NVIDIA closed its fiscal second quarter of 2027 with the largest revenue figure it has ever reported: $96.2 billion, up 106% from a year earlier and 18% from the prior quarter. Yet the number that will shape how investors read this print is not the record on the top line. It is the one management put next to it: a gross margin guided down, on purpose and in advance. This analysis walks through what NVIDIA reported for the quarter ended July 26, 2026, how it compares with the $81.6 billion record it set in the prior quarter, and why a company still accelerating chose to tell investors its margin was about to step lower.
Key Takeaways
- NVIDIA reported record revenue of $96.2 billion for fiscal Q2 2027 (+106% year over year, +18% sequentially), above a consensus near $92 billion and above its own prior $91 billion guide.
- GAAP gross margin was 75.0%, but management guided it down in a stair-step: 74.0% (±50bp) for Q3, with a 71-72% trough guided for Q4, citing memory and component cost inflation.
- GAAP diluted EPS of $2.46 came in HIGHER than non-GAAP EPS of $2.22 — an inversion caused by $7.771 billion of gains on equity securities that the non-GAAP basis excludes.
- GAAP net income of $59.688 billion exceeded operating cash flow of $24.077 billion by roughly $35.6 billion; the gap is accounted for in full by the non-cash equity gains plus a receivables and inventory build.
- Reporting is now two segments — Data Center ($89.0 billion) and Edge Computing ($7.2 billion) — with Data Center at about 92.5% of revenue.
The quarter in numbers: a $96.2B record
Revenue of $96.2 billion set a company record and beat the consensus figure near $92 billion. It also came in above NVIDIA's own prior-quarter guide of $91 billion. Sequentially, revenue rose 18% on the $81.6 billion reported three months earlier; year over year, it more than doubled, up 106%.
The earnings figures grew even faster than revenue, and here the two accounting bases diverge in an unusual way. GAAP diluted EPS was $2.46, up 128% year over year and 3% sequentially. Non-GAAP diluted EPS was $2.22, up 120% year over year and 19% sequentially. GAAP net income was $59.688 billion; non-GAAP net income was $53.954 billion.
| Metric | Q2 FY2027 | YoY | QoQ |
|---|---|---|---|
| Revenue | $96.2B | +106% | +18% |
| GAAP gross margin | 75.0% | +2.6pts | +0.1pt |
| GAAP operating income | $63.734B | +124% | +19% |
| GAAP net income | $59.688B | +126% | +2% |
| GAAP diluted EPS | $2.46 | +128% | +3% |
| Non-GAAP diluted EPS | $2.22 | +120% | +19% |
One caution on the beat: a per-share earnings-surprise percentage is not published here, because the consensus EPS figure for this specific quarter was not yet captured in the earnings record when this was written. The cleaner statement of the beat is the top line — $96.2 billion reported against a consensus near $92 billion. For context, NVIDIA has beaten consensus on non-GAAP EPS in each of the four preceding quarters.
| Quarter | EPS surprise |
|---|---|
| Q2 FY2026 | +2.13% |
| Q3 FY2026 | +1.99% |
| Q4 FY2026 | +3.62% |
| Q1 FY2027 | +4.34% |
| Q2 FY2027 | not available |
Growth bought with margin: the stair-step down
This is the defining tension of the quarter, and it is unusually clean. Revenue is accelerating and the Q3 guide of $108.0 billion (±2%) implies another 12% sequential step. At the same time, gross margin is guided lower — not in a single drop, but in a stair-step. Reported gross margin was 75.0% (both GAAP and non-GAAP) in Q2. The Q3 outlook calls for 74.0%, plus or minus 50 basis points. The stated mechanism is memory and component cost inflation, which NVIDIA is absorbing now.
Each point of margin is roughly $1.1B of gross profit
The give-up is easier to weigh once it is put in dollars. At the $108.0 billion Q3 guide, each percentage point of gross margin is worth roughly $1.08 billion of gross profit.
Sizing the margin walk-down
At the $108.0 billion Q3 revenue guide, one percentage point of gross margin is about $1.08 billion of gross profit. Walking from the 75.0% just reported toward the low-70s implied by later guidance — a move of roughly three and a half points — represents on the order of $3.8 billion of self-disclosed gross profit given up per quarter at that revenue level.
For the first time in this cycle, NVIDIA is visibly trading gross-margin rate for volume, and it is telling investors so ahead of time. There are two defensible ways to read that. One is that the company is paying up for component supply to capture share of an accelerating build-out — spending margin points to move more units into a market it does not want to under-serve. The other is that the era of frictionless pricing power, in which cost increases were simply passed through, is meeting its first real limit. Both readings fit the same facts; the print itself does not settle which is right.
