Network and Concentration Research
Research on upstream and downstream paths, shared counterparties, and supplier or customer concentration.
Publications
AI Data Centers Shift the Bottleneck to Electricity Infrastructure
Network path
How is the AI data center buildout shifting infrastructure bottlenecks from GPUs toward grid connections, transformers, generation, switchgear, copper, cooling, and backup power?
AI data center growth is shifting the critical path toward deliverable electricity, exposing bottlenecks in grid connections, transformers, turbines, switchgear, copper, cooling, and backup power.
AI Power Demand Strains the Gas Turbine Supply Chain
Network path
How is AI data center power demand propagating below gas turbine OEMs into hot-section components, specialty alloys, coatings, replacement parts, and maintenance?
AI data center power demand is pushing the gas turbine bottleneck below GE Vernova, Siemens Energy, and Mitsubishi Power into cast airfoils, nickel superalloys, coatings, replacement parts, and turbine maintenance.
China's Control of Battery Midstream Processing
Network path
Which parts of the battery supply chain remain most concentrated in China even when cell manufacturing or raw-material sourcing moves elsewhere?
China's battery advantage now sits in qualified midstream conversion: graphite anodes, LFP cathodes, electrolyte inputs, separators, and the customer qualification needed to replace them.
Co-Packaged Optics Shifts the AI Networking Bottleneck
Network path
Which supplier functions gain importance as AI networking moves from pluggable optics toward co-packaged optics?
Co-packaged optics shifts AI networking value toward high-power lasers, silicon photonics, advanced packaging, fiber attachment, and precision manufacturing.
Indium Phosphide Capacity Constrains AI Networking
Network path
Where can capacity become constrained below AI switches and optical modules as 800G, 1.6T, and future 3.2T networking expands?
The AI networking bottleneck is moving upstream from switches and optical modules into indium phosphide substrates, epitaxy, and lasers, where capacity is expanding much more slowly than demand for 800G, 1.6T, and future 3.2T optics.
TSMC's Third-Tier Materials and Equipment Suppliers
Network path
Which second and third order suppliers sit beneath TSMC's 2026 capacity build, and where does TSMC appear economically more important to the upstream supplier than the supplier is to TSMC's total cost base?
TSMC's 2026 capex surge reaches beyond the obvious equipment names into Japanese wafer, photoresist, CMP, vacuum, packaging and tool-component suppliers where customer dependence can be more concentrated.
U.S. Polysilicon Capacity Lacks Guaranteed Domestic Demand
Network path
Why can domestic US polysilicon capacity remain underutilized even while semiconductor reshoring increases strategic demand for the material?
US polysilicon policy is trying to secure a material essential to both chips and solar, but the economics still depend on downstream wafer buyers, qualification, and whether domestic material is actually preferred over foreign alternatives.
Block Bootstrapping for Supply-Chain Strategy Tests
Network path
How can a quant estimate uncertainty for a supply-chain strategy without using bootstrap resamples that break temporal dependence, shared-customer events, and network clusters?
Resample time blocks and preserve relevant customer, supplier, or event clusters instead of independently shuffling stock rows that are economically connected.
Customer Payment Delays Turn Suppliers Into Lenders
Customer concentration
How can slower payment by a major customer weaken supplier cash conversion and increase short-term financing needs even when reported revenue remains stable?
Longer customer payment terms raise receivables and funding needs even when sales are unchanged, turning revenue concentration into working-capital concentration.
ASML Is Most Exposed to a Micron Purchasing Cut
Customer concentration
Among five quantified Micron supplier relationships, which supplier has the greatest customer concentration to Micron if Micron-specific purchasing weakens?
ASML has the largest displayed dependence on Micron by supplier revenue share at 7.64%, followed by Lam at 5.61%, KLA at 4.59%, Applied Materials at 2.96%, and Shin-Etsu at 1.83%.
Beta-Neutral Supply-Chain Trading Baskets
Shared counterparty
How can a quant express a supply-chain hypothesis as a market-neutral or factor-controlled basket without turning directional relationship percentages into mechanical portfolio weights?
Use customer and supplier relationships to choose an economically coherent candidate set, then separate that selection logic from the beta, sector, style, volatility, liquidity, and covariance controls used to size the actual trade.
