When Does Nvidia Report Earnings? Wednesday, August 26, After the Close. The Bar Is $91 Billion

TL;DR

• Nvidia reports fiscal Q2 2027 on Wednesday, August 26, 2026, after the close. The quarter ended July 26, 2026.
• Results land at roughly 1:20pm PT / 4:20pm ET. The conference call is at 2:00pm PT / 5:00pm ET (10:00pm BST, 5:00am Singapore on Thursday).
• The bar is management's own: revenue of $91 billion, plus or minus 2%, with gross margin guided to 74.9% GAAP and 75% non-GAAP.
• Context for that number: Q1 delivered a record $81.62 billion, up 85% year over year and 20% sequentially, with data centre revenue of $75.2 billion, which is more than 92% of everything Nvidia sells.
• Consensus already sits above the guide, at roughly $93.5 billion and $2.13 of EPS. Clearing $91 billion is not the test.
• The best preview available lands three weeks earlier. AMD reports on August 4, and it is the only other company that sells AI compute rather than buying it.
• The detail most previews will miss: that $91 billion guide assumes no data centre compute revenue from China at all.

When Does Nvidia Report Earnings?

After the close on Wednesday, August 26, 2026. Results are expected at about 1:20pm PT / 4:20pm ET, with the conference call 40 minutes later at 2:00pm PT / 5:00pm ET.

If you are watching from outside the US, that call is 10:00pm BST in the UK and 5:00am on Thursday in Singapore.

The gap between the release and the call matters more for Nvidia than for most companies. The press release carries the revenue and the guide, but the segment detail and the China commentary come on the call, and Nvidia has a history of the second half of that hour mattering more than the first.

What Nvidia Has Guided To

Line	Fiscal Q2 2027 guidance
Revenue	$91 billion, plus or minus 2%
GAAP gross margin	74.9%
Non-GAAP gross margin	75.0%
China data centre compute	Assumed to be zero

And the quarter it is being measured against:

Line	Fiscal Q1 2027 actual
Revenue	$81.62 billion, up 85% y/y and 20% sequentially
Data centre	$75.2 billion, more than 92% of total revenue
Hyperscale	$38 billion
AI, cloud, industrial and enterprise	$37 billion
Edge computing	$6.4 billion

$91 billion against $81.62 billion is another 11% sequential step, on a base that already grew 85% year over year. That is the scale of what is being asked.

And note where the street has parked itself. Consensus is around $93.5 billion of revenue and $2.13 of EPS, above the top of management's own guided band. Analysts normally hide inside a guided range; here they are betting on a beat before it happens, which is the same unusual setup Palantir carries into Monday evening. It means clearing $91 billion is the floor, not the test.

One number to keep in proportion: data centre is more than 92% of Nvidia's revenue. That is a share, not a growth rate. This is not a diversified semiconductor company with a strong data centre division. It is a data centre company with some other lines attached.

The Board

The bar is management's own $91 billion, and it assumes nothing from China.

The China Assumption Is the Free Option

This is the part worth understanding before the print, because it changes how you read a beat.

Nvidia's $91 billion guide assumes it books no data centre compute revenue from China. Not reduced. None.

That does two things at once. It removes a large and genuinely unpredictable variable from management's forecast, which is why the guide can be as precise as plus or minus 2% on a number this size. And it means any China revenue at all arrives as pure upside against a bar that did not include it.

So the question to ask on August 26 is not "did Nvidia beat $91 billion", because a company guiding this precisely usually clears its own number. It is where the beat came from. A beat built on hyperscale demand is a different signal from a beat built on a China licence change, and only one of them tells you anything about the durability of the underlying demand.

Why the Segment Split Now Matters

Nvidia recently moved to a reporting framework that separates Data Center into Hyperscale and AI, cloud, industrial and enterprise, with Edge Computing alongside.

That split is more useful than it sounds. Hyperscale is the handful of enormous customers building frontier clusters, and it is the line most exposed to concentration risk: a small number of buyers, each capable of pausing. The enterprise line is broader, stickier and slower, and it is the one that would tell you AI spending is spreading beyond the companies that report their capex on a quarterly call.

At $38 billion against $37 billion in Q1, those two were almost exactly balanced. Which one carries the growth this quarter is the most informative number in the release, and it is not the headline.

What Actually Decides the Reaction

Three things, and the revenue line is not among them.

1. The next guide, not this quarter. This has been the season's consistent lesson. Apple set a record gross margin and fell about 8% because the September guide came in light. Roblox beat on EPS and lost 29% on guidance alone. The October quarter guide is the number that moves Nvidia.

2. Gross margin at 75%. Nvidia is guiding to a margin most hardware companies could not reach in their best quarter. It is also buying memory into the cost squeeze that has hit everyone downstream. Holding 75% while component costs rise would be a genuine achievement. Guiding it lower would be the story.

3. Whether the demand is still contracted. Microsoft answered the capex question with $678 billion of backlog and was rewarded for it. Nvidia sells to those same buyers. Any disclosure that firms up forward commitments rather than forward hopes does more for the stock than the revenue line will.

One Precedent Worth Remembering

Q1 was a strong report and the stock fell. Revenue up 85%, data centre up 92%, and shares slid anyway.

That is the pattern to plan around. When a company is priced for near-perfection, the print is not an opportunity, it is a hurdle. Nvidia has cleared its own guidance repeatedly and been sold on the day more than once, because the debate has moved on from whether it can grow to whether the customers can keep funding the growth.

The scenario work on all of this, the consensus spread, the margin-versus-memory-costs question and the logged plays, is in the full Nvidia Q2 FY2027 earnings preview.

The Options Angle

• Implied volatility into an Nvidia print is always expensive, and it is at its worst in the final week. Buying calls or puts the day before means paying peak premium into a guaranteed volatility crush.
• Nvidia is a whole-market event, not a single-stock event. A large move drags the Nasdaq with it, so index positions are Nvidia positions on August 26 whether you intended that or not.
• For holders, a covered call into the pre-earnings volatility bid is the cleanest way to get paid for a print you are going to sit through anyway.
• Three and a half weeks is a long time to hold a view. The AMD report on August 4 and the July jobs report on August 7 both land first, and either can reprice the whole complex before Nvidia says a word.

The One-Line Read

Nvidia reports fiscal Q2 2027 after the close on Wednesday August 26, with results near 4:20pm ET and the call at 5:00pm ET, against its own guidance of $91 billion and a 75% gross margin: the guide assumes zero China data centre compute revenue, which makes any China contribution free upside and makes the composition of the beat more informative than the beat, and as Apple and Roblox both demonstrated this week, the number that decides the reaction will be the October guide rather than anything about the quarter that just ended.