On the cost pressure, CFO Colette Kress addressed gross margins directly on the earnings call. Her point, as reported in call coverage, was that component costs have risen significantly, that pricing conditions in memory are extreme, and that the magnitude of the increase has exceeded the company's prior expectations and is set to rise further into next year. Investors weighing the two readings above should treat that as management's characterisation of the pressure, not an independent fact — the substance is attributed to Kress and to the call.
The GAAP/non-GAAP inversion: when adjusted is the conservative number
Most quarters, a company's non-GAAP earnings figure sits above its GAAP figure, because the adjustments strip out costs. This quarter, NVIDIA's did the opposite: GAAP diluted EPS ($2.46) was higher than non-GAAP ($2.22). The cause is a single item — $7.771 billion of gains on equity securities, which sits inside GAAP net income and which the non-GAAP basis excludes.
Why non-GAAP excludes the $7.771B equity gain
An equity-securities gain is an increase in the value of investments NVIDIA holds. It has nothing to do with selling GPUs, so the non-GAAP basis removes it to show the operating business on its own. The consequence is worth stating plainly: here, the adjusted number is the more conservative one, because it takes out an investment gain rather than adding back a cost. Adjustments do not always flatter.
That direction is the opposite of what a reader would have seen elsewhere this earnings season. In AMD's Q2, where the non-GAAP figure came in higher, the adjustments removed recurring operating costs and lifted the reported figure above GAAP — the usual pattern. NVIDIA's inversion runs the other way for a specific, non-operating reason. Neither basis is inherently the "true" one; each includes and excludes different things, and the useful exercise is knowing which is which rather than picking a favourite.
The cash gap: where $35.6B went
GAAP net income was $59.688 billion in the quarter, while operating cash flow was $24.077 billion — a gap of $35.611 billion. That is a large difference, and it is fully accounted for by three items drawn from the cash-flow statement and balance sheet.
Working capital: receivables and inventory build
The first component is the $7.771 billion of gains on equity securities already discussed: it is non-cash, so it lifts GAAP net income without bringing cash in. The second is a $22.346 billion increase in accounts receivable — a use of cash, as revenue is billed but not yet collected. The third is a $5.784 billion increase in inventories, also a use of cash.
| Component | Amount |
|---|---|
| Non-cash gains on equity securities | $7.771B |
| Increase in accounts receivable | $22.346B |
| Increase in inventories | $5.784B |
| Sum | $35.901B |
| Net income vs operating cash flow gap | $35.611B |
Those three items sum to $35.901 billion, which accounts for the $35.611 billion gap in full; the small overshoot nets against ordinary non-cash add-backs such as depreciation and stock-based compensation, which run the other way and which exist but do not need itemising here. At quarter end, accounts receivable stood at $63.059 billion and inventories at $31.575 billion.
This is the same story as the margin walk-down, seen from the balance sheet rather than the income statement. NVIDIA is building inventory and financing customer receivables to feed an accelerating ramp; that consumes cash now against revenue expected later. Stated as mechanics, working capital absorbed the cash. Whether that build is comfortable is a judgement for the reader — nothing in the statements characterises the receivable increase as a collection issue, and reading it as one would go beyond what the figures show.
Free cash flow stepped down from Q1
Purchases of property and equipment were $2.677 billion in the quarter, so free cash flow — operating cash flow less capital expenditure — was approximately $21.4 billion. That is a marked step down from the $48.6 billion of free cash flow the company reported in Q1 FY2027, and the sequential decline is worth stating plainly alongside the working-capital explanation above. Over the first six months of fiscal 2027, operating cash flow was $74.421 billion, which implies Q1 alone generated roughly $50.3 billion — far more than Q2, precisely because the working-capital build landed in the second quarter.
NVIDIA still returned capital heavily. Approximately $26.0 billion went to shareholders in the quarter through share repurchases and cash dividends, with roughly $99.0 billion remaining under the buyback authorization at quarter end. The board declared a quarterly cash dividend of $0.25 per share, payable October 1, 2026.
Two segments now: Data Center at ~92.5% of revenue
Starting from Q1 FY2027, NVIDIA reports two segments rather than the older breakout. Data Center is one; Edge Computing — which combines what were previously reported as Gaming, Professional Visualization, Automotive, and OEM & Other — is the other. Separate Gaming, ProViz, and Automotive figures are no longer reported for this quarter.
Data Center vs Edge Computing
Data Center revenue was $89.0 billion, up 117% year over year and 18% sequentially. Edge Computing revenue was $7.2 billion, up 27% year over year and 13% sequentially. The two sum to the $96.2 billion total.
| Segment | Q2 FY2027 | YoY | Share of total |
|---|---|---|---|
| Data Center | $89.0B | +117% | ~92.5% |
| Edge Computing | $7.2B | +27% | ~7.5% |
| Total | $96.2B | +106% | 100% |
The concentration is the point to sit with: Data Center is about 92.5% of revenue, extending the concentration theme that ran through the prior-quarter analysis. A business this dependent on one segment rises and falls with that segment's demand, and that is a structural feature of the results rather than a one-quarter quirk.