Can Supply-Chain Data Build Better Stock Peer Groups Than Industry Codes?
Shared counterparty
Can a quant improve cross-sectional normalization, relative-value research, and peer benchmarking by defining peers from customer and supplier relationships instead of relying only on industry classifications?
Shared customers, shared suppliers, and similar dependency profiles can define dynamic economic peer groups for normalization, relative value, stat arb, and risk modeling that compete directly with fixed sector classifications.
Can Supply-Chain Data Improve Statistical Arbitrage Pair Selection?
Shared counterparty
Can customer and supplier relationships reduce the stat-arb pair search space and produce more economically interpretable candidate pairs without assuming that network similarity guarantees mean reversion?
Use shared customers, shared suppliers, and network similarity as an economic filter for stat-arb candidate generation, then require the same out-of-sample spread, liquidity, and cost tests as any other pairs strategy.
Choosing the Next Company to Research
Network path
How can supply-chain relationships tell an investor which outside customer, supplier, or counterparty deserves to be researched next after the target company itself?
Research the outside company most capable of changing the unresolved part of your thesis. For a supplier that is often a major customer; for a customer it can be a critical supplier or bottleneck.
Choosing Where to Own an Investment Theme
Network path
How can an investor decide where in a supply chain to own an investment theme when customers, suppliers, and equipment providers offer different combinations of demand exposure, cyclicality, concentration, and risk?
Supply-chain structure lets investors choose where in an investment theme to take exposure by comparing direct customers, upstream suppliers, equipment providers, and second-order beneficiaries rather than defaulting to the obvious stock.
Customer Concentration Should Widen the Revenue Scenario Range, Not Automatically Lower the Forecast
Customer concentration
How should customer concentration change the range of plausible revenue outcomes without turning a large customer relationship into an automatically bearish forecast?
A dominant customer increases the sensitivity of the revenue path to one outside company. That should widen upside and downside scenarios before it automatically changes the central forecast.
Customer Diversification Within One Industry Cycle
Customer concentration
How much protection does customer diversification provide when several major customers compete with one another but remain exposed to the same industry cycle?
Selling to several competing customers can reduce dependence on one buyer while leaving a supplier exposed to the same industry's capital spending, inventory, technology, and demand cycle.
Diversifying Within Semiconductors by Customer Ecosystem
Shared counterparty
Can an investor improve diversification within one sector by selecting companies whose important customers and immediate demand paths differ?
Two semiconductor holdings can repeat the same major customer while another same-sector company introduces different direct demand paths. Customer-network diversification is therefore a separate layer from ticker count and sector classification.
ETF Overlap With Existing Stock Dependencies
Shared counterparty
How can an investor judge whether an ETF adds new economic paths to an existing stock portfolio instead of simply packaging more securities around dependencies the investor already owns?
An ETF can be diversified by itself while repeating the same customer, supplier, foundry, and theme dependencies already present in an investor's individual-stock portfolio.
Famous Customers Do Not Make Supplier Stocks Safer
Customer concentration
How should investors balance the benefits of serving large high-quality customers against the concentration and bargaining risks created when those customers represent a large share of supplier revenue?
No. Strong customers can improve demand visibility and validate a supplier's product, but large customer dependence can also increase concentration, bargaining-power, financing, and event risk.
Finding Picks-and-Shovels Stocks With Supply-Chain Data
Customer concentration
How can an investor use customer relationships to find suppliers that can participate across several competing winners inside the same investment theme?
Look for a supplier that sells relevant inputs to several competing leaders rather than depending on one winner. The best candidates reduce winner-selection risk without pretending that customer diversity removes the broader industry cycle.
Five Micron Suppliers Account for 25.45% of the Displayed Cost Base
Supplier concentration
How much of Micron's displayed cost base is associated with the five quantified supplier relationships for ASML, Lam Research, Applied Materials, KLA, and Shin-Etsu Chemical, and how concentrated is that visible exposure?
The five quantified Micron supplier relationships in the supplied data sum to 25.45% of Micron's cost base. ASML and Lam alone account for 17.43 percentage points, showing that the visible quantified exposure is concentrated toward the top.
Graph Laplacian Regularization for Supply-Chain Stock Models
Network path
Can customer and supplier relationships improve multi-stock prediction by regularizing connected company models toward one another without forcing every stock into one global model or a graph neural network?
