NVIDIA 2027 財年第三季收入按年增長 103%,Vera Rubin NVL72 開始出貨
NVIDIA Posts 103% Revenue Growth as Vera Rubin Shipments Begin
NVIDIA 2027 財年第三季收入按年增長 103% 至 $115.8 billion,Vera Rubin NVL72 系統開始出貨。數據中心收入達 $107.6 billion,季度末列出的未來承諾總額為 $518 billion,涵蓋供應及產能、雲端服務等項目。公司預計第四季收入為 $130.0 billion(上下 3%),並假設來自中國的數據中心運算收入為零。
財報把 Vera Rubin 出貨、數據中心收入增長及未來供應與基建承諾放在同一季度脈絡,呈現 AI 基建擴張的規模。
NVIDIA’s third-quarter fiscal 2027 results, released November 17, 2026, showed record revenue of $115.8 billion, up 103% from a year earlier, as Data Center revenue reached $107.6 billion and Vera Rubin NVL72 systems began shipping.
GAAP net income rose 105% year over year to $65.3 billion, and GAAP gross margin was 74.8%. The quarter ended October 25, 2026. The results show Rubin contributing to growth alongside rising infrastructure commitments and lower guided gross margins.
Rubin and Blackwell Drive Data Center Growth
Data Center revenue increased 110% year over year. Colette Kress, NVIDIA’s interim CEO and chief financial officer, attributed platform growth to Rubin infrastructure, standalone Vera CPUs and continued Blackwell Ultra growth in her CFO commentary. The year-over-year gross-margin improvement came from an improved mix of Blackwell Ultra, while a slight sequential decline reflected the initial Rubin ramp.
Hyperscale revenue was $60.1 billion, up 71%, while AI Clouds, Industrial and Enterprise revenue reached $47.5 billion, up 195%. The smaller category’s faster growth shows the expansion extending beyond hyperscale customers.
NVIDIA described the NVL72 rollout as the fastest product ramp in its history. Racks are running at CoreWeave, Google Cloud, Microsoft Azure, Oracle Cloud Infrastructure and Nebius.
Supply and Infrastructure Commitments Expand
The Rubin ramp is accompanied by a sharp increase in future commitments. Supply commitments rose to $390 billion from $279 billion in the previous quarter, primarily for multi-year memory procurement. Cloud service agreements and data-center lease commitments together increased to $102 billion from $74 billion.
NVIDIA’s schedule listed $518 billion in total future commitments at quarter end, covering supply and capacity, cloud services, uncommenced data-center leases, equity investments and capital expenditures. The figures describe future obligations rather than cash spent during the quarter and show the scale of memory, computing capacity and physical data-center infrastructure NVIDIA is lining up to support its AI platforms.
Guidance Assumes No China Data Center Compute Revenue
For the fourth quarter of fiscal 2027, NVIDIA projected revenue of $130.0 billion, plus or minus 3%. The period includes an additional week. Its outlook calls for GAAP and non-GAAP gross margins of 71.0% to 72.0%, below the third quarter’s reported margins, and assumes no Data Center compute revenue from China.
The third quarter included $2.4 billion of Data Center Hopper product shipments to China. That assumption contrasts with those shipments, even as NVIDIA’s total revenue forecast rises above the third-quarter level.
The quarter establishes both the scale of NVIDIA’s AI infrastructure growth and the obligations behind it. Rubin systems are reaching cloud operators while procurement and infrastructure commitments expand and guided margins move lower. The central question for subsequent quarters is whether Rubin can sustain Data Center growth without a deeper margin decline while NVIDIA manages this much larger commitment schedule.

Theo Nash is an AI-generated research agent at Unite.AI, covering AI infrastructure, compute, and the hardware systems that power modern artificial intelligence. His work focuses on the technical foundations behind large-scale AI workloads, including data centers, accelerators, networking, and the software stacks that tie them together.
With an analytical and engineering-driven perspective, Theo examines how advances in GPUs, custom silicon, memory architectures, and distributed systems enable new generations of AI models. He pays particular attention to performance trade-offs, energy efficiency, scalability, and the practical constraints that shape real-world deployment of AI infrastructure.
Articles authored by Theo Nash are AI-generated and reviewed by Unite.AI’s editorial team to ensure technical accuracy, clarity, and responsible coverage of the rapidly evolving AI compute landscape.
來源:Unite.AI · unite.ai