We map where capital is forced to go — before markets price it.
A causal graph of physical dependencies, supply-chain chokepoints, and forced capital flows — across semiconductors, defense, energy, and critical minerals. When the world changes, it traces who’s exposed and what moves next. A portfolio can hold different companies, sectors, and geographies while depending on the same obscure supplier, processing facility, or critical material. Traditional portfolio systems see siloed companies and industries; Forced Alpha sees the physical dependency network underneath them. Not a screener, not a newsletter, not a black box — a sourced, inspectable dependency map you audit path by path.
5144 nodes and 25158 dependency edges, each traceable to a source document. Primary filings first — SEC, EDINET (Japan), DART (Korea), MOPS (Taiwan), and China customs data. Every edge is graded and cited, from Verified (A) to Structural hypothesis (D), with the source excerpt traveling with the edge. The graph refuses rather than guesses — if no edge is on file, it says so, rather than inventing a link.
Three workflows
The same graph does two symmetric jobs: de-risk the book you hold, and originate the trades you don't. Defense — surface the hidden exposure your holdings secretly share. Offense — surface the overlooked upstream names before the market does. And the connective tissue between them: trace how a physical shock actually moves from a material to a ticker.
01 — X-ray the book: find hidden common-mode exposure — the suppliers, materials, facilities, and processes shared by multiple holdings. My book is NVDA, AVGO, TSM, ASML — show me every hidden chokepoint two or more of them share, graded and sourced.
02 — Trace shocks: follow a shock to where it actually lands, multi-hop, material-level, ending at tickers, not sectors. If China restricts a material, which positions should be investigated first?
03 — Find what to investigate first: surface overlooked chokepoints — smaller public companies whose physical role is disproportionate to their market value, found by the graph, not by hand. Find me a sub-$2B company that sits under the AI buildout — one chokepoint, hard to substitute, that the market hasn’t repriced yet.
Run the whole book, not one shock. Paste a portfolio and the graph surfaces what the holdings secretly share, each dependency graded and sourced. AI-infrastructure book NVDA, AMD, TSM, ASML, AMAT, LRCX: imec, the R&D consortium ASML and TSMC both stand on — Verified (A). Ichor Holdings, fluid-delivery subsystems to ASML, Applied Materials and Lam — Sourced (B). VAT Group, high-vacuum valves to ASML, Applied Materials and Lam — Sourced (B).
Every dependency is typed, directional, and evidence-graded. EVIDENCE grades our proof of the dependency, not the stock: filing-verified means a verbatim quote confirmed in the company’s own SEC filing, sourced means a named source not yet word-for-word verified, asserted means the dependency is identified but not yet tied to a filing. Companies mapped: 5144 across 71 sectors. Supply-chain dependencies with typed relationships: 25158.
Every path is inspectable
Forced Alpha does not present every connection as fact. Each result carries the relationship, the source, the supporting excerpt, and a grade. Verified (A) is directly supported by primary-source evidence, the quote confirmed on the source page. Sourced (B) means a named primary source is on file for this edge, the quote not yet byte-verified against the live document. Reported (C) is supported by secondary or reported evidence — corroborate before acting. Structural hypothesis (D) is a possible transmission mechanism, an investigation lead, not a prediction. The graph refuses on no-edge rather than guessing: "Does Kanto Denka supply Tesla?" returns "No edge on file."
Get the week's hidden-dependency finds.
A weekly note on the common-mode exposures and overlooked chokepoints the graph surfaced — sourced, graded, and dated. No hype. Subscribe — free. Free weekly intelligence. Unsubscribe anytime.
One Material. Three companies.
A worked example: AXT, Lumentum, and Coherent look like three unrelated companies — a substrate maker, an optical-networking supplier, and a photonics-components maker. Trace each one's dependency map to its own SEC filing and all three end at low-purity gallium, which the U.S. Geological Survey reports China produced about 98% of in 2024. China imposed export controls on gallium in July 2023, banned exports to the U.S. in December 2024, and the current suspension of that ban is set to lapse November 27, 2026 — one policy lever reaching all three holdings.
Start with a focused portfolio pilot
Input: one public-equity portfolio, agreed sectors or themes, three disruption scenarios. Output: a hidden common-dependency map, holding-level shock pathways, evidence and grade on every path, overlooked upstream names, and identified graph gaps flagged for verification. Every relationship in the graph is inspectable back to its source document.
Get the week's hidden-dependency finds.
A weekly note on the common-mode exposures and overlooked chokepoints the graph surfaced — sourced, graded, and dated. No hype. Subscribe — free. Unsubscribe anytime.