How to Find Shared Customers Across Portfolio Holdings
September 10, 2026
Altsets
Research by Altsets Research
Find hidden demand overlap by identifying external customers connected to multiple holdings, then separate structural overlap from quantified exposure.
Data used:Altsets Supply Chain Intelligence: 90k+ entities, 400k+ relationships, 20+ years of history.
Key findings
- Micron and Nvidia share Quanta Computer as a visible downstream customer in the bounded semiconductor portfolio network.
- The Nvidia to Quanta edge is quantified while the Micron to Quanta edge is not, illustrating why structural overlap and quantified overlap should remain separate concepts.
Two holdings can look unrelated at the ticker or sector level while depending on the same downstream customer. A shared-customer screen asks which external customers connect to more than one company in a portfolio or research set. The semiconductor portfolio example containing ASML, Micron Technology, and Nvidia reveals one visible shared external customer: Quanta Computer.
Micron and Nvidia both connect downstream to Quanta in the supplied network. The Nvidia to Quanta relationship includes estimates of 0.53% of Nvidia revenue and 1.76% of Quanta's cost of goods sold. The Micron to Quanta edge is visible, but the supplied view does not provide comparable metrics for it.
The useful conclusion is structural: Micron and Nvidia share a downstream customer in the displayed network. That is enough to create a research question without pretending the overlap is fully quantified.
Why shared customers matter
Traditional portfolio diversification usually groups holdings by sector, industry, country, factor, or market capitalization. A shared customer creates another type of economic overlap. Two suppliers can belong to different reported industries while still depending on the same buyer, production program, end market, or capital-spending cycle.
If the shared customer changes demand, both holdings may be exposed to the same economic signal. The shared-customer screen is therefore useful for finding common demand dependencies beneath conventional portfolio labels.
Start with overlap, then measure it
The first step is structural:
- map customers for each holding;
- normalize company identities;
- count customers connected to multiple holdings;
- separate internal portfolio relationships from external counterparties.
Only after finding overlap should the analyst ask how material it is. This matters because shared connections can be weak, metrics can be missing, and one portfolio holding may have a much larger relationship with the customer than another.
The Quanta example
Within the bounded ASML, Micron, and Nvidia portfolio map, Quanta is the only visible external counterparty connected downstream to more than one holding. That makes it a useful worked example. The Nvidia to Quanta edge provides 0.53% of Nvidia revenue and 1.76% of Quanta COGS.
The Micron to Quanta edge lacks comparable displayed metrics in the supplied view. The correct conclusion is not that Quanta is the largest hidden demand concentration in the portfolio. The correct conclusion is that Quanta is a confirmed shared downstream connection worth investigating further.
Missing values should not be manufactured
A common temptation is to assign zero exposure when an edge lacks a metric. That would be incorrect.
The Micron to Quanta relationship exists in the map. Its missing displayed percentages remain unknown.
This means a shared-customer screen can have two layers: structural overlap, where the shared relationship is known and quantified overlap, where comparable economic metrics are available. The distinction should remain visible.
Shared customers can reveal hidden end-market concentration
A shared customer may point to a broader common driver. Multiple component suppliers selling to the same server manufacturer, automaker, consumer electronics company, or industrial buyer may be exposed to the same demand cycle even if the suppliers sit in different industries. The shared customer is therefore a bridge between company-level relationships and hidden economic clusters. The next question becomes whether other customers in the same end market create additional overlap.
Turning overlap into portfolio research
After finding a shared customer: identify which holdings connect to it, retrieve supplier revenue percentage for each relationship where available, retrieve customer cost percentage where relevant, add relationship size, check whether the relationships involve the same end market, compare the overlap with portfolio weights, and investigate whether other customers create the same demand cluster. This can reveal that holdings diversified by ticker are less diversified by underlying demand source.
Shared customers versus shared suppliers
The two screens answer different questions. A shared customer points to common demand exposure. A shared supplier points to common input or disruption exposure.
Both can exist in the same portfolio, but they should not be merged into one generic network count. The direction of the edge matters because the economic interpretation changes.
When this use case is strongest
Shared-customer analysis becomes especially useful when holdings sell into concentrated end markets, customer capex cycles matter, one large buyer influences multiple suppliers, portfolio holdings appear diversified on sector labels, or an earnings event at the common customer has just occurred. It becomes weaker when the overlap is unquantified, stale, or tied to unrelated business segments.
A repeatable shared-customer workflow
- Define the portfolio or company set.
- Fix the analysis date.
- Map first-degree customers for every holding.
- Normalize aliases and subsidiaries.
- Count external customers touching multiple holdings.
- Separate known overlap from quantified overlap.
- Rank quantified shared customers by the metric that matches the question.
- Investigate the common demand driver behind the overlap.
The hidden supply-chain concentration portfolio study contains the original Quanta example. This guide isolates the shared-customer method so it can be applied to other portfolios and research sets.
For relationship direction and missing-value treatment, read the Altsets supply-chain data methodology. Browse Supply-Chain Data Use Cases for other methods.