NVIDIA has once again presented staggering figures. The company announced record-breaking results for its fourth quarter of fiscal year 2026 (ended on January 25, 2026) and for the entire fiscal year, with a message that aims to go beyond the financial headline: the company believes the market has entered a new cycle, the era of Generative AI, where inference becomes the driving force and the “cost per token” turns into a strategic metric.
In the quarter, NVIDIA reported $68.127 billion in revenue, a 20% increase from the previous quarter and a 73% rise compared to a year ago. For FY2026, revenues totaled $215.938 billion, representing a 65% year-over-year growth. Gross margin remained at very high levels: 75.0% (GAAP) and 75.2% (non-GAAP) for the quarter, and 71.1% (GAAP) and 71.3% (non-GAAP) for the fiscal year.
In diluted earnings per share, the company reported $1.76 (GAAP) and $1.62 (non-GAAP) for the quarter; and $4.90 (GAAP) and $4.77 (non-GAAP) for the full year. This snapshot reinforces Jensen Huang, founder and CEO,’s consistent message: “computing demand is growing exponentially” and the industry is entering a tipping point where agents adopt enterprise-scale solutions en masse.
Data Center: the Driving Force (and Showcasing Blackwell and Rubin)
If there’s one number that sums up NVIDIA’s moment, it’s Data Center. The segment reported $62.3 billion in the fourth quarter, a 22% quarter-over-quarter increase and a 75% increase year-over-year. For FY2026, the segment reached $193.7 billion, a 68% increase over the previous year. Essentially, the data center is no longer “a segment”: it is the core economic driver of the company.
The technological narrative aligns with this reality. Huang stated that Grace Blackwell with NVLink is “the king of inference” and advocated for a “magnitude order” reduction in the cost per token. Meanwhile, Vera Rubin will extend that advantage. As part of their milestones, NVIDIA reiterated their Rubin platform, describing it as a set of six new chips aimed at reducing inference costs per token by up to 10 times compared to Blackwell. The company also announced that cloud providers like AWS, Google Cloud, Microsoft Azure, and Oracle Cloud Infrastructure will be among the first to deploy Rubin-based instances.
Additionally, NVIDIA highlighted infrastructure innovations indicating a shift in approach: it’s no longer just about accelerating computation, but about building a complete “factory.” In this vein, they introduced BlueField-4 as the foundation of a new storage platform designed for “context memory” in inference (Inference Context Memory Storage Platform), recognizing that agent performance depends on both computing and data pathways, networks, and memory/storage access.
Next Quarter Outlook: $78 Billion and a Key Detail on China
NVIDIA also provided guidance for Q1 FY2027: $78 billion in revenue, with a margin of ±2%. The company added an important note: they are not including data center compute revenue from China in this outlook. Regarding margins, they expect 74.9% (GAAP) and 75.0% (non-GAAP), with operating expenses around $7.7 billion (GAAP) and $7.5 billion (non-GAAP).
They also announced a significant accounting change for analysts: starting from Q1 FY2027, NVIDIA will include stock compensation in its non-GAAP metrics. This adjustment aims to provide a more comprehensive view of recurring costs in a talent-intensive company.
Capital Return: $41.1 Billion and a New Dividend
Alongside growth, NVIDIA reported a substantial capital return: during FY2026, it returned $41.1 billion to shareholders through buybacks and dividends, and it maintains $58.5 billion available under its buyback authorization.
Additionally, it will pay a quarterly dividend of $0.01 per share on April 1, 2026, to shareholders of record as of March 11, 2026.
Gaming, RTX PRO, and Automotive: Parallel Growth with Nuances
Although the market focus is on AI, other divisions also provided signals of strength.
- Gaming: quarterly revenue of $3.7 billion, a 47% year-over-year increase, but down 13% from the previous quarter due to natural inventory moderation after the holiday season. For the full year, Gaming grew 41% to $16 billion.
- Professional Visualization: $1.3 billion in the quarter (a 74% quarterly increase and 159% annual growth) and $3.2 billion for the year (up 70%). The company highlighted the release of RTX PRO 5000 72GB, based on Blackwell, targeting high-end models and scalable agent workloads.
- Automotive and Robotics: $604 million this quarter (+2% QoQ, +6% YoY) and $2.3 billion for the year (+39%). NVIDIA emphasized its DRIVE ecosystem, partnerships with manufacturers, and open models and tools for robotics and “physical AI.”
Overall, the narrative remains consistent: the data center drives volume, while other divisions reinforce the platform strategy (software, tools, open models, industry collaborations) to sustain expansion.
Frequently Asked Questions (FAQ)
How much did NVIDIA earn in Q4 FY2026 and how does that compare to last year?
It posted $68.127 billion in revenue, up 73% year-over-year and 20% from the previous quarter.
What was NVIDIA’s largest business segment in FY2026 and what was its figure?
The Data Center segment was the largest, with $193.7 billion for FY2026 and $62.3 billion in Q4 alone.
What guidance did NVIDIA give for Q1 FY2027 and what about China?
They guided $78 billion (±2%) in revenue, specifying that they are not assuming data center compute revenue from China in this forecast.
When will NVIDIA pay its next dividend and what is the amount?
It will pay $0.01 per share on April 1, 2026 (record date: March 11, 2026).