Cross-sector pattern

The Chokepoint Concentration Pattern

Physical and institutional single points of failure appear at exactly the moment dependencies on them peak — and they are owned, not accidental.

↓ .md Take this into your AI — the full analysis + graph as markdown, ready to paste into ChatGPT, Claude, Gemini or any AI.

A chokepoint is the most pervasive structural object in the corpus. It is a single node — a company, a place, a piece of software — that an entire system has become unable to route around. The recurring finding is that these points appear precisely when the dependencies running through them are at their maximum, and that they are almost never accidental. Someone owns them.

The instances

  • Semiconductors. TSMC fabricates the advanced chips the digital economy runs on; ASML is the sole maker of the EUV lithography machines TSMC needs to do it. One company makes the machines, one company makes the chips. TSMC Geopolitical Chokepoint is the single most connected node in the entire corpus — it bridges AI infrastructure, semiconductors, energy, defense, and geopolitics.
  • Energy. The Strait of Hormuz carries roughly a fifth of seaborne oil through a waterway narrow enough to close.
  • Critical minerals. China controls the refining and processing of nearly all the strategic minerals that Western fabs, batteries, and weapons systems depend on — a chokepoint on inputs rather than outputs.
  • Financial data. The Bloomberg Terminal’s three interlocking layers — workflow dependency, data network effects, and analyst training — create switching costs deep enough to be near-inescapable.
  • AI. NVIDIA’s GPU dominance is protected less by raw hardware performance than by the CUDA Fortress — a software ecosystem that makes leaving expensive even when alternatives exist on paper.

What makes a chokepoint, structurally

Three properties recur across every instance:

  1. Irreplaceability in the short run — alternatives exist only on multi-year timescales, if at all (the collapse of would-be alternatives, like a European battery champion, repeatedly confirms the dependency).
  2. Ownership — a chokepoint is leverage, and leverage gets used. The corpus tracks deliberate activation events, such as the weaponization of rare-earth exports, where a latent dependency is switched into a geopolitical instrument.
  3. Recursion — chokepoints nest. Control of a chokepoint in one layer becomes the foundation for a chokepoint in the next, which is how concentration compounds rather than merely accumulates.

Why it matters

A chokepoint is efficient until the day it is a vulnerability, and the transition between those two states is instantaneous. The strategic question the corpus keeps surfacing is not “where are the chokepoints” — those are visible — but “what happens when several of them are stressed at once,” since their owners and their failure modes are more correlated than any single industry view reveals. This pattern is the mechanical core of the Great Simultaneous Concentration.

Hub concepts in the graph

TSMC Geopolitical ChokepointASML EUV MonopolyBloomberg Terminal Three-Layer Lock-inCUDA Fortress

The evidence

Explorations whose graphs this pattern is drawn from.

| 112 nodes · 410 edges

How are global supply chains restructuring post-COVID — nearshoring, friendshoring, and the end of just-in-time

Why Your Stuff Stopped Arriving on Time — and Why Fixing It Is Harder Than It Looks

| 102 nodes · 327 edges

How do carbon markets actually work, and are they effective or just greenwashing at scale

Do Carbon Markets Actually Work, or Are They Just Expensive Theater?

| 138 nodes · 423 edges

How are oil majors (Shell, BP, Exxon, Saudi Aramco) actually positioning for the transition, and who's serious

Why Big Oil Says "We're Going Green" But Keeps Drilling: A Plain-Language Guide

| 117 nodes · 450 edges

How does global monetary policy actually work, and what are the structural fragilities in the system

How the World's Money System Works — and Where It Gets Wobbly

Companies that instantiate it

The other patterns