When a 10-K Says Customer A: How Supply-Chain Data Helps
September 5, 2026
Altsets
Research by Altsets Research
Use named supplier-customer relationships to investigate anonymous customer-concentration disclosures without falsely assigning a filing percentage to a specific company.
Data used:Altsets Supply Chain Intelligence: 90k+ entities, 400k+ relationships, 20+ years of history.
Key findings
- Nvidia publicly discloses large anonymous direct-customer concentration while named supply-chain relationships provide candidate counterparties and channel context, not automatic identities.
- A filing concentration percentage should not be attached to a named relationship unless period, customer role, accounting basis, and additional evidence support the identification.
Do not assign an anonymous filing percentage to a named company without direct evidence. Supply-chain data can identify plausible customer candidates for follow-up, but it cannot by itself prove that a mapped company is the filing's "Customer A."
Public-company filings often disclose customer concentration without naming the customers. Nvidia's fiscal 2026 annual report, for example, said one direct customer represented 22% of total revenue and another represented 14%, but did not identify either customer by name.
A supply-chain relationship graph can add named counterparties, but it should not be used to attach the filing's anonymous percentages to a specific company without supporting evidence. That distinction turns supply-chain data into a research tool instead of a guessing engine.
What "Customer A" actually tells you
A concentration disclosure establishes economic importance. It can tell you that a small number of direct customers represent a meaningful share of revenue. What it often does not tell you is the customer name, whether the customer is an ODM, OEM, distributor, cloud provider, or system integrator, which end customer ultimately consumes the product, and whether the relationship is direct or passes through another channel.
Nvidia's filing is unusually useful because it also defines these customer categories. The company says direct customers can include AIBs, distributors, ODMs, OEMs, cloud service providers, AI model makers, and system integrators. Indirect customers can purchase through those direct customers. That means the anonymous concentration percentage and the economic end user do not have to refer to the same entity.
Named relationship data solves a different part of the problem
The supplied Altsets Nvidia graph contains named downstream relationships with companies including Amazon, Microsoft, Super Micro Computer, Quanta Computer, and Samsung Electronics. Two displayed edges carry directional percentages: Nvidia to Quanta: 0.53% of Nvidia revenue and 1.76% of Quanta's cost base and Nvidia to Samsung: 0.04% of Nvidia revenue and 0.10% of Samsung's cost base. Those percentages are useful precisely because they are tied to named relationships.
But neither displayed percentage resembles the anonymous 22% or 14% direct-customer concentration disclosed in Nvidia's fiscal 2026 filing. The correct response is not to force a match.
Do not reverse-engineer an identity from one number
Suppose a filing says Customer A is 22% of revenue. A separate dataset contains a named customer with a large relationship. That is not enough to declare the named company Customer A.
The figures may measure different things. They may refer to different periods. The filing may use recognized revenue while the relationship dataset uses an estimate.
A direct customer can also sit between the supplier and the eventual end customer. A defensible identification requires stronger evidence.
What evidence can raise confidence
A filing-reader workflow can combine several clues:
- named supplier-customer relationships;
- revenue-concentration disclosures;
- accounts-receivable concentration;
- product announcements;
- purchase commitments;
- customer-specific statements;
- shipment or procurement records;
- geography;
- timing;
- relationship history.
The strongest conclusion may still be "candidate customer" rather than a definitive identity. That is a useful result.
Direct and indirect customers complicate the mapping
Nvidia explicitly distinguishes direct from indirect customers. Its fiscal 2026 filing says indirect customers primarily purchase through direct customers such as system integrators and distributors. That matters because a large cloud or AI company can drive economic demand without being the legal counterparty responsible for the largest direct revenue line.
A named supply-chain graph can therefore contain several economically connected companies around the same demand pool. The filing percentage should not automatically be assigned to whichever brand name looks most important.
Why relationship data is still valuable
The inability to name Customer A immediately does not make the graph less useful. It changes the job. The graph can generate candidate identities, classify customer types, reveal direct and indirect network structure, show which named relationships already have economic metrics, create a list of filings and disclosures to investigate, and track whether a candidate relationship existed at the relevant historical date. That is far more useful than searching the web for guesses about an anonymous disclosure.
A practical anonymous-customer workflow
- Record the filing period and concentration percentage.
- Record whether the disclosure describes a direct or indirect customer.
- Map named customer relationships at the same point in time.
- Keep relationship metrics separate from filing revenue percentages.
- Classify the named counterparties by channel role.
- Search official filings and product disclosures for customer-specific evidence.
- Use geography and timing as supporting evidence, not proof.
- Label unresolved identities as candidates.
- Revisit the mapping when new disclosures appear.
The Nvidia customer-concentration analysis shows what can be concluded from two named quantified relationships. This guide addresses the separate problem created when a filing discloses a large customer without naming it.
For relationship methodology and point-in-time constraints, read the Altsets supply-chain data methodology. Browse Supply-Chain Data Use Cases for other filing and investing workflows.