Partial HHI: Measuring Supplier Concentration With Incomplete Data

Key findings

• Five quantified Micron supplier relationships cover 25.45% of the displayed cost base and contribute approximately 196.9 HHI points on a 0-to-10,000 scale.
• Normalizing only those five suppliers produces a roughly 3,040 HHI for the observed subset, but that statistic must not be presented as Micron's full supplier HHI.

Research question: How can investors calculate a useful supplier-concentration index when only part of a company's procurement network has quantified cost shares?

Calculate a partial supplier-concentration index by squaring each known supplier cost share and summing the results, then label it as a partial HHI rather than the company's full HHI. The calculation can still compare the concentration visible in the dataset, but it cannot measure suppliers that are missing or unquantified. Micron provides a useful example because five quantified upstream relationships are visible in the supplied Altsets network:

Together, those five relationships represent 25.45% of Micron's displayed cost base. That is enough to calculate a lower-bound contribution to supplier concentration.

HHI squares each supplier share

The Herfindahl-Hirschman Index adds the square of each exposure share. Using a 0 to 10,000 scale, the five observed Micron supplier shares contribute approximately 196.9 HHI points. That is not Micron's full supplier HHI. It is the HHI contribution from the five observed relationships.

Why 196.9 is a lower bound

Any additional supplier with a positive cost share adds another squared term. The missing relationships therefore cannot reduce the 196.9 points already contributed by the observed suppliers. The true full-company supplier HHI must be at least that high if the relationship percentages use compatible denominators. How much higher it is depends on the suppliers that are not included in the calculation.

Do not normalize the observed set and call it company concentration

The five known suppliers represent 25.45% of the displayed cost base. If those five shares are normalized to 100% of the observed subset, their HHI is roughly 3,040. That number answers a different question: How concentrated is the known five-supplier subset internally?

It does not answer: How concentrated is Micron's entire supplier base?

Those two calculations should never be labeled the same way.

ASML drives most of the observed concentration

ASML's 11.91% cost share contributes about 141.8 HHI points by itself. Lam Research contributes about 30.5 points. Applied Materials contributes about 14.7.

KLA contributes about 8.1. Shin-Etsu contributes about 1.8. The squaring operation makes larger relationships matter disproportionately. That is why HHI can reveal something a simple supplier count cannot.

Supplier count and concentration can move in opposite directions

A company can add many tiny suppliers without meaningfully reducing dependency on one dominant supplier. It can also replace one large relationship with several medium-sized ones and materially reduce HHI. This makes the index useful for tracking structural changes.

The investor still needs consistent historical coverage to compare periods properly. A changing dataset cannot be mistaken for a changing supply chain.

HHI does not measure substitutability

A low HHI can coexist with high operational risk. Five suppliers may serve five completely different functions. Micron's upstream set illustrates this problem.

ASML, Lam Research, KLA, Applied Materials, and Shin-Etsu participate in different parts of semiconductor manufacturing. The fact that several suppliers exist does not mean they can substitute for one another. HHI measures concentration of economic exposure. It does not measure technical redundancy.

A better concentration dashboard uses multiple metrics

A useful supplier-risk screen can show known cost-share coverage, lower-bound HHI, largest supplier cost share, top-three known supplier share, number of quantified suppliers, number of structural-only suppliers, functional category, and substitution difficulty. That combination is more informative than publishing one concentration score without context.

Top-three concentration adds another view

For the five displayed Micron suppliers, the three largest customer-cost shares are ASML, Lam Research, and Applied Materials. Together they account for 21.27% of Micron's cost base in the displayed data. Again, that is a known-share statistic. It does not imply the three are Micron's three largest suppliers overall unless coverage is complete enough to support that claim.

A repeatable partial-HHI workflow

1. Confirm all percentages use the same denominator.
2. Record the share of total costs covered by known relationships.
3. Square each known supplier share.
4. Add the squared terms.
5. Label the result as an observed or lower-bound HHI contribution.
6. Do not normalize incomplete coverage and call it full-company HHI.
7. Track coverage separately from concentration.
8. Add product-function and substitution analysis before drawing resilience conclusions.
9. Use point-in-time coverage for historical comparisons.

The economic supplier-ranking guide explains how different relationship metrics produce different rankings. For relationship definitions and how to interpret them, read the Altsets supply chain data methodology. Browse Supply Chain Data Use Cases for other portfolio and company-risk methods.

Sources

• Altsets supply chain dataset
• Micron fiscal Q3 2026 filing

Related research

• A Shared Customer Does Not Make Two Suppliers Substitutes

• Rank Commercial Relationships by Dollars, Revenue or Cost Share

• Why Supplier Diversification Does Not Fix Customer Concentration

• HBM4’s Second Supply Chain: Logic Base Dies and Package Readiness

Methodology

Read the methodology for this research.

Cite this research

Altsets Research. "Partial HHI: Measuring Supplier Concentration With Incomplete Data." Published August 3, 2026. https://www.altsets.com/research/partial-hhi-supplier-concentration

Copy citationCanonical URLMethodology