Inside Nvidia's Q2
Margins slipped, and the stock rose anyway

📄 The numbers — nearly every line beat

Nvidia reported fiscal 2027 second-quarter results (quarter ended July 26, 2026) after the close on August 26, 2026.

On the numbers alone there is little to fault. Revenue cleared the estimate by more than $4 billion and more than doubled year over year. A single quarter at $96.2 billion exceeds what most large corporations book in a full year.

Guidance was strong too.

That last line is the most aggressive item in the release. The company itself guided next fiscal year's growth to 70%, more than $100 billion above Wall Street's model. Announcements like that usually send a stock sharply higher.

This one didn't.

📉 Four quarters of falling — and why the pattern broke

Nvidia has beaten estimates five quarters running. Here is how the stock responded each time.

Four consecutive quarters of "beat, then drop." And this time the stock arrived at the print already down seven straight sessions — its longest losing streak since 2022. Expectations going in were explicit: major outlets ran headlines predicting a fifth consecutive decline.

Instead it rose. Shares jumped 7.63% in overnight trading right after the conference call, added 6% in Thursday's premarket, and held gains near 4% in the regular session on August 27 — ending the losing streak.

That said, the rally did not erase every concern. One line in this print missed, and it wasn't revenue.

🧊 The real news was margin, not revenue

Gross margin came in at 75.0%, exactly in line with guidance. The issue is the outlook.

Q3 gross margin guidance: 74.0% ±50bp. A one-point step down.

One point sounds small, but against $108 billion of quarterly revenue it means roughly $1 billion of profit disappears. More important is the direction. Nvidia's margin has never been a line investors had to worry about. A 75% gross margin is abnormally high for semiconductors, and was itself evidence of the company's near-monopoly position.

Now that number falls for the first time. The company named the cause directly: rising memory costs.

So the new information in this print was never revenue — everyone knew revenue would be large. The new information was a signal that the quality of the growth had begun, marginally, to deteriorate.

This time the market chose not to lead with that. Not because the bear case was wrong, but because it ranked lower. A one-point margin step-down is a settled fact; FY2028 guidance is still a promise — and the market weighted the promise more heavily. Whether that was the right call gets answered by whether margin holds 74% next quarter, which is why margin heads the checklist in §8.

🔄 The cost squeezing that margin lifted Micron 688%

Here is the most interesting angle in this print. Viewed from the other side, the memory cost pressuring Nvidia's margin looks like something else entirely.

Same AI cycle, same twelve months. Yet Nvidia captured barely half the gain of its own tech sector (+39.7%) and only narrowly cleared the S&P 500 (+18.7%). Micron — which sells memory to Nvidia — rose 34 times as much.

The multiple is the more telling part. Even after a 688% move, Micron's P/E of 21.2× sits below Nvidia's 32.1×. That means actual earnings grew as fast as the share price, or faster — results drove the stock, not multiple expansion. The mechanics of this memory supercycle are covered in the Micron earnings deep dive.

That said, it is too early to call the trend one-directional. On August 24, Chinese memory maker CXMT surged over 460% on its Shanghai debut, reaching a $540 billion market cap — and memory stocks sold off hard (Micron -5.83%). CXMT holds just 8% DRAM share, fourth globally, but the $8.6 billion it raised is earmarked for capacity, raising fears of downward pressure on DRAM prices. Nvidia fell 2.91% the same day.

In other words, the memory price is both a cost burden for Nvidia and a thermometer for AI demand. Read it in only one direction and you get it wrong.

🔮 The FY2028 +70% guide — two sides of a supply constraint

The strongest card in this release was the preliminary FY2028 revenue growth outlook of roughly 70% — $690–700 billion, more than $100 billion above the $570 billion analysts had modeled.

But the company attached a condition: demand is "much higher" than what it can supply, and the outlook assumes a supply constraint. That sentence reads two ways.

One more condition is easy to miss in the Q3 guide. The $108 billion figure excludes China data center compute revenue entirely. China risk is therefore absent from guidance — which cuts both ways: a downside removed, but also an upside option not reflected. The regulatory backdrop is covered in the US chip war.

🎯 What a $306 target does and doesn't tell you

How does Wall Street see it? Data as of August 26, 2026.

On the surface this is overwhelmingly positive. Apply the tests from how to read an analyst report, though, and the reading shifts.

• +45.8% upside is three times the market average. Across the 525 stocks marketbrief tracks, average implied upside is +15.0%. Nvidia's +45.8% may express conviction — or it may simply be targets that haven't been marked down as fast as the price fell. Price targets tend to trail the stock.
• The absence of Sell ratings is not information. Among those same 525 stocks, zero carry a Sell consensus. That is a structural property of the metric, not a feature of Nvidia.
• An 8% spread, however, is real signal. The five most recent targets cluster tightly at $325–$352, against a market-wide median spread of 20.5%. Unusually settled — which means the narrative is stable, and also that there is room for these to move together if that narrative breaks.

The factor scores say the same thing. Profitability grades A (top 5.5%) and growth grades A (top 13.9%), both elite — while momentum is B-, at the 48.7th percentile: dead average. Best-in-class results, ordinary price action. That gap is Nvidia right now.

🧭 What to check next — monitoring instead of forecasting

What matters next quarter? Instead of a price forecast, here are verifiable indicators.

※ Written August 27, 2026. Earnings figures are based on Nvidia's official FY2027 Q2 release (quarter ended 2026-07-26) published August 26, 2026, and major news coverage; price, valuation, and analyst data are as of the 2026-08-26 close. The August 27 move (+7.63% after hours, around $218.77 intraday) reflects reporting as of roughly 10:00 a.m. ET on August 27, is an intraday figure rather than a closing price, and may change by the close. Prior-quarter reactions in the table are day-of approximations. FY2028 guidance is a preliminary company outlook and is not guaranteed. This content is informational and is not a recommendation to buy or sell any security. Single-stock investing carries risk of loss of principal; investment decisions and their consequences rest with the investor.