Yes. Graph-Laplacian or multi-task regularization can let economically connected stocks borrow statistical strength while keeping separate company models and avoiding a full graph neural network.
Graph Neural Networks for Supply-Chain Trading
Network path
When does a graph neural network add enough value to justify its complexity for supply-chain trading research compared with simple network statistics, tree models, or linear cross-sectional features?
Graph neural networks can model direction and multi-hop firm relationships, but simple network features remain the baseline a complex architecture should beat under strict point-in-time testing.
Graph-Aware Cross-Validation for Supply-Chain Models
Network path
How should a quant split training and test data when customer and supplier relationships connect firms across the dataset and graph features can transmit information across an ordinary random cross-sectional split?
Connected companies are not independent rows, so quants should distinguish time, firm, and network-cluster generalization instead of assuming a random stock split proves a relationship model works on unseen parts of the economy.
Hidden Customer Crowding Across Quant Strategies
Shared counterparty
Can a multi-strategy quant portfolio measure crowding in customer and supplier space so independently designed alpha models do not unknowingly depend on the same external company?
Yes. Different models can own different stocks and still converge on the same customer, supplier, or bottleneck, creating hidden dependency crowding that ordinary strategy labels and normal-period return correlations may not reveal.
Hidden Supply-Chain Concentration in Diversified ETFs
Shared counterparty
How can an ETF or index remain diversified by holdings and sector weights while still containing repeated customer, supplier, or bottleneck dependencies across its underlying companies?
An ETF can spread capital across many securities while several holdings still depend on the same customers, suppliers, foundries, or bottlenecks, creating a second layer of concentration beneath fund weights.
Is Supply-Chain Investing Only Useful for Semiconductor Stocks?
Network path
Does supply-chain investing provide useful information outside semiconductors, and what kinds of cross-industry relationships can change an investment decision?
Battery, automotive, enterprise technology, telecom, retail, and cloud relationships show that dependency-aware investing is a way of analyzing companies across industries rather than a semiconductor-only strategy.
Limiting Portfolio Dependence on One Company
Shared counterparty
How can an investor control repeated portfolio dependence on one outside customer, supplier, foundry, or infrastructure company without forcing unlike relationship metrics into one false exposure score?
Treat repeated customers, suppliers, and other outside nodes as portfolio exposures that can be deliberately limited even when the company creating the concentration is not directly owned.
More Suppliers Do Not Always Make a Company Safer
Supplier concentration
Does a large supplier network make a company more resilient, or can a business still contain critical single-source and limited-source bottlenecks despite having many visible suppliers?
Supplier count can look reassuring while critical inputs still depend on a small number of qualified sources, so real resilience depends on replaceability rather than the length of the vendor list.
Network Centrality as a Supply-Chain Quant Factor
Network path
How can a quant test whether supply-chain network centrality contains information about risk, price efficiency, or returns beyond firm size, industry structure, and uneven relationship-data coverage?
Degree, weighted degree, betweenness, and other graph measures can encode economically interesting network position, but they need point-in-time construction and controls for company size, industry, liquidity, analyst attention, and data coverage.
Network Dependence Reduces Effective Sample Size
Network path
How should a quant think about effective sample size, standard errors, and event counts when supply-chain relationships make cross-sectional stock observations statistically dependent?
Connected companies can share customers, suppliers, industries, and shocks, so the row count in a cross-sectional network model can materially exceed the amount of independent statistical evidence.
One-Customer Growth vs. Ecosystem Growth
Customer concentration
How can an investor distinguish supplier growth driven mainly by one large customer from broader growth supported by several meaningful and economically independent demand paths?
Compare the economic weight of the largest customers, how those weights change, and whether the customers share the same underlying demand cycle. Headline revenue growth can hide a supplier becoming more concentrated underneath.
Portfolio Catalyst-Calendar Diversification
Shared counterparty
Can an investor diversify not only securities and sectors but also the external customer and supplier earnings events capable of moving several holdings at once?
A portfolio can spread external event risk by avoiding too many holdings whose important customers or suppliers report at the same time. Supply-chain data reveals those non-owned catalysts before they appear in the holdings list.
Portfolio Demand-Shock vs. Supply-Shock Risk
Shared counterparty
How can an investor separate portfolio vulnerability to customer-demand shocks from vulnerability to supplier availability, cost, and production shocks?
