What Can Supply-Chain Data Not Tell You About a Stock?
September 14, 2026
Altsets
Research by Altsets Research
It cannot tell you fair value, future stock direction, exact shock pass-through, technical substitutability, or causality by itself. It maps economic relationships that still need valuation, product, fundamental, and market-expectation context.
Data used:Altsets Supply Chain Intelligence: 90k+ entities, 400k+ relationships, 20+ years of history.
Key findings
- The supplied Micron-Nvidia and ASML-Micron metrics show economic dependence, but neither supplier revenue share nor customer cost share is a stock beta, expected return, exact earnings sensitivity, or disruption pass-through coefficient.
- Relationship topology also does not prove technical substitution or causality, so product evidence, event controls, valuation, and market expectations remain necessary after the graph identifies the relevant counterparties.
Supply-chain data cannot tell you by itself whether a stock is cheap, whether the share price will rise, how much of a disruption will pass through to earnings, or whether one supplier can actually be replaced. It tells you where economic relationships exist and, when metrics are available, how important those relationships appear on each side. The investment conclusion still requires valuation, product knowledge, company fundamentals, market expectations, and evidence about how the relationship works.
A relationship percentage is not a stock-return forecast
The supplied Altsets data shows Nvidia representing 17.62% of Micron revenue in the displayed relationship. That is a large customer relationship for Micron, but it does not mean Micron stock should move 17.62% when Nvidia changes guidance. The percentage measures supplier revenue exposure, not equity beta, earnings sensitivity, or expected return.
The same distinction applies to the supplied ASML-Micron relationship. ASML accounts for 11.91% of Micron's cost base in the displayed data. That can make ASML an important company to investigate when thinking about Micron's production dependencies, but the percentage does not tell the investor what a one-week ASML disruption would do to Micron gross margin. Inventory, timing, substitutability, existing capacity, contractual arrangements, and management response all sit between the relationship metric and the financial outcome.
The graph does not prove technical substitutability
Two suppliers can sell into the same customer and still provide completely different products. Two customers can buy from the same supplier without competing for the same capacity. A structural relationship shows that an economic connection exists. It does not prove that one company can replace another or that both companies are exposed to the same operational bottleneck.
That is why replacement-supplier research needs product and qualification evidence in addition to the graph. An investor can use relationship data to find the candidate companies worth investigating, then use filings, product documentation, management commentary, and industry research to determine whether the products are actually substitutable. The network narrows the question. It does not answer the engineering problem.
The graph does not prove causality
If Nvidia reports strong demand and Micron rises afterward, the customer relationship creates a plausible transmission path. It does not prove that Nvidia caused the Micron move. Semiconductor stocks can react to the same industry news, factor flows, macro information, or AI sentiment at the same time.
The same problem appears in historical backtests. Connected stocks can be correlated because they share an industry rather than because information traveled through the direct relationship. A serious event study compares connected companies with appropriate controls and asks whether the relationship explains anything beyond sector and market effects. Economic linkage makes the hypothesis more credible, but it does not eliminate the need for falsification.
Missing data is not proof that the relationship is small
A structural relationship without a displayed relationship size, supplier revenue percentage, or customer cost percentage should remain unquantified. Missing is not zero. A company with several structural suppliers is not automatically less dependent on them than a company whose relationships happen to have estimated economic metrics.
This matters when ranking stocks. A dense quantified graph can look more diversified simply because more relationships have measurements. Coverage, evidence state, and metric availability need to remain visible so the investor does not confuse data completeness with business quality.
Supply-chain data is one lens, not the whole thesis
A company can have attractive network positioning and still be an unattractive stock because valuation is extreme, margins are deteriorating, management is allocating capital poorly, or expectations are already too high. The reverse can also happen. A concentrated supply chain can be acceptable if the company has strong bargaining power, technically superior suppliers, contractual protection, or a valuation that already reflects the risk.
The graph is best used to answer a narrower set of questions: who depends on whom, how concentrated are the observed relationships, which outside companies deserve attention, where can a shock plausibly travel, and what relationship changed. Those answers can materially improve an investment thesis without pretending the network contains every variable needed to price a security.
The conclusion is that the data maps dependence, not destiny
Supply-chain data is powerful because it reveals economic connections that sector labels and price charts can miss. It does not independently determine fair value, future returns, causal impact, technical substitutability, or exact shock pass-through. Use the graph to find the relationships that matter, then use other evidence to decide what those relationships mean for the stock.
The exposure-is-not-beta guide explains why relationship exposure should not be treated as stock sensitivity. The trust-incomplete-data guide explains how missingness and evidence limits should constrain the conclusion.
For relationship definitions and evidence limits, read the Altsets methodology.