Nvidia Earnings Highlight Supply-Constrained Growth

Nvidia (NVDA) reported on Aug. 26, 2026, strong growth in AI data-center sales and unveiled its first full-year FY2028 guidance framework, which it described as supply-constrained. Combined with robust partner demand, this outlook suggests significant upside potential despite margin pressure from rising memory costs.

Quarter Results and Guidance

For the quarter ended July 26, 2026, Nvidia posted revenue of $96.2 billion, up about 106% year over year, with adjusted non-GAAP earnings per share of $2.22, exceeding consensus estimates near $2.09–$2.10. AI data-center revenue reached $89.0 billion, rising about 117% year over year and 18% sequentially, accounting for roughly 92–93% of total sales. Hyperscale cloud customers generated $48.7 billion, while AI clouds, industrial, enterprise, and sovereign AI (ACIE) contributed $40.3 billion, with ACIE as the fastest-growing segment. Non-hyperscaler customers, including sovereign and enterprise buyers, “neo clouds,” and industrial users, now represent about half of data-center revenue and are expanding faster than hyperscalers.

GAAP gross margin rose to about 75%, up from roughly 72.7% a year earlier. Net margin and return on equity remained above 60% and 110%, respectively. Nvidia guided Q3 FY2027 revenue to $108.0 billion, plus or minus 2%, assuming no data-center compute revenue from China due to export restrictions. It set GAAP and non-GAAP gross margins at about 74%, noting margins could trough near 71–72% in the fiscal fourth quarter. This guidance implies roughly 89% year-over-year revenue growth for the quarter.

Supply Commitments, Demand, and Margins

Nvidia provided its first full-year FY2028 guidance framework, targeting approximately 70% revenue growth year over year. Management described this outlook as supply-constrained, primarily by memory availability and cost. This growth rate exceeds analysts’ models by about 26 percentage points, implying FY2028 revenue near $690–700 billion versus prior consensus around $570 billion.

Supply and capacity commitments rose sharply to roughly $279–350 billion by quarter end, up from about $119 billion the previous quarter. Management expects supply constraints to persist through at least the end of FY2028. CFO Colette Kress cited “extreme pricing conditions in memory,” with price increases exceeding earlier expectations, pressuring gross margins. The company’s outlook reflects HBM4 pass-through, passing some memory cost increases to customers, with an implied FY2028 gross-margin range near 72–73%.

Partner demand signals reinforce this outlook. Dell Technologies reported record AI server orders of about $60.9 billion, AI server revenue near $16.4 billion, and an AI backlog around $95 billion. Dell guided to roughly $50 billion in AI revenue for FY27, with AI-optimized servers up about 100% year over year. Dell emphasized a tight supply environment, prioritizing securing components and pricing to protect margins.

Dell also highlighted AI network architectures optimized for GPU-to-GPU “elephant flows,” using flow-control technologies like PFC/DCQCN with RoCEv2 to maximize utilization for large AI training workloads. It was the first to ship rack systems engineered on Nvidia’s Vera Rubin platform, which Nvidia indicated began production shipments during the quarter and could represent a substantial share of upcoming data-center revenue. Nvidia now sells both chips and a full software stack for AI infrastructure, increasing customer lock-in. Hyperscalers continue to ramp Blackwell Ultra and other accelerators, supporting sustained demand.

Together, Nvidia’s FY2028 framework, rising supply commitments, and partner order books suggest materially higher long-term AI revenue potential than many models assume, even as HBM-related cost pressures temper margin expansion. Despite repeated fundamental beats, Nvidia’s shares have fallen after each of its last five quarterly earnings announcements, reflecting valuation multiple adjustments amid rapid revenue and profit growth.