Categories Technology

NVIDIA Beats Big: $115.8B Revenue, $130B Q4 Guide

NVIDIA reported its third-quarter fiscal 2027 results after the market closed on Tuesday, November 17, 2026, and the headline numbers were hard to miss: revenue of $115.8 billion, up 103% from a year earlier, and a fourth-quarter revenue guide of $130 billion. The company beat Wall Street’s revenue expectations by roughly $7 billion, extended its streak of earnings beats to 13 straight quarters, and still managed to give investors something to worry about in the fine print.

That tension is the real story here. This was not a clean, everything-is-perfect quarter. It was a quarter where demand for AI computing kept climbing faster than anyone expected, where the core business got even more concentrated, and where the guidance for next quarter looked stronger on the surface than it does once you adjust for the calendar and the margin outlook. If you own the stock, run a business that buys AI tools, or just want to understand where the AI buildout stands right now, you need both halves of that story.

This article is not financial or investment advice. It is a plain-English breakdown of what NVIDIA reported, what the numbers actually mean once you work through them, and why they matter beyond the stock price.

The headline numbers: a $7 billion beat

NVIDIA’s fiscal third quarter ended October 25, 2026. Revenue came in at $115.8 billion, up 103% year over year and up 20% from the prior quarter. That beat the Bloomberg consensus of about $109.1 billion by roughly $6.7 billion, and it beat NVIDIA’s own August guidance of $108 billion, plus or minus 2%, by a comfortable margin.

Non-GAAP earnings per share were $2.68, against expectations of about $2.47. GAAP earnings per share were $2.70. GAAP net income reached $65.3 billion, and operating income was $77.5 billion, up 115% from a year earlier. Free cash flow was $43.8 billion in the quarter, about double the prior quarter’s level.

Gross margin was 74.8% on a GAAP basis and 74.9% on a non-GAAP basis. That is a very high margin by any standard in hardware, and it is one reason the market pays so much attention to what happens to it next. More on that in a moment, because the margin outlook for the fourth quarter is where the report gets interesting.

The stock’s first reaction was positive but not wild. Shares closed November 16 at $156.14, then closed November 17 at $155.46, down 0.44% in the regular session. After the release hit, the stock rose about 2% in after-hours trading, with some coverage putting the extended-trading move closer to 3.5%. Either way, the move was a modest thumbs-up, not a celebration. That muted reaction tells you something important: the market had already priced in a very strong quarter, and it was reading the guidance just as closely as the headline beat.

The beat, laid out in one table

A beat only makes sense when you can see what was expected, what was guided, and what actually happened side by side. Here is that comparison for the quarter just reported, plus the fourth-quarter outlook.

Measure NVIDIA’s August guide Wall Street consensus Actual / new guide
Q3 revenue $108 billion (±2%) ~$109.1 billion $115.8 billion
Q3 non-GAAP EPS Not guided ~$2.47 $2.68
Q3 gross margin (non-GAAP) Not guided as a single point Not the focus 74.9%
Q4 revenue — ~$125.25 billion (LSEG) $130 billion (±3%)
Q4 gross margin — Not the focus 71–72% (GAAP and non-GAAP)

Read that table from top to bottom and the pattern is clear. NVIDIA beat on revenue by a wide margin and beat on earnings per share by about 21 cents. Then it guided fourth-quarter revenue above consensus — $130 billion against about $125.25 billion expected — while guiding gross margin down to 71–72%, a clear step down from the 74.9% non-GAAP margin it just posted.

That combination is unusual. Normally a company that beats revenue by $7 billion and guides the next quarter above consensus would see its margin outlook hold steady or improve. NVIDIA did the opposite on margin, and that is why the after-hours reaction stayed measured instead of turning into a surge.

Data Center is now the whole story

Data Center revenue was $107.6 billion in the quarter, up 110% from a year earlier. That is about 93% of NVIDIA’s total sales. Put differently, for every $100 NVIDIA took in during the quarter, about $93 came from selling AI computing hardware and related products to data centers.

That concentration matters for two reasons. First, it explains why NVIDIA’s results have become a proxy for the entire AI buildout. When the big cloud providers and AI labs spend more on computing, NVIDIA’s revenue follows almost immediately. When they pause, there is very little elsewhere in the business to cushion the fall. Gaming, professional visualization, and automotive still exist, but at this scale they are rounding errors next to Data Center.

