Nvidia reported $96.2 billion of revenue for its second quarter of fiscal 2027 on August 26, 2026, including $89.0 billion from Data Center. The same release set formal Q3 revenue guidance at $108.0 billion, plus or minus 2%, and formal Q3 GAAP and non-GAAP gross-margin guidance at 74.0%, plus or minus 50 basis points.
The earnings call then added a different disclosure state: management’s preliminary expectation for fiscal 2028 revenue to grow about 70% year over year, alongside demand commentary, a supply bottleneck and a longer-range margin path. For FP&A, the control issue is not whether those statements matter. It is whether each one enters actuals, formal guidance, a management-expectation line or a scenario watchlist without being relabelled as something more certain.
What changed and what it means
Collapsing formal Q3 guidance, preliminary FY28 expectations and demand/supply commentary into one outlook can overstate forecast certainty and obscure variance ownership.
- Decision affected
- Decide which Nvidia Q2 and FY28 disclosures enter actuals, formal guidance, management expectations or scenario/watchlist states before rebasing the forecast.
- Evidence in brief
- Nvidia’s August 26 release establishes Q2 actuals and formal Q3 revenue and margin outlook; the earnings call separately states preliminary FY28 growth, supply and margin expectations.
- What remains unresolved
- Nvidia has not provided a formal quarterly bridge from the 70% FY28 growth expectation to Q3 guidance or quantified how additional supply would change that long-range revenue path.
- Next verification
- Preserve the disclosure-state labels now, then reconcile Q3 actuals and any updated FY28 or margin commentary against the August 26 baseline.
Key takeaways
- Nvidia’s Q2 FY27 revenue of $96.2 billion is a closed-period actual; the earlier $91.0 billion, plus or minus 2%, Q2 outlook is now a variance benchmark.
- The $108.0 billion, plus or minus 2%, Q3 revenue outlook and 74.0%, plus or minus 50 basis points, Q3 gross-margin outlook are formal near-term guidance.
- The approximately 70% FY28 revenue-growth figure is a preliminary management expectation, not the same evidence state as the formal Q3 range.
- Customer-demand commentary, supply limits, memory pricing and the Q4/FY28 margin path belong in driver assumptions and scenarios unless Nvidia formally changes the guidance state.
Q2 actuals closed the old $91bn guidance state
Nvidia’s May 20 Q1 results release guided Q2 revenue to $91.0 billion, plus or minus 2%. That created a formal range of $89.18 billion to $92.82 billion. The August result of $96.2 billion was therefore about $5.2 billion above the prior midpoint and about $3.4 billion above the top of the issued range.
That prior range should now stay in the variance file, not in the current forecast. The $96.2 billion result belongs in actuals, while the $91.0 billion midpoint and its range remain the external benchmark against which the completed quarter can be reconciled. Data Center’s $89.0 billion is also an actual. Nvidia did not give a comparable formal Q2 Data Center revenue range in the May outlook, so FP&A should not manufacture a segment “beat versus guidance” where no segment guide existed.
The $108bn Q3 range is formal guidance
The Q2 earnings release labels the next-quarter figures as Nvidia’s Q3 fiscal 2027 “Outlook.” Revenue is expected to be $108.0 billion, plus or minus 2%, which gives a formal range of $105.84 billion to $110.16 billion. GAAP and non-GAAP gross margins are expected to be 74.0%, plus or minus 50 basis points.
Those are the cleanest external planning anchors in the disclosure. An internal model can use the midpoint, the full range or its own independent forecast, but the source label should remain attached. If FP&A replaces a range with one internal point, it should preserve the issue date, range bounds, comparison basis and owner so a later actual-to-forecast review can tell whether the variance is against management guidance or an internal planning assumption. That discipline fits the actuals-to-forecast handoff in a governed budgeting and forecasting cycle.
Keep the 70% FY28 expectation in a separate state
On the August 26 earnings call, CFO Colette Kress described Nvidia’s preliminary expectation for fiscal 2028 revenue to grow approximately 70% year over year. She immediately paired that with a supply constraint, saying supply is expected to remain a bottleneck at least through the end of fiscal 2028.
