Evaluate NVIDIA’s dependence on customers and suppliers as two separate risks: measure customer concentration using revenue shares across clearly defined periods, and assess supply exposure by production stage, geography, commitments, and the flexibility of alternatives. NVIDIA’s fiscal 2027 second-quarter filing discloses substantial customer concentration, while its public filings name important suppliers but do not quantify each supplier’s share of spending or production.
Start by separating the questions
A customer-concentration measure asks how much revenue comes from a small number of buyers. A supply-dependence assessment asks whether NVIDIA can obtain the manufacturing, memory, packaging, and assembly capacity it needs, and how costly or slow it would be to adjust if a provider or region were disrupted.
As an Amazon Associate I earn from qualifying purchases.
Those are related but distinct exposures. A large customer changing orders can affect revenue timing and capacity planning. A disruption at a critical supply stage can limit production even if NVIDIA names multiple providers elsewhere in its supply chain. These are potential pathways implied by the business model and risks described in NVIDIA’s filings, not claims that a specific disruption has occurred.
Use the company’s fiscal periods carefully. NVIDIA’s fiscal 2027 Q2 was the quarter ended July 26, 2026; its fiscal 2026 year ended January 25, 2026. The latest quarterly figures below are from NVIDIA’s Form 10-Q filed August 27, 2026, and the annual supplier baseline is from its fiscal 2026 Form 10-K.
#1 Best Overall
- AI Performance: 767 AI TOPS
- OC mode: 2632 MHz (OC mode)/ 2602 MHz (Default mode)
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Axial-tech fan design features a smaller fan hub that facilitates longer blades and a barrier ring that increases downward air pressure
- A 2.5-slot design maximizes compatibility and cooling efficiency for superior performance in small chassis
How concentrated are NVIDIA’s disclosed direct customers?
NVIDIA’s Form 10-Q reports that one direct customer represented 16% of total revenue in fiscal 2027 Q2. For the first half of fiscal 2027, three direct customers represented 16%, 15%, and 13% of total revenue. These customers were primarily associated with Compute & Networking, and their identities are not disclosed.
The annual comparison is a different measurement window: NVIDIA’s fiscal 2026 Form 10-K reports that two direct customers represented 22% and 14% of full-year revenue, primarily in Compute & Networking. The filings do not establish that the customers in one period are the same customers in another.
Rank #2
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Powered by GeForce RTX 5070 Ti
- Integrated with 16GB GDDR7 256bit memory interface
- PCIe 5.0
- WINDFORCE cooling system
| Reporting period | Direct-customer concentration disclosed | How to read it |
|---|---|---|
| Fiscal 2027 Q2, quarter ended July 26, 2026 | One customer: 16% of total revenue | Quarterly share; primarily associated with Compute & Networking. Customer identity not disclosed. |
| First half of fiscal 2027 | Three customers: 16%, 15%, and 13% of total revenue | Half-year shares; primarily associated with Compute & Networking. Customer identities not disclosed. |
| Fiscal 2026, year ended January 25, 2026 | Two customers: 22% and 14% of total revenue | Full-year shares; primarily associated with Compute & Networking. Customer identities not disclosed. |
Do not add percentages from different rows: the quarterly, half-year, and full-year figures have different denominators and customer mixes. They show that material revenue concentration is recurring across reported periods, not that one particular buyer held a stable share over time.
Why direct-customer shares do not show the whole customer picture
Billing customers and end users may differ
NVIDIA’s definition of direct customers includes add-in-board providers, distributors, original design manufacturers (ODMs), original equipment manufacturers (OEMs), cloud service providers, AI model makers, and system integrators. Indirect customers buy through those direct customers and can include cloud providers, AI clouds, AI model makers, enterprises, and public-sector entities. Hardware may pass through intermediaries, so the party invoiced by NVIDIA need not be the ultimate user.
Rank #3
- Powered by the NVIDIA Blackwell architecture and DLSS 4. System Requirements: Minimum 850W PSU with 16-pin 12V-2x6 (12VHPWR) connector required. Verify before purchasing.
- Military-grade components deliver rock-solid power and longer lifespan for ultimate durability. Compatibility: 348mm (13.7") length, 3.6 slots, 4.3 lbs. Confirm case clearance and slot spacing. GPU bracket included.
- Protective PCB coating helps protect against short circuits caused by moisture, dust, or debris
- 3.6-slot design with massive fin array optimized for airflow from three Axial-tech fans
- Phase-change GPU thermal pad helps ensure optimal thermal performance and longevity, outlasting traditional thermal paste for graphics cards under heavy loads
Indirect exposure is estimated, not a second set of directly reported shares
NVIDIA says a significant amount of revenue comes from a limited number of indirect customers, with some individual indirect customers representing 10% or more. The company estimates this downstream revenue using factors including purchase orders, product specifications, and internal sales data. It also describes meaningful revenue attributable to an unnamed AI research and deployment company through cloud services purchased from NVIDIA customers.
Keep that disclosure separate from the direct-customer percentages. The indirect amounts are estimates and can relate to the same sales channel represented by direct-customer revenue; adding them mechanically to direct shares would risk double-counting. NVIDIA’s Form 10-Q says, “Our revenue is concentrated among a limited number of direct and indirect customers and this trend may continue.”
Rank #4
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Powered by GeForce RTX 5060
- Integrated with 8GB GDDR7 128bit memory interface
- PCIe 5.0
- WINDFORCE cooling system
Keep customer geography separate from supply-chain geography
For fiscal 2027 Q2, 38% of total revenue came from customers headquartered outside the United States; for the first half of fiscal 2027, the corresponding share was 30%, according to NVIDIA’s Form 10-Q. These percentages classify revenue by the headquarters of direct customers. They do not identify where every product was shipped or where the ultimate end user is located.
Recommended Free Tools
Supplier geography measures something different: where manufacturing and other supply-chain stages are concentrated. Do not treat the customer-headquarters percentages as a proxy for production location or as evidence of supplier concentration.
Best Value
- Powered by the NVIDIA Blackwell architecture and DLSS 4 OC mode: 2640MHz/Default mode: 2610MHz (Boost Clock)
- Military-grade components deliver rock-solid power and longer lifespan for ultimate durability
- Protective PCB coating helps protect against short circuits caused by moisture, dust, or debris
- 3.125-slot design with massive fin array optimized for airflow from three Axial-tech fans
- Phase-change GPU thermal pad helps ensure optimal thermal performance and longevity, outlasting traditional thermal paste for graphics cards under heavy loads
Map suppliers by the stage they serve
NVIDIA’s fiscal 2026 Form 10-K names providers across several stages. Read the list as evidence of external supply relationships, not as a ranking or a breakdown of production shares.
| Supply-chain stage | Providers or capability named in NVIDIA’s fiscal 2026 Form 10-K | What the disclosure establishes |
|---|---|---|
| Wafer fabrication | Taiwan Semiconductor Manufacturing Company (TSMC) and Samsung Electronics | Named wafer foundries; no supplier-by-supplier production or spending percentages are stated. |
| Memory | SK hynix, Micron Technology, and Samsung Electronics | Named memory providers; no supplier-specific allocation is stated. |
| Assembly, testing, and packaging | Contract manufacturers and independent subcontractors including Hon Hai Precision Industry (Foxconn), Wistron, and Fabrinet | Named participants in these activities; no supplier-specific allocation is stated. |
| Advanced packaging | NVIDIA says it uses CoWoS technology for semiconductor packaging | The filing identifies the technology but does not establish a supplier share for it. |
NVIDIA describes its supply chain as mainly concentrated in Asia and says it has expanded supplier relationships to build redundancy and resilience, while expanding in the United States and Latin America. It says, “We have expanded our supplier relationships to build redundancy and resilience in our operations to provide long-term manufacturing capacity aligned with growing customer demand.” This describes a diversification effort; it does not establish that a named provider can immediately replace another. Capacity, qualification, technology, and transition timing can differ by stage.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What the $279 billion commitment figure does—and does not—mean
NVIDIA reported $279 billion in supply and capacity commitments as of July 26, 2026, compared with $119 billion in the prior quarter as referenced in its fiscal 2027 Q2 Form 10-Q. The commitments primarily relate to memory and manufacturing facilities for data-center infrastructure systems and support long-term demand across current and future architectures.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →This is a forward commitment figure, not cash already spent, an amount assigned to one supplier, or a supplier-share estimate. NVIDIA says some supplier agreements permit inventory procurement against criteria it defines and may be canceled, rescheduled, or adjusted before firm orders; changes can result in additional costs. The scale of commitments is relevant to capacity planning and the cost of changing plans, but the total alone does not show how much is fixed, what can be altered, or how exposure is allocated among providers.
Use a repeatable framework to evaluate dependence
- Fix the period and denominator. Label every customer figure as quarterly, year-to-date, or full-year revenue. Compare like periods where possible, and do not combine overlapping windows.
- Separate direct buyers from downstream users. Record reported direct-customer shares separately from estimated indirect-customer exposure. Note whether identities are disclosed and whether the measures may overlap.
- Check the segment concentration. Note which business area a large customer share is primarily associated with. A concentration in Compute & Networking can make results more sensitive to demand and purchasing patterns in that part of the business.
- Map supply stages and locations. Track wafer fabrication, memory, packaging, assembly, and testing separately. Ask where each stage is concentrated and whether the filing identifies alternatives at that same stage.
- Read commitments alongside flexibility. Distinguish forward capacity obligations from purchases already made. Look for cancellation, rescheduling, adjustment rights, and possible change costs without assuming those rights make commitments cost-free.
- Assess substitutability rather than counting supplier names. Multiple names across the chain do not prove spare capacity, technical equivalence, or an immediate replacement path. Look for evidence about stage-specific capacity and transition options before concluding that redundancy removes risk.
- Track operational mechanisms. NVIDIA says inaccurate demand estimates, changes in customer purchases, production complexity, and supply constraints can contribute to volatility, delays, inventory provisions, and higher costs. It also says its business depends on consistent and reliable supply from overseas partners, especially in Taiwan and South Korea. Consider those risks alongside customer shares, not as separate proof of a particular future outcome.
How to interpret the disclosures without overstating them
- Supported: Customer concentration is material in the reported periods, and NVIDIA describes significant indirect-customer exposure.
- Supported: External manufacturing and supply relationships, including geographic exposure in Asia, are strategically important to the business.
- Not established by the supplier list: Each supplier’s share of NVIDIA’s spending, production, or capacity, or the extent to which one supplier can replace another.
- Not established by concentration alone: That a major customer will reduce purchases, that a disruption will occur, or that the company is necessarily a good or poor investment. Those conclusions require evidence beyond the concentration metrics.
A named infrastructure project is context, not an identity clue
On August 17, 2026, NVIDIA announced that OpenAI would be the customer for the PORTS-Pike AI compute campus in Ohio, with SB Energy building, owning, and operating the data center under a 20-year lease. NVIDIA said it would secure land, power, and shell capacity and invest $1.5 billion in SB Energy. This illustrates how customer demand can connect to infrastructure and partner arrangements. The project announcement does not identify OpenAI as one of the unnamed customers in NVIDIA’s concentration disclosures.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
