Accelerated Computing Direct Customer Revenue Concentration measures how much of NVIDIA's reported revenue comes from customers that buy directly from NVIDIA. This issuer-reported metric is useful, but its accounting perimeter matters because a direct customer can be a distributor, original design manufacturer, system integrator, cloud service provider, AI model maker, or other intermediary rather than the final user of the computing capacity.
For fiscal Q2 2027, NVIDIA said one direct customer represented 16% of total revenue, primarily in Compute & Networking. A 16% largest-direct-customer share leaves 84% of reported revenue attributable to all other direct and indirect customer relationships combined, but that residual does not reveal end-customer concentration. A year earlier, two direct customers represented 23% and 16% of quarterly revenue.
The first-half concentration was broader than the quarterly headline
The quarterly 16% figure describes only the largest disclosed direct customer in that three-month period. For the first half of fiscal 2027, NVIDIA reported three direct customers at 16%, 15%, and 13% of total revenue.
Those disclosed shares sum to 44% of first-half revenue. That does not mean the remaining 56% was economically diversified across many unrelated end customers. It means 44% was attributable to three direct purchasing relationships that individually crossed NVIDIA's disclosure threshold.
The prior-year first half had two disclosed direct customers at 20% and 15%, or 35% in total. The comparison shows why concentration should be tracked across more than one quarter. Customer mix can change quickly as large infrastructure deployments move between buyers and periods.
Direct customer concentration can understate downstream dependence
NVIDIA separately describes indirect customers as companies that purchase products through direct customers. These can include cloud service providers, AI clouds, AI model makers, enterprises, and public-sector entities.
That distinction creates an important analytical boundary. One large downstream customer can buy cloud services or infrastructure through several direct counterparties. NVIDIA therefore cautions that indirect-customer revenue is an estimate based on purchase-order information, product specifications, internal sales data, and other sources.
In fiscal Q2 2027, NVIDIA said one AI research and deployment company contributed a meaningful amount of revenue by purchasing cloud services from NVIDIA's customers. The company did not identify that amount as the same thing as the 16% direct-customer share.
For investor analysis, the 16% figure should therefore be read as channel concentration, not a complete measure of economic end-customer concentration.
Accounts receivable adds a second concentration lens
Revenue concentration measures sales recognized during a period. Accounts-receivable concentration measures who owes NVIDIA money at a point in time.
As of July 26, 2026, five direct customers accounted for 22%, 14%, 13%, 11%, and 10% of NVIDIA's accounts-receivable balance. That is a different denominator from revenue, so the percentages should not be mixed into one concentration ratio.
The receivables data can still be useful as a cross-check. A customer that is a large share of both current sales and unpaid receivables creates a different exposure from a customer whose purchases are large but settle quickly.
Geography follows the direct customer's headquarters
NVIDIA also assigns geographic revenue using the headquarters location of the direct customer. The end customer and shipping destination can be elsewhere.
That means direct-customer concentration, geographic mix, and end-market demand are three different views of the same commercial chain. Treating them as interchangeable can produce false conclusions about where demand originates or which downstream buyer is driving it.
For market-platform context, pair this page with Hyperscale Revenue and ACIE Revenue. For the segment accounting perimeter, use Compute & Networking Segment Revenue.
Primary source: NVIDIA fiscal Q2 2027 Form 10-Q.
Part of the Accelerated Computing Platform Economics
See It in Company Research
These companies are examples of how the concept is reported or discussed in public filings. Definitions can differ by issuer; these links open company research rather than a normalized metric comparison.
- NVDAOpen operating-model research →10 of 10 reviewed concepts in Accelerated Computing Platform EconomicsData Center demand and customer mix4 of 4 bridge concepts supportedContinue through this bridge:AI Clouds, Industrial & Enterprise RevenueData Center Revenue GrowthHyperscale Revenue
Continue Research
Continue from the concept into the Grizzly Bulls research surface that best matches the next question. These links are research continuations, not recommendations or required steps.
Compare semiconductor stocks
Compare direct-customer concentration with market-platform mix without assuming it reveals end-customer concentration.
Explore more topics in the Financial Research Encyclopedia.