That combination matters for model design. The 70% figure can be recorded as a named management expectation, with the fiscal-year period and supply constraint attached. It should not be copied into the same field as the formal Q3 revenue range. A long-range plan may choose to use it in a base case, an upside case or a monitored scenario, but that is an internal FP&A decision. The public disclosure itself does not turn the 70% expectation into a quarterly guide or remove the supply dependency.
Demand is a driver signal, not another revenue guide
Kress also said customer forecasts point to Nvidia’s growth doubling next year, while Nvidia expects about 70% growth because it is supply-constrained. Those two statements describe different sides of the model. Customer forecasts are a demand signal. Nvidia’s own approximately 70% expectation is management’s current long-range revenue view under the supply constraint.
FP&A should therefore avoid using the customer-demand statement as a second revenue forecast. A cleaner model records demand, available supply and expected revenue separately. Demand can exceed supply without creating recognised revenue, and a supply-constrained revenue expectation can rise if capacity improves. The useful watchlist is the gap between those drivers and the conditions that could close it.
Gross-margin assumptions need two horizons
The formal Q3 gross-margin guide is 74.0%, plus or minus 50 basis points. The call then extends the margin discussion beyond Q3. Management said extreme memory pricing has exceeded prior expectations, expects margins to bottom in Q4 in a 71% to 72% range, and expects them to settle at 72% to 73% in fiscal 2028 as executed price increases take effect.
Those later figures are useful, but they belong in a different horizon from the formal Q3 guide. The Q3 range can sit in the near-term guidance column. The Q4 and FY28 margin path should sit in a management-expectation or scenario column with the stated drivers: memory pricing, component costs, supply conditions and executed price increases. If any of those drivers change, the assumption can be revised without rewriting what Nvidia formally guided for Q3.
Map each Nvidia disclosure before rebasing the forecast
| Disclosure | Evidence state | Safe FP&A treatment | Control to retain |
|---|---|---|---|
| Q2 FY27 revenue: $96.2bn | Reported actual | Closed-period actuals and variance bridge | Quarter, measure and source |
| Q2 Data Center revenue: $89.0bn | Reported segment actual | Segment actuals; no invented segment guidance variance | Segment definition and comparison base |
| Prior Q2 revenue outlook: $91.0bn ±2% | Superseded formal guidance | Historical variance benchmark | Issue date, range and midpoint |
| Q3 revenue outlook: $108.0bn ±2% | Current formal guidance | Near-term external reference range | Range, midpoint, issue date and China assumption |
| Q3 gross margin: 74.0% ±50 bps | Current formal guidance | Near-term margin reference range | GAAP/non-GAAP basis, range and issue date |
| FY28 revenue growth: about 70% | Preliminary management expectation | Long-range assumption or scenario, explicitly labelled | Fiscal period, supply constraint and source date |
| Customer forecasts point to faster growth | Demand signal | Demand-driver scenario, not booked revenue | Conversion, capacity and timing assumptions |
| Q4 71%–72%; FY28 72%–73% margin path | Longer-range management expectation | Margin scenario tied to memory cost and pricing drivers | Component costs, price actions and update trigger |
What FP&A should lock before the next refresh
- Close Q2 against the old guide. Keep actual, midpoint variance and range outcome as separate fields.
- Load Q3 guidance without collapsing its range. Preserve $108.0 billion ±2% and 74.0% ±50 basis points as the issued external reference.
- Label FY28 separately. Record approximately 70% growth as a preliminary management expectation with the supply constraint attached.
- Keep demand and supply as drivers. Do not turn customer forecasts or capacity constraints into another revenue line.
- Version the margin scenario. Track the Q4 and FY28 margin path against memory pricing, component costs and executed price increases.
The forecasting value in Nvidia’s August 26 disclosure comes from preserving those states, not from choosing the largest number. Q2 is actual. Q3 has formal revenue and margin guidance. FY28 has a preliminary revenue expectation and a longer-range margin path. Demand and supply explain the gap between them. Keeping those labels intact gives FP&A a model that can be updated when Nvidia’s next disclosure changes the evidence rather than rewriting history.