Second, the customer concentration inside Data Center is striking: three direct customers represented 36% of NVIDIA’s sales in the quarter. That is how the hyperscale market works, with a handful of cloud giants buying at enormous scale — but it does mean NVIDIA’s fortunes are tied closely to the spending plans of a few customers. If one of them slows its buildout, the effect would show up fast.

Interim CEO Colette Kress, who also serves as chief financial officer, pointed to roughly 75% year-over-year growth into fiscal 2028, a step up from the roughly 70% framing the company had used before. She also said demand is still accelerating as supply catches up. That is a strong statement, and it fits the numbers: revenue more than doubled year over year, and the quarter still came in well above the company’s own guide. The constraint, in NVIDIA’s telling, has been supply, not demand. The Vera Rubin NVL72 computing platform has now entered production shipments, which is the next step in that supply story — a newer platform moving from announcement to actual deliveries.

Why the $130 billion guide is softer than it looks

A $130 billion revenue guide sounds enormous, and it is. It is also about $4.75 billion above what Wall Street expected. But two adjustments change how strong it really is, and both are worth working through yourself, because they are the kind of detail that gets lost in headlines.

The 14-week quarter

The fourth quarter of fiscal 2027 is a 14-week quarter, compared with 13 weeks in the third quarter. That extra week matters a lot when you are comparing the two numbers.

Here is the math. Third-quarter revenue of $115.8 billion over 13 weeks works out to about $8.91 billion per week. Fourth-quarter guidance of $130 billion over 14 weeks works out to about $9.29 billion per week. That is a weekly run-rate increase of roughly 4%, not the 12% increase you get if you compare $130 billion with $115.8 billion directly.

That does not make the guide weak. A 4% weekly increase on a base this large is still about $5.3 billion more revenue than 14 weeks at the old weekly pace would produce. But it does change the picture: the headline jump from $115.8 billion to $130 billion is mostly the calendar doing the work.

If you only remember one number from this section, remember the 4%. That is the honest growth rate inside the guide.

The margin step-down

The second adjustment is the margin guide. NVIDIA guided fourth-quarter gross margin to 71–72% on both a GAAP and non-GAAP basis, down from 74.9% non-GAAP in the quarter just reported. That is a drop of roughly three percentage points, which on $130 billion of revenue is worth roughly $3.9 billion of gross profit.

Why would margin fall while revenue rises? The company did not pin the step-down to a single cause in the headline figures, but the most widely discussed explanation is cost pressure in memory and other components. High-bandwidth memory and related parts have become more expensive as demand for AI hardware has surged, and those costs land directly in the cost of goods sold. A platform transition can also weigh on margins in the short term, as newer products ramp and older ones mix through.

Whatever the precise mix, the direction is what investors focused on. A company that can charge what NVIDIA charges and still sees its margin guide fall by three points is telling you that its input costs are rising faster than its prices, at least for now. That does not erase a $115.8 billion quarter. It does mean the next quarter’s profit growth will be slower than its revenue growth, and it puts a question mark next to the margin trajectory that was not there before.

The buyback and dividend box

NVIDIA also used the report to make a clear statement about capital returns. The board approved a $150 billion increase to the share buyback authorization. The company pays a $0.25 quarterly dividend. In the third quarter alone, it returned $28.1 billion to shareholders.

Those numbers deserve a pause, because they show how much cash this business is generating. Free cash flow of $43.8 billion in a single quarter is enough to fund the dividend, fund a very large buyback program, and still leave room for the research and development spending that keeps the product pipeline moving. The $150 billion authorization increase is not money spent — it is permission to spend, over time, as the board chooses — but the size of it signals confidence that the cash generation will continue.

For a regular reader, the practical takeaway is simpler: buybacks and dividends are how a company with more cash than it can profitably reinvest returns money to shareholders. They do not make the business better or worse on their own.

The quarter vs. the transition: keep them separate

This was the first earnings report under interim CEO Colette Kress, who took over after Jensen Huang resigned on November 12, 2026. That leadership change is a major story in its own right, and you can read the site’s own coverage of it here: Jensen Huang resigns as NVIDIA CEO November 12, 2026.

