Supply Chain Fragility — Hidden Chokepoints
Strip the oil-price narrative; map the non-obvious dependencies routing through a single gap
Most macro analysis of geopolitical conflict focuses on one variable: oil price. The supply chain fragility framework argues this is systematically wrong — the structurally more dangerous variables are the ones no one is pricing. In the case of Strait of Hormuz disruption, the non-obvious chokepoints are helium (semiconductor production) and fertilizer (food supply), both routing through the same 21km gap as oil.
Helium cannot be stockpiled — it leaks through containers — making it uniquely vulnerable to supply interruption. South Korea sources 65% of its helium from the Gulf region and manufactures two-thirds of the world's memory chips. A 30% global helium supply cut translates directly to a 30% cut in global semiconductor capacity with a 2-6 month lag before alternative supply normalises. This is a direct threat to AI infrastructure build-out plans that depends on uninterrupted GPU production.
The framework's analytical move is to build a dependency map from a single geographic chokepoint outward, tracing second- and third-order effects: Strait disruption → helium shortage → memory chip production halt → GPU supply constraint → AI infrastructure capex cannot be deployed → AI boom timeline compresses. The energy-GDP correlation (virtually lock-step over 40 years) provides a rough magnitude estimate: losing 20% of global LNG implies a 5-10% fall in gross world product.
- Narrative focuses on the obvious variable (oil); structural risk lives in the non-obvious non-substitutable inputs (helium, fertilizer).
- Stockpilability is a key vulnerability multiplier — inputs that cannot be stored (helium) have zero buffer against supply interruption.
- Geographic concentration compounds fragility — a single 21km strait routing 20-30% of oil, helium, and fertilizer simultaneously creates correlated, not independent, risks.
- Energy consumption and GDP are empirically lock-step over 40 years — energy disruption translates to GDP disruption at roughly equal magnitude.
- Second-order effects (helium → chips → GPUs → AI capex) travel faster than markets reprice them — front-run the dependency chain, not the headline.
- Identify the geographic chokepoint under stressMap which physical infrastructure or geographic bottleneck is the subject of geopolitical risk. In the Hormuz case, the chokepoint is a 21km strait. Characterise the bottleneck by what flows through it, not just the headline commodity. Draw the full commodity flow map from the single chokepoint.Pro tipStart with shipping lane data and pipeline infrastructure maps — most chokepoints route multiple commodity types, not just the one in the news.
- Filter for non-substitutable, non-stockpilable inputsOf all commodities routing through the chokepoint, identify those where (a) no near-term substitute exists and (b) stockpiling is physically impossible or cost-prohibitive. Helium fails both tests simultaneously — no substitute in semiconductor lithography and it leaks through any container. These are the true fragility signals.Pro tipHelium, rare earth elements, and certain industrial gases are the canonical non-stockpilable inputs. Oil is highly stockpilable — strategic reserves exist — which is why it is systematically overweighted as a risk signal.WarningDo not conflate 'no substitute in current production process' with 'no substitute ever' — timeline matters. A 2-6 month shortage is actionable; a 10-year substitute development timeline is not the same risk.
- Trace the second-order production dependency chainFor each non-substitutable input, trace the production dependency chain downstream. Helium → semiconductor lithography → memory chip production → GPU manufacturing → AI data center build-out. Quantify each link: South Korea produces 2/3 of world memory chips and sources 65% of helium from the Gulf.Pro tipUse production concentration data (what % of global X comes from geography Y) to size the magnitude of each dependency link.
- Estimate disruption magnitude and durationUse domain expert data (not general macro models) to estimate the minimum shutdown duration and time to supply normalisation. For helium: Cornblutch's estimate is 2-3 months minimum shutdown, 6 months to normalise. For energy: Keen's 40-year energy-GDP correlation implies a 1:1 ratio — 20% LNG loss maps to roughly 5-10% gross world product decline.Pro tipIndustrial capacity recovery timelines are often longer than market consensus — the Saudi LNG attack destroyed 2 of 14 units with a 5-year rebuild timeline and only 5 companies globally capable of doing it.WarningExpert duration estimates are minimums, not means — model the tails, not just the central case.
- Identify which assets are underpriced for the dependency riskCompare the dependency-chain disruption magnitude to current market pricing of affected assets. If GPU supply constrained for 3-6 months, which AI infrastructure positions are implicitly assuming