Map customer demand shocks separately from supplier cost and availability shocks to reveal how portfolio holdings can fail for different reasons.
Portfolio Supply-Chain Exposure Beyond 13F Holdings
Shared counterparty
What economic dependency information can supply-chain data reveal that cannot be observed from an institutional Form 13F holdings report alone?
A Form 13F shows reportable securities held by an institutional manager. Supply-chain data shows the outside companies those holdings depend on, so a portfolio can have major Nvidia exposure without owning Nvidia stock.
SK hynix Has Greater Nvidia Exposure Than Micron
Shared counterparty
Between SK Hynix and Micron, which memory supplier has the larger economic relationship with Nvidia across supplier revenue share, relationship size, and Nvidia customer cost share?
Nvidia represents 27.88% of SK Hynix revenue versus 17.62% of Micron revenue. The SK Hynix relationship is also larger by relationship size and by Nvidia cost share in the supplied data.
Stocks That Are Both Customers and Suppliers
Network path
How should an investor analyze a company that sits in the middle of the supply chain and is economically dependent on both upstream suppliers and downstream customers?
It means the company carries two-sided network exposure. Upstream suppliers can affect its production and costs while downstream customers affect its revenue and demand, so the thesis needs to be researched in both directions.
Supply-Chain Data for Stocks With Little Price History
Network path
Can supply-chain relationships provide useful priors for newly listed shares, ADRs, spin-offs, or newly covered securities before enough price history exists for conventional quantitative features?
A new security can have little market history while the underlying company already has customers, suppliers, products, and economic peers, letting network features provide cold-start context until security-specific evidence accumulates.
Supply-Chain Network Motifs as Quant Features
Network path
Can recurring small supply-chain structures become systematic features that distinguish common demand, common bottlenecks, and customer diversification beyond ordinary degree or centrality measures?
Yes. Repeating local graph shapes such as shared-customer wedges, shared-supplier wedges, fan-in hubs, and fan-out customer structures can encode economic patterns that degree and centrality scores collapse into one number.
Supply-Chain-Aware Covariance Models
Network path
Can supply-chain relationships improve covariance estimation and portfolio risk forecasts by identifying economically connected stock pairs whose correlations may differ from random pairs or rise during shared-node shocks?
Use customer and supplier relationships as structure for covariance shrinkage, residual clustering, and stress scenarios rather than converting relationship percentages directly into correlations.
TSMC Nearly Matches Samsung and Intel in Shin-Etsu Exposure
Customer concentration
How concentrated is Shin-Etsu Chemical's displayed semiconductor customer exposure across TSMC, Samsung Electronics, and Intel, and does TSMC dominate the group?
TSMC represents 4.02% of Shin-Etsu revenue in the supplied data, almost equal to Samsung and Intel combined at 4.22%. The three displayed semiconductor customers sum to 8.24% of Shin-Etsu revenue.
What Does Dependency Analysis Show That Sector, Factor, and Sentiment Data Miss?
Network path
What information does supply-chain dependency analysis add to sector, factor, sentiment, technical, and fundamental views of a stock or portfolio?
Dependency analysis maps the outside companies and commercial relationships capable of changing an investment outcome, adding a different layer of awareness from sector labels, factor exposures, sentiment, and price behavior.
Portfolio Exposure to TSMC
Network path
How can an investor measure direct and indirect portfolio dependence on TSMC without confusing TSMC stock weight, company relationships, second-order exposure, and Taiwan geographic risk?
Measure TSMC dependence beyond direct TSM stock weight by separating portfolio holdings with direct relationships, second-order paths, structural exposure, and Taiwan-specific geographic risk.
How to Find Shared Customers Across Portfolio Holdings
Shared counterparty
How can investors find hidden portfolio concentration created by shared customers?
Find hidden demand overlap by identifying external customers connected to multiple holdings, then separate structural overlap from quantified exposure.
Adding a Stock Can Reduce Portfolio Diversification
Shared counterparty
How can an investor determine whether a candidate stock adds a genuinely new economic path to the portfolio or simply adds another position tied to customers, suppliers, and bottlenecks already owned?