But there is a timing detail that matters for how you read these results. The quarter ended October 25, 2026, which was before Huang’s resignation. The numbers in this report — the $115.8 billion, the $107.6 billion in Data Center, the $2.68 in non-GAAP earnings per share — describe a business operating under Huang’s leadership through the end of October. Only the guidance and the earnings call describe the transition period under Kress.

That distinction keeps you honest in both directions. You should not credit the new interim leadership with a quarter that ended before the change, and you should not blame it for the margin guide either, at least not yet. The guidance is the first forward-looking statement under Kress, and it is fair to judge her on it. The quarter itself belongs to the business as it was run through October 25.

Board chair Harvey Jones said the board does not intend to reinstate Huang, and he declined to answer questions about the ongoing internal investigation. That brings us to the part of this story where care with language matters most.

The honesty box: what is reported vs. what is known about Huang’s exit

On November 8, NVIDIA’s board opened an independent investigation into reports that Huang was targeted by a foreign intelligence effort. That is the correct way to describe it: reports, and an investigation into those reports. It is not a finding. It is not a conclusion. No outcome has been announced, and the board has declined to discuss it on the earnings call.

What is confirmed is the leadership change itself. Huang resigned on November 12, 2026. Kress, who was already chief financial officer, is serving as interim CEO. Jones has said the board does not intend to reinstate Huang. Those are facts. Everything else about why the resignation happened, beyond what the board has said publicly, is still in the category of reported allegations and unanswered questions.

China: $2.4 billion in, zero assumed going forward

China is the other number in this report that needs careful reading. Data Center Hopper shipments to China were $2.4 billion in the quarter just reported. That is real revenue, already booked, from a product line that has been caught up in export controls and policy debates for years.

For the fourth quarter, NVIDIA’s guidance assumes no Data Center compute revenue from China at all. Zero. That is a conservative assumption, and it is worth understanding why a company would guide to zero on a line that just produced $2.4 billion.

The reason is policy uncertainty. Export rules and the political environment around advanced chips sold into China can change quickly, and a company guiding a quarter does not want to promise revenue that a rule change could take away. Guiding to zero removes that risk — any China Data Center compute revenue that does come through would be upside.

What this means if you buy AI tools, not NVIDIA stock

Most readers of this site do not own a meaningful amount of NVIDIA stock. They run small businesses, freelance, create content, or build side projects. So why should a $115.8 billion earnings report matter to them?

It matters because NVIDIA’s numbers are a leading indicator for the price and availability of the AI tools you use every day. Here is how that chain works, in plain terms.

First, GPU supply. When NVIDIA says demand is still accelerating as supply catches up, that is another way of saying the hardware that powers AI services is still scarce relative to demand. Scarcity in hardware does not immediately raise the price of your monthly AI subscription, but it does shape how fast providers can expand, how quickly new models can be trained, and how much capacity is available for new features. If supply catches up, as Kress suggested it is starting to, you should expect more capacity and more competition among AI providers over time. If it does not, expect the current leaders to stay ahead on speed and availability.

Second, AI tool pricing. The cost of running an AI service is tied closely to the cost of the computing behind it. When NVIDIA’s margins come under pressure from memory costs, that is a signal that the underlying cost of AI computing is rising, not falling, at least in the short term. Providers can absorb that cost for a while, pass it on through higher prices, or offset it with efficiency gains. None of those outcomes is guaranteed. What you can do is watch for price changes in the tools you rely on, and avoid building a business that depends on a single AI tool staying cheap forever. NVIDIA is also pushing cheaper open models of its own — see our look at NVIDIA’s Nemotron Transcribe open model — which is one more reason tool prices may not stay high forever.

Third, data-center demand. A $107.6 billion Data Center quarter, up 110% year over year, means the big cloud providers and AI labs are still spending heavily on infrastructure. That spending is what funds the next generation of models, the next round of features, and the next wave of tools aimed at small businesses and creators. It is also what makes the AI market feel crowded and fast-moving. If you are choosing tools right now, the fact that the infrastructure spend is still growing tells you the pace of new releases is unlikely to slow down soon.

A practical checklist for freelancers and small businesses

You do not need to trade the stock to act on this report. You can use it to make better decisions about the tools you already pay for.

