Nvidia's 75% gross margin: how much room is there for competition?
Nvidia’s 75% gross margin makes small changes worth billions. Explore customer concentration, AI demand and a transparent margin sensitivity calculation.
Wealth / The Compound journalPhoto: Christian Wiediger / Unsplash (Unsplash Licence) · Illustrative image.
Archival cover: Christian Wiediger / Unsplash; published on Unsplash in 2019. Original photograph · Unsplash Licence. Capture date not independently verified. The consumer graphics card pictured is not a current data-centre accelerator.
Research date: 18 October 2026. Reported figures are historical; illustrative calculations are identified below.
Nvidia’s investment case depends on more than selling additional computing equipment. It also depends on how much profit the company retains from each dollar of sales as customers gain scale, competing systems improve and new product generations arrive.
A business can remain central to AI infrastructure while its economics become less exceptional. Equally, lower prices per unit of useful computing can expand demand enough to support both the supplier and its customers. Investors need to distinguish those possibilities rather than treat an expanding AI market as a complete valuation argument.
Start with the right fiscal period
Nvidia’s second-quarter fiscal 2027 results cover the period ended 26 July 2026. The fiscal-year label does not mean these are forecasts for calendar 2027.
The company reported revenue of $96.2 billion, Data Center revenue of $89.0 billion and a 75.0% GAAP gross margin. Using the rounded revenue figures, Data Center represented approximately 92.5% of sales. GAAP diluted EPS was $2.46; the separately reported non-GAAP figure was $2.22. Those measures should not be swapped to create a more favourable comparison.
Gross margin measures revenue remaining after the cost of sales. It is not the same as operating margin or the amount that can be distributed to shareholders.
What a five-point margin change means
Compound’s sensitivity below holds revenue fixed at the rounded reported $96.2 billion. Only gross margin changes. These are hypothetical cases, not company guidance or estimates of the next quarter.
| Gross-margin assumption | Implied gross profit | Difference from the 75% baseline |
|---|---|---|
| 75% | $72.15bn | — |
| 72% | $69.26bn | −$2.89bn |
| 70% | $67.34bn | −$4.81bn |
At this sales level, each percentage point of gross margin is worth approximately $962 million of quarterly gross profit. That is why apparently small changes deserve attention.
There is an important counterpoint. Revenue at a 70% gross margin would need to reach approximately $103.07 billion to produce the same $72.15 billion of gross profit as the baseline: about 7.1% higher sales. Growth can offset compression. It cannot be assumed to do so automatically, and equal gross profit would still not guarantee equal operating profit if research or other expenses increased.
Customer diversity needs careful interpretation
Nvidia’s quarterly SEC filing distinguishes direct purchasers from indirect customers. A distributor or system assembler may buy the equipment, while a cloud operator or AI developer ultimately drives demand. Counting invoice recipients is therefore not the same as measuring economic diversification.
The filing reported one direct customer at 16% of quarterly revenue and said a limited number of indirect customers remained significant. It also identified product mix as a reason for the year-on-year gross-margin improvement. Those disclosures support a more useful research question than simply asking how many companies buy GPUs: how dependent is future purchasing on a small group of end users’ budgets?
A financing squeeze at an important buyer could influence equipment demand even if the broader technology continues to improve. Conversely, spending from a wider range of commercially successful customers would make the demand story more durable.
The strongest bull and bear arguments
The bullish case is that buyers evaluate total useful output, not just chip prices. A system that helps customers complete more valuable work with less engineering effort can sustain pricing power. If demand expands faster than suppliers can satisfy it, customers may prioritise reliable deployment over small hardware savings.
The bearish case is that today’s strongest customers have an incentive to reduce dependence on an expensive supplier. Alternatives need not replace every workload to influence negotiations. Product transitions, inventory commitments and less favourable sales mix could also weaken margins without causing an outright revenue decline.
For context on the buyer’s side, Microsoft’s own earnings call describes substantial spending on CPUs and GPUs. That is evidence of infrastructure investment, not proof that all of that spending becomes Nvidia revenue or that it will recur indefinitely.
A better next-quarter checklist
Read revenue and gross margin together. Check customer concentration using both direct and indirect disclosures. Examine inventory provisions and purchase commitments before attributing every margin change to competitive pressure. Finally, ask whether customers are generating returns that can support continued purchases.
An excellent supplier can still be a disappointing investment at the wrong price. This analysis deliberately makes no claim about today’s valuation: it identifies the operating evidence needed before applying a valuation multiple.
Follow Nvidia’s continuous stock research page and Compound’s earnings archive for the reported financial record.
Sources and methodology
Calculations are Compound’s own, using the stated reported figures and rounded inputs where identified. These articles are general investment research, not personal financial advice.