Test the next stock by the economic paths it adds to the portfolio, not just its ticker or sector, because a new position can reduce issuer risk while increasing shared customer or bottleneck exposure.
Customer Exposure Revenue Scenarios
Customer concentration
How can investors combine multiple customer revenue exposures into a first-order supplier revenue sensitivity scenario without calling the result a forecast?
Combine multiple customer revenue shares into explicit demand scenarios so customer-mix risk becomes measurable without turning a linear sensitivity calculation into a forecast.
Direct and Indirect Customer Double Counting
Network path
How can investors avoid counting the same end demand twice when a supply-chain graph contains both direct buyers and indirect end customers?
Classify legal buyers, channel partners, and end customers before adding customer exposure so the same underlying demand is not counted more than once.
Researching Anonymous Major Customers in 10-K Filings
Customer concentration
How can investors research unnamed major customers in a filing without falsely assigning the disclosed concentration percentage to a named company?
Use named supplier-customer relationships to investigate anonymous customer-concentration disclosures without falsely assigning a filing percentage to a specific company.
Customer Contracts and Supply-Chain Demand Visibility
Customer concentration
How should investors combine customer concentration with long-term volume commitments, pricing terms, and contract durability?
Combine economic customer concentration with multi-year volume commitments, pricing structure, deposits, and cancellation terms to distinguish visible demand from flexible or spot demand.
Earnings Surprise Propagation Across Supply Chains
Network path
How can investors trace an earnings surprise through direct and second-order supply-chain relationships without turning the network into an automatic forecast?
Use direct and second-order relationship exposure plus reporting sequence to decide where an earnings or guidance surprise deserves follow-up research.
HPE's Four-Customer Concentration
Customer concentration
What do HPE's displayed customer relationships reveal about the difference between customer-industry diversity and the economic concentration of the events an investor should monitor?
Four displayed HPE customer relationships span very different industries, but Microsoft represents about 59.84% of their combined USD relationship size, showing why customer breadth and economic-weight diversity are different.
Customer Concentration Can Strengthen an Investment Thesis
Customer concentration
When can a large customer relationship improve demand visibility and strategic confidence enough to strengthen an investment thesis rather than function only as a concentration risk?
A large customer can create fragility, but it can also improve demand visibility when the relationship is strategically deep, expanding, and tied to a real product roadmap.
Hidden Supply-Chain Concentration in Stock Portfolios
Shared counterparty
An investor-focused framework and semiconductor case study for finding shared suppliers, customers, internal edges, and hidden portfolio dependencies.
Relative-Value Screens Built Around a Shared Customer
Shared counterparty
How can investors use shared-customer exposure to create a relative-value research pair without turning relationship percentages into a trading hedge ratio?
Use common customer exposure to choose economically comparable suppliers, then investigate why their fundamentals or prices diverge without using relationship percentages as trading hedge ratios.
Customer Concentration Is Not the Same as End-Market Concentration
Customer concentration
How can one large customer relationship still contain multiple end-market exposures, and how should investors separate customer concentration from product and end-market diversification?
Use the LG Energy Solution to Tesla relationship to separate customer concentration from EV, energy-storage, product-mix, and manufacturing diversification.
Economic Baskets Built From Supply-Chain Exposure
Customer concentration
How can investors build a basket around measurable supply-chain exposure instead of a broad thematic label?
Construct baskets around measurable customer or supplier relationships and weight members by economic exposure instead of using only themes or market capitalization.
Diversification Can Fail When Volatility Spikes
Shared counterparty
How can shared customer, supplier, or bottleneck dependencies make otherwise diversified holdings behave like one event cluster during a shock even when normal historical correlation looked low?
Several holdings can become one event cluster when a shared customer, supplier, or bottleneck becomes the source of new information, even when historical correlation previously looked low.
Custom Stock Indexes With Supply-Chain Risk Limits
Shared counterparty
How can investors cap shared customer, supplier, and network exposure when constructing a custom stock index instead of relying only on sector and country limits?
Add customer, supplier, and network exposure limits to conventional custom-index rules so sector, country, and position caps do not hide economic dependence between constituents.
Customer Concentration vs. Credit Risk
Customer concentration
How can investors separate customer revenue dependence from accounts-receivable credit concentration using supply-chain relationships and financial statements?