Start by listing the AI tools you pay for monthly, and what each one actually does for your revenue. If a tool is nice to have but does not directly help you earn, deliver, or save time, it is a candidate to cut if prices rise. Next, check whether you have a backup for your most important AI tool. If your main writing, image, or video tool went up 20% in price tomorrow, would you have an alternative ready, or would you be stuck?

What to watch next

Three things will tell you whether this quarter was a high-water mark or a step on the way to something bigger.

The first is the margin. If the fourth-quarter margin lands at 71–72% as guided and then stabilizes or recovers, the step-down will look like a temporary cost squeeze. If it keeps falling in the quarter after that, the story changes. A business that can grow revenue 100% year over year but cannot hold its margin is a different business from one that can do both.

The second is the weekly run-rate. The fourth quarter is 14 weeks, so the headline comparison will flatter the growth rate. Watch the weekly number instead. If the first quarter of fiscal 2028, a normal 13-week quarter, holds a weekly run-rate near $9.3 billion, the business has held its new level. If it falls back toward $8.9 billion, the fourth-quarter jump was mostly the calendar.

The third is the CEO search. Kress is interim CEO, and the board will need to name a permanent successor. That choice will shape NVIDIA’s strategy, its relationships with its largest customers, and its handling of the China question.

The bottom line

NVIDIA’s third quarter was a very strong quarter with a cautious guide. Revenue more than doubled, earnings beat expectations by a clear margin, and the company returned $28.1 billion to shareholders while authorizing another $150 billion in buybacks. At the same time, the fourth-quarter guide grows only about 4% on a weekly basis once you adjust for the 14-week calendar, and the margin guide steps down to 71–72%.

That is not a contradiction. It is what a business looks like when demand is still enormous, supply is still catching up, and input costs are rising. The market’s muted after-hours reaction — up about 2%, not a surge — reflected that balance.

For the full picture, NVIDIA’s investor relations release is the primary source for the figures in this article, and you can also read the reaction coverage from Motley Fool and Market Unclouded for how the market read the numbers.

Frequently Asked Questions

How much revenue did NVIDIA report for Q3 fiscal 2027?

NVIDIA reported revenue of $115.8 billion for the third quarter of fiscal 2027, which ended October 25, 2026. That was up 103% from a year earlier and up 20% from the prior quarter. It beat the Bloomberg consensus of about $109.1 billion by roughly $6.7 billion.

What is NVIDIA’s Q4 revenue guidance?

NVIDIA guided fourth-quarter fiscal 2027 revenue to $130 billion, plus or minus 3%. That is above the LSEG consensus of about $125.25 billion. The fourth quarter is a 14-week quarter, so the weekly run-rate increase over the third quarter is only about 4%, not the 12% the headline comparison suggests.

Why did NVIDIA guide gross margin down to 71–72%?

NVIDIA guided fourth-quarter gross margin to 71–72% on both a GAAP and non-GAAP basis, down from 74.9% non-GAAP in the third quarter. The company did not attribute the step-down to a single cause in the headline figures, but rising memory and component costs are the most widely discussed explanation. On $130 billion of revenue, a three-point margin drop is worth roughly $3.9 billion of gross profit.

Who is NVIDIA’s CEO right now?

Colette Kress is serving as interim CEO, in addition to her role as chief financial officer, after Jensen Huang resigned on November 12, 2026. Board chair Harvey Jones has said the board does not intend to reinstate Huang. The third-quarter results cover a period that ended October 25, before the resignation, so the numbers describe the business as run through that date.

How much did NVIDIA return to shareholders in Q3?

NVIDIA returned $28.1 billion to shareholders in the third quarter. The board also approved a $150 billion increase to the share buyback authorization, and the company pays a $0.25 quarterly dividend. Free cash flow was $43.8 billion in the quarter, about double the prior quarter.

Does NVIDIA’s Q4 guidance include revenue from China?

No. The fourth-quarter guidance assumes no Data Center compute revenue from China. In the quarter just reported, Data Center Hopper shipments to China were $2.4 billion. Guiding to zero removes policy risk from the forecast — any China Data Center compute revenue that does come through would be upside to the guide.

Leave a Reply

Your email address will not be published. Required fields are marked *