NVIDIA Q4 FY26 earnings: $68 billion revenue proves AI boom is far from over
NVIDIA smashed Q4 FY26 estimates with $68B revenue, $62.3B from data centers. Full breakdown of results, Vera Rubin, and what it means for AI.
Whether you're looking for an angel investor, a growth advisor, or just want to connect — I'm always open to great ideas.
Get in TouchAI, startups & growth insights. No spam.
TL;DR: NVIDIA reported $68.1 billion in Q4 FY26 revenue, up 73% year over year, with data center sales alone hitting $62.3 billion. Q1 FY27 guidance came in at $78 billion, blowing past the $72 billion Wall Street expected. The AI spending cycle shows no sign of slowing.
NVIDIA reported its Q4 FY26 results on February 25, 2026. The numbers were not just good. They were company records, quarter after quarter, in an era where the bar keeps getting raised.
Revenue for the quarter came in at $68.13 billion, a 73% jump from $39.3 billion in Q4 FY25, and 20% above Q3 FY26. Wall Street had penciled in $66.2 billion. NVIDIA cleared it by nearly $2 billion.
GAAP earnings per share hit $1.76. Non-GAAP EPS came in at $1.62, well ahead of the $1.53 analysts had expected.
Net income almost doubled, reaching $42.96 billion, compared to $22.1 billion in Q4 FY25. That is a 94% increase in a single year.
Gross margin held up at 75.0% on a GAAP basis and 75.2% on a non-GAAP basis. That is an improvement of 160 basis points sequentially over Q3 FY26's 73.6%, and 170 basis points above the year-ago period.
Here is a quick look at how the actual results compared to what Wall Street expected:
| Metric | Estimate | Actual | Result |
|---|---|---|---|
| Revenue | $66.2B | $68.13B | ✓ Beat |
| Non-GAAP EPS | $1.53 | $1.62 | ✓ Beat |
| Data Center Revenue | $60.69B | $62.3B | ✓ Beat |
| Net Income | $35.8B | $42.96B | ✓ Beat |
| Q1 FY27 Guidance | $72.0B | $78.0B | ✓ Beat |
Every single line item beat. That is not an accident at this point. NVIDIA has now delivered a string of quarters where the company did not just meet expectations but reset them.
Data center revenue was $62.3 billion for the quarter, up 75% year over year and 22% from Q3 FY26. This single segment now accounts for over 91% of NVIDIA's total revenue.
The driver is Blackwell. NVIDIA's current-generation GPU architecture, named after mathematician David Blackwell, shipped in record volume through the quarter. Hyperscalers including Microsoft, Google, Amazon, and Meta have been racing to build out AI training and inference infrastructure, and they are all doing it with NVIDIA chips.
The demand dynamic here is important. Companies are not just buying GPUs to train large language models. They are buying them to run inference at scale, to power AI agents, and to process increasingly complex workloads in real time. That inference use case alone creates a durable demand floor that did not exist three years ago.
Gaming revenue came in at $3.7 billion, up 47% year over year. That is a meaningful number, and it shows the consumer GPU market is healthy. But compared to the data center story, it is almost a footnote.
Gaming was down 13% sequentially from Q3, which typically reflects post-holiday seasonality. The GeForce RTX 50 series is in the market, and demand for AI-accelerated gaming features continues to grow.
Automotive contributed $604 million in revenue, up 6% year over year. For the full fiscal year, automotive hit $2.3 billion, up 39%.
This segment is worth watching. NVIDIA's DRIVE platform is being adopted by an expanding list of automakers for autonomous driving and in-vehicle AI. The numbers are small relative to data center, but the growth trajectory is real and the market is early.
Professional visualization added $511 million in Q4, up 10% year over year. Steady, not spectacular.
Stepping back to the full fiscal year gives you a sense of how extraordinary NVIDIA's run has been.
| Metric | FY25 | FY26 | Change |
|---|---|---|---|
| Total Revenue | $130.5B | $215.9B | ↑65% |
| Data Center Revenue | $115.2B | $193.7B | ↑68% |
| Gaming Revenue | $11.4B | $16.0B | ↑41% |
| Automotive Revenue | $1.7B | $2.3B | ↑39% |
| Net Income | $72.9B | $120.1B | ↑65% |
| GAAP EPS | $2.94 | $4.90 | ↑67% |
$215.9 billion in annual revenue. $120.1 billion in annual net income. Those are numbers that put NVIDIA in company with a very short list of companies in the history of public markets.
