model companies need to shift to domestically produced chips, making demand for domestic chips and wafer foundry services more certain, and recommended watching segments constrained by capacity bottlenecks. By contrast, cloud providers and model companies that rent Nvidia computing power overseas may be affected.
At a recent earnings briefing, Hua Hong Semiconductor executives said this year’s AI boom has been exceptionally strong, spreading from memory into specialty processes and logic technologies, and that the company’s performance growth has benefited substantially from artificial intelligence. According to the company’s assessment, more than 50% of its end products have been boosted by AI development, with AI-related end markets growing very rapidly. On the other hand, high memory-product prices have dampened consumer demand, producing an offsetting effect overall.
Tencent executives acknowledged during the company’s third-quarter 2026 earnings call that the amount of computing power actually available is currently declining and that model training will be affected in the short term. Over the longer term, more Chinese companies will rely on domestic supply chains and domestically produced chips, while the latest restrictions are expected to accelerate the development of China’s supply chains.
AI Trends