Can You Trust Supply-Chain Data When Companies Do Not Disclose Everything?
September 14, 2026
Altsets
Research by Altsets Research
Supply-chain evidence is useful when confidence is calibrated to what is actually known: structural relationships, directional metrics, missing values, dates, and unresolved counterparties should remain distinct.
Data used:Altsets Supply Chain Intelligence: 90k+ entities, 400k+ relationships, 20+ years of history.
Key findings
- Apple's 2025 Form 10-K describes a large, complex global supply chain and single or limited sources for critical components without providing investors a complete normalized map of every current relationship.
- Incomplete coverage does not make known relationships useless, but the strength of the investment conclusion should never exceed the evidence supporting the relationship, metric, product scope, or counterparty identity.
Yes, but only as incomplete and graded evidence. Supply-chain data can support useful decisions without observing every relationship, provided unknown values are not treated as zero and weak or structural-only edges are not presented as precise exposure.
That does not make supply-chain data useless. It changes how the data should be used. The investor should treat relationship evidence as graded information rather than assume every edge has the same certainty or completeness.
Public companies themselves disclose incomplete supply-chain pictures
Apple's 2025 Form 10-K describes a large and complex global supply chain, says the company relies on single or limited sources for many critical components, and notes that a significant majority of manufacturing is performed through outsourcing partners across several countries.
Those disclosures are important, but they do not provide a complete current map of every supplier, every component, every factory, and every commercial value. Public filings are written for financial disclosure, not to hand investors a normalized dependency graph.
The gap between useful disclosure and complete disclosure is normal.
Missing metrics should remain missing
One of the easiest ways to make relationship data misleading is to force every relationship into a number.
A structural supplier or customer edge can still be valuable because it tells the investor that an economic path exists. If the dataset does not contain a reliable relationship size, supplier revenue percentage, or customer cost percentage, the correct representation is to preserve that missingness.
The investor can then decide whether the relationship deserves more public research rather than treating an invented estimate as fact.
Quantified relationships deserve a different level of confidence
When a relationship has a directional economic metric, the investor can ask more precise questions. A supplier revenue percentage can show how important a customer is to the supplier. A customer cost percentage can show how much of the customer's cost base is associated with the supplier.
Even then, the metric does not answer every question. It may not identify the exact product, facility, contract term, replacement difficulty, or future durability of the relationship.
Quantification improves the analysis without eliminating the need for context.
A relationship should be corroborated when the investment conclusion is important
The stronger the conclusion, the stronger the evidence should be.
If an investor is only using a structural edge to decide which filing to read next, the relationship can function as a research lead. If the investor is about to reduce a large position because of a customer dependency, the relationship deserves additional confirmation from filings, product announcements, supplier lists, management commentary, or other relevant evidence.
This is the same standard good fundamental research applies to any important claim.
Anonymous customers are a special case
Companies sometimes disclose large customers without naming them. A similar percentage appearing elsewhere can be a clue, but it is not proof that the anonymous customer has been identified.
Defensible identification can require extensive cross-referencing across dated filings, supplier and customer disclosures, historical relationship evidence, channel structure, entity changes, and other records. Some anonymous relationships may remain unresolved.
A serious dataset should preserve that uncertainty rather than publish a confident identity because two numbers happen to resemble each other.
Incompleteness does not erase comparative value
An investor does not need every relationship in the economy before one verified dependency becomes useful. If two portfolio holdings clearly share an important customer, that overlap exists even if other unknown relationships remain outside the dataset.
The limitation affects what the investor can conclude, not whether the known relationship contains information.
This is similar to fundamental analysis more broadly. A filing never tells the investor everything about a company, but the disclosed facts can still improve a decision.
The right question is whether the evidence supports the claim
A structural relationship can support "these companies are connected." A directional percentage can support "this customer is economically important to this supplier." Public product evidence can support "the relationship currently involves this product or business function."
Those are different claims with different evidence requirements.
The investor gets into trouble when one type of evidence is stretched to answer a stronger question than it actually supports.
The conclusion is calibrated trust
Supply-chain data should not be treated as an oracle. It should be treated as an evidence layer with explicit boundaries.
The most trustworthy workflow preserves direction, missingness, dates, structural-only relationships, and uncertainty around product scope or anonymous counterparties. The value comes from knowing more than you knew before without pretending to know more than the evidence can support.
The filings versus supply-chain-data guide explains why the two evidence sources complement each other. The missing-metrics guide explains why an unquantified relationship should not be treated as zero.
For relationship definitions and evidence limits, read the Altsets methodology.