To put the data center growth in context: NVIDIA's CEO noted on the earnings call that the data center business has grown 13x since ChatGPT launched in late 2022. That is not a trend line. That is a step change in how the world buys compute.
Jensen Huang opened his prepared remarks with a declaration that AI has crossed a new threshold.
"Computing demand is growing exponentially. The agentic AI inflection point has arrived. Grace Blackwell with NVLink is the king of inference today, delivering an order-of-magnitude lower cost per token, and Vera Rubin will extend that leadership even further."
He went further on the call, framing compute as the new currency of the AI economy:
"Compute is revenues. Without compute, there is no way to generate tokens. Without tokens, there is no way to generate revenue. So in this new world of AI, compute equals revenues."
That last line is the clearest articulation yet of why enterprise AI spending is not discretionary. If your product generates tokens to generate revenue, compute is cost of goods sold, not a capital expenditure you can delay.
The "agentic AI" framing matters. The shift from AI models that answer questions to AI agents that take autonomous actions across systems represents a qualitative change in compute requirements. Agents run inference continuously, not just when a user types something. That means more GPU-hours per deployment, which means more NVIDIA chips.
Every NVIDIA earnings call now includes a chapter on what comes next, and this one was no exception.
CFO Colette Kress confirmed on the call that NVIDIA "shipped our first Vera Rubin samples to customers earlier this week." Vera Rubin is the next-generation rack-scale platform that succeeds Grace Blackwell. The platform comprises six new chips, and NVIDIA says it will deliver up to a 10x reduction in inference token cost compared to Blackwell.
That number, if it holds, is significant. Lower cost per token means AI products become economically viable at larger scale, which means more deployment, which means more demand for the next chip generation. NVIDIA has structured its product roadmap as a compounding flywheel.
Vera Rubin is expected to reach broader availability later in 2026. The fact that samples are already in customer hands suggests the timeline is on track.
The stock reaction was measured. NVIDIA shares initially rose about 3% in after-hours trading after the results dropped on February 25. During the conference call itself, shares dipped as much as 1.5%, and by end of evening the stock was roughly flat.
That might seem like a muted response to a record quarter. But it reflects something important about how NVIDIA trades right now. The expectations were already extremely high. Beating a $66 billion consensus by $2 billion is impressive. It is also, at this point, roughly what the market expects NVIDIA to do.
The $78 billion guidance for Q1 FY27 was the real news. Analysts had expected $72 billion. Coming in 8.3% above consensus on forward guidance is not a small thing. It tells you that the hyperscaler capex cycle is accelerating, not plateauing.
"Nvidia hits earnings record, Jensen Huang touts AI 'inflection point'" -- @axios
"Nvidia smashes Q4 with $68 billion in revenue, and a Q1 outlook that quashes AI bubble talk" -- @FortuneMagazine
The analyst community came away constructive. Eleven analysts revised their earnings estimates upward for the coming period following the results.
UBS raised its price target from $235 to $245 and reiterated a "Buy" rating. KeyBanc analyst John Vinh maintained an "Overweight" rating with a $275 price target.
The broad theme across analyst notes was that the Q1 guidance settled the debate, at least for now, about whether AI capex was about to roll over. The hyperscalers have not blinked. Microsoft, Google, Amazon, and Meta have all provided capex guidance in recent weeks that points to continued or accelerating spend on AI infrastructure through 2026. NVIDIA is the primary beneficiary of that spend.
One note of complexity: starting in Q1 FY27, NVIDIA will include stock-based compensation in its non-GAAP results, aligning with evolving accounting practices. That changes the comparability of non-GAAP EPS going forward. Investors will need to account for that when modeling future quarters against current consensus estimates.
The $78 billion Q1 FY27 guidance is the number that tells you the most about where NVIDIA thinks its business is going.
The midpoint of $78 billion represents an additional 14.5% sequential growth from Q4 FY26's $68.1 billion. That is a company adding roughly $10 billion in quarterly revenue in a single quarter, on top of an already record base.
To put that in perspective: Apple, one of the largest companies in the world by revenue, generates roughly $90-95 billion per quarter. NVIDIA is approaching that level, and it is a chip company that did not exist in its current form five years ago.
The implied full-year FY27 run rate, if NVIDIA can sustain and grow from $78 billion, would push annual revenue north of $300 billion.
The ±2% range on that guidance is also worth noting. NVIDIA has historically been conservative in how it guides. A range centered on $78 billion, with upside optionality, is not a management team that sees softness coming.
What could cause a miss? The two plausible scenarios are export controls tightening further, which would limit sales to China and certain other markets, and a faster-than-expected deceleration in hyperscaler capex. Neither appears imminent based on current signals, but both are real risks.
NVIDIA reported $68.13 billion in total revenue for Q4 FY26, the quarter ending January 26, 2026. That beat the analyst consensus estimate of $66.2 billion and represented 73% growth from $39.3 billion in Q4 FY25.
Data center revenue hit $62.3 billion in Q4 FY26, up 75% year over year and 22% sequentially from Q3 FY26. It was the single largest contributor, accounting for over 91% of NVIDIA's total quarterly revenue. The segment was driven by continued strong demand for Blackwell GPUs from hyperscalers and cloud providers.
NVIDIA posted GAAP diluted EPS of $1.76 and non-GAAP EPS of $1.62 for Q4 FY26. The non-GAAP figure beat analyst expectations of $1.53. The GAAP net income for the quarter was $42.96 billion, up 94% from $22.1 billion in Q4 FY25.
NVIDIA guided for $78 billion in revenue for Q1 FY27, with a tolerance of plus or minus 2%. That was significantly ahead of the $72 billion analyst consensus estimate, representing an 8.3% beat on forward guidance. The company also expects non-GAAP gross margins of approximately 75% for the quarter.
NVIDIA's full fiscal year 2026 revenue totaled $215.9 billion, up 65% from $130.5 billion in FY25. Data center revenue for the full year was $193.7 billion, up 68%. Full year net income was $120.1 billion, up 65% from $72.9 billion in FY25.
Jensen Huang said "the agentic AI inflection point has arrived" and described Grace Blackwell as "the king of inference today" for delivering an order-of-magnitude lower cost per token. He also made the pointed observation that "compute is revenues," arguing that in an AI-driven economy, access to compute is not optional but a direct input to revenue generation.
Vera Rubin is NVIDIA's next-generation rack-scale GPU platform, the successor to Grace Blackwell. NVIDIA says it is designed to deliver up to a 10x reduction in inference token cost compared to Blackwell. CFO Colette Kress confirmed on the Q4 FY26 earnings call that the first Vera Rubin samples were shipped to customers the week of February 25, 2026. Broader availability is expected later in 2026.
NVIDIA's gaming segment generated $3.7 billion in Q4 FY26 revenue, up 47% year over year. The quarter-over-quarter figure was down 13% from Q3 FY26, which reflects typical post-holiday seasonality. For the full FY26 year, gaming revenue was $16.0 billion, up 41% from $11.4 billion in FY25.
NVIDIA shares initially gained about 3% in after-hours trading following the earnings release on February 25, 2026. During the conference call, shares pulled back as much as 1.5% before settling roughly flat by end of evening. The muted reaction reflects elevated expectations going into the print. The stronger-than-expected Q1 guidance of $78 billion was widely cited as the most meaningful positive signal for investors.
The two most commonly cited risks are export controls and hyperscaler capex deceleration. U.S. government restrictions on advanced chip exports to China have already curtailed NVIDIA's addressable market in that region. Any tightening of those rules would reduce available revenue. On the demand side, if Microsoft, Google, Amazon, or Meta slow their AI infrastructure spending, NVIDIA would feel that immediately. Neither risk appears acute based on current signals, but both remain live considerations for long-term investors.
Meta leases Google TPUs in a multiyear deal worth billions. Combined with AMD and Nvidia pacts, Meta is reshaping AI chip supply dynamics.
NVIDIA launches the Rubin computing architecture at MWC Barcelona with six new chips spanning agentic AI, physical AI, autonomous vehicles, robotics, and biomedical.
NVIDIA invests $2 billion each in Lumentum and Coherent, signaling that AI infrastructure's next bottleneck is connectivity — not compute — as optical interconnects become critical for data center scaling.