Separate revenue dependence from accounts-receivable concentration so a major customer relationship is not automatically treated as the same level of collection risk.
Supplier-Customer Pairs Do Not Guarantee Portfolio Diversification
Network path
When an investor owns both a supplier and its customer, how much independent diversification does the second position really add when both holdings sit on the same commercial relationship?
Two different stocks can still sit on the same economic path, so a supplier-customer pair should be treated as intentional overlap rather than assumed to be fully independent diversification.
Tracking Anonymous Customer Concentration Across Filings
Customer concentration
Why can anonymous customer labels not be treated as stable identities across reporting periods, and what evidence burden makes a defensible historical concentration analysis difficult?
Explain why anonymous customer labels cannot be treated as permanent identities and why defensible historical analysis requires extensive cross-referencing across dated relationship and disclosure evidence.
Which of My Stocks Depend on the Same Supplier?
Shared counterparty
How can investors find portfolio holdings that depend on the same upstream supplier and distinguish a meaningful common dependency from a weak or unquantified network overlap?
Reverse-map the supplier sets of portfolio holdings to find common upstream companies that can create hidden concentration across otherwise unrelated stocks.
Hidden Economic Clusters in Supply-Chain Networks
Network path
How can investors find groups of companies that share the same economic demand cycle even when sector and country classifications make them look diversified?
Map overlapping supplier, customer, manufacturing, server, and cloud relationships to find economic communities that conventional sector and country classifications can miss.
The Most Important Outside Company in a Portfolio
Shared counterparty
How can an investor rank non-owned customers, suppliers, and other network nodes by their economic importance across multiple portfolio holdings?
Rank non-owned customers, suppliers, and other economic nodes by how broadly and materially they connect to portfolio holdings, creating an external-company watchlist that ordinary position reports cannot show.
SK ecoplant's $3.1 Billion Upstream Relationship With SK hynix
Network path
What economically significant relationship sits one step upstream of SK Hynix?
Altsets estimates SK ecoplant's relationship with SK Hynix at 36.49% of supplier revenue, 3.1B USD, and 18.81% of SK Hynix COGS, revealing a material second-degree dependency.
Diversifying Around a Core Stock Without Selling
Network path
How can an investor keep a high-conviction core stock while choosing additional holdings that reduce repeated customer, supplier, and network dependencies?
Keep a high-conviction position while choosing additional holdings that reduce repeated customer, supplier, and network dependencies instead of relying only on sector or return correlation.
Partial HHI for Incomplete Supplier Data
Supplier concentration
How can investors calculate a useful supplier-concentration index when only part of a company's procurement network has quantified cost shares?
Calculate a lower-bound supplier HHI from known customer-cost shares while keeping observed coverage, normalized-subset concentration, and full-company concentration separate.
How to Find Second-Order Supply-Chain Exposure
Network path
How can investors find meaningful second-order supply-chain exposure without overinterpreting network connections?
Expand beyond direct suppliers and customers to identify economically meaningful two-hop paths without turning network proximity into a false exposure score.
Nvidia Customer Concentration: Quanta and Samsung Compared
Customer concentration
Do Nvidia's visible downstream customer relationships show material customer concentration?
Two quantified Nvidia customer relationships show 0.53% of Nvidia revenue tied to Quanta and 0.04% tied to Samsung, illustrating why famous customer names do not prove concentration.
How to Use Supply-Chain Data to Try to Disprove an Investment Thesis
Network path
How can investors use supplier, customer, and network evidence to search for facts that would falsify an investment thesis instead of only confirming it?
Use supplier, customer, and network evidence as a falsification tool by defining what relationship facts would weaken or contradict a stock thesis before searching for confirmation.
Position Sizing for Supply-Chain Dependency Risk
Customer concentration
When two stocks have similar conviction, how can concentrated customer, supplier, or network dependence justify different portfolio weights without turning relationship data into a mechanical sizing formula?
Use concentrated customer, supplier, and network dependencies as an additional portfolio risk input when deciding how much capital to place behind an otherwise attractive stock thesis.
Supply-Chain vs. Sector and Country Diversification
Network path
How can a portfolio look diversified by sector and country while several holdings remain exposed to the same supply-chain demand cycle?
Compare sector and country diversification with supplier, customer, end-market, and network overlap so globally distributed holdings are not assumed to be economically independent.