Nvidia Q4 Earnings Data Center AI Drives Surge

NVIDIA Corp. reported record fourth-quarter revenue on Feb. 25, 2026, driven by accelerating data-center AI sales. The company unveiled the NVIDIA Rubin platform and issued first-quarter guidance that frames fiscal 2027 margins and revenue expectations.

Quarter Results and Data Center Revenue

Nvidia posted fourth-quarter revenue of $68.1 billion, up 20.0% sequentially and 73.0% year-over-year. Data Center revenue reached $62.3 billion, rising 22.0% sequentially and 75.0% year-over-year. For fiscal 2026, Data Center revenue totaled $193.7 billion, a 68.0% increase from the prior year.

GAAP gross margin for the quarter was 75.0%, with a non-GAAP margin of 75.2%. GAAP net income was $43.0 billion, with diluted earnings per share of $1.76. For the full fiscal year, Nvidia reported revenue of $215.9 billion, up 65.0% year-over-year, and GAAP net income of $120.1 billion, with diluted EPS of $4.90. The annual GAAP gross margin was 71.1%, with a non-GAAP margin of 71.3%.

Other segments were smaller: gaming revenue was $3.7 billion, down 13.0% sequentially but up 47.0% year-over-year; professional visualization revenue rose 74.0% sequentially and 159.0% year-over-year to $1.3 billion; automotive revenue increased 2.0% sequentially and 6.0% year-over-year to $604 million.

Guidance, Products, and Shareholder Returns

Nvidia forecasted first-quarter fiscal 2027 revenue of $78.0 billion, plus or minus 2.0%, excluding any Data Center compute revenue from China. The company projected a GAAP gross margin of 74.9% ±50 basis points and a non-GAAP margin of 75.0% ±50 basis points, which includes a 0.1% stock-based compensation impact. Operating expenses are expected to total about $7.7 billion on a GAAP basis and $7.5 billion non-GAAP, including roughly $1.9 billion in stock-based compensation. This quarter marks the first time Nvidia includes stock-based compensation in its non-GAAP measures.

Nvidia introduced the Rubin platform, which comprises six new chips designed to reduce inference token costs by up to 10 times compared with the Blackwell platform. The company also announced the RTX PRO 5000 72GB Blackwell GPU and Alpamayo AI models for autonomous vehicles. Cloud partners deploying Vera Rubin include AWS, Google Cloud, Microsoft Azure, and Oracle Cloud Infrastructure. Nvidia named Cadence, Siemens, and Synopsys as collaborators for AI efforts in India.

In fiscal 2026, Nvidia returned $41.1 billion to shareholders through repurchases and dividends. The company has $58.5 billion of remaining repurchase authorization. The board declared a $0.01 per-share dividend payable April 1, 2026.

With Data Center revenue accounting for the bulk of sales, the Rubin platform’s efficiency gains and the company’s guidance position near-term demand in that business as the key factor shaping Nvidia’s fiscal 2027 outlook.