The Q3 outlook and the China assumption
NVIDIA guided Q3 FY2027 revenue to $108.0 billion, plus or minus 2%, which implies roughly 12% sequential growth on the $96.2 billion just reported. It guided gross margin to 74.0% (±50bp), GAAP operating expenses to about $9.2 billion (non-GAAP about $9.0 billion), and a full-year tax rate of 16.0% to 18.0%.
One assumption inside that guide deserves its own line. NVIDIA's outlook explicitly assumes no Data Center compute revenue from China. That is continuity with the prior-quarter framing: because nothing from China is in the model, there is nothing in the model to lose from it. Any change in access would be an addition to, not a subtraction from, the guided figures. The assumption is NVIDIA's own, stated in its outlook.
What management expects for fiscal 2028
Several of the most-quoted numbers from this print are forward-looking statements made by management, and they should be read as expectation or guidance rather than as fact. On the call, CFO Colette Kress said NVIDIA expects revenue to grow approximately 70% in fiscal 2028 and characterised that outlook as supply-constrained. She guided gross margin to a trough of 71-72% in Q4 FY2027 before settling at 72-73% in fiscal 2028, as executed price increases take effect in the first quarter of fiscal 2028. CEO Jensen Huang said demand is higher than that roughly 70% growth outlook, that supply is broadly constrained, and that the company is working with suppliers to expand capacity. Management also characterised Vera Rubin — expected to be about 20% of Q3 Data Center revenue — as its fastest product ramp to date.
Items drawn from the call are reported commentary rather than lines in the press release, and each is attributed to the named executive above. As with any forward-looking statement, these are expectations that may not be realised; past performance and prior guidance do not guarantee future results.
Guidance is management's, not a forecast
On the tone of the quarter, Huang framed the moment in the release itself:
"AI has reached its inflection point. It's doing useful work. Its tokens are productive and profitable. Now, compute is revenue.
— Jensen Huang (NVIDIA Q2 FY2027 press release)
That is the company's characterisation of demand, not an independent assessment. It belongs in the record as management's view.
How the market read it — after hours only
A note on the reaction, framed carefully. NVIDIA reported after the close on Wednesday, August 26, 2026, and this analysis was written the following day, before the first full post-earnings trading session had ended. The report-day close was $209.66, down 1.59% on the day — but that close came before the results were released, so it is not a reaction to them. There is no verified post-earnings closing price yet.
The only defensible reference to the immediate move is an after-hours one: Yahoo Finance reported NVDA at $225.30, up 7.46%, in after-hours trading following the release. That is an after-hours figure, not a close, and it can change by the time the session ends. Going into the print, the stock was well off its recent high — down 6.94% from the $225.30 local peak of August 13 to the $209.66 report-day close, though still up 10.34% from the July 29 trough of $190.01. NVIDIA has declined after each of its four preceding earnings reports; that is history, and whether the pattern held this time is not something the record can yet show.
Frequently asked questions
How much revenue did NVIDIA report in Q2 FY2027? Revenue was a record $96.2 billion, up 106% year over year and 18% sequentially, above a consensus near $92 billion and above NVIDIA's own prior $91 billion guide.
Why is NVIDIA's gross margin going down to 74%? Gross margin was 75.0% in Q2 and is guided to 74.0% (±50bp) in Q3, with a 71-72% trough guided for Q4. Management cited memory and component cost inflation that NVIDIA is absorbing now, with executed price increases taking effect in the first quarter of fiscal 2028.
Why is NVIDIA's GAAP EPS higher than its non-GAAP EPS? GAAP diluted EPS ($2.46) exceeded non-GAAP ($2.22) because the non-GAAP basis excludes $7.771 billion of gains on equity securities — a non-operating item. This quarter, the adjusted number is the more conservative one.
What explains the gap between NVIDIA's net income and operating cash flow? GAAP net income of $59.688 billion exceeded operating cash flow of $24.077 billion by $35.611 billion. That gap is accounted for by $7.771 billion of non-cash equity gains, a $22.346 billion increase in accounts receivable, and a $5.784 billion increase in inventories.
What are NVIDIA's two reporting segments now? Data Center ($89.0 billion) and Edge Computing ($7.2 billion). Edge Computing combines the former Gaming, Professional Visualization, Automotive, and OEM & Other lines. Data Center is about 92.5% of revenue.
Does NVIDIA's Q3 outlook include any China Data Center revenue? No. NVIDIA's outlook explicitly assumes no Data Center compute revenue from China, so nothing from China is built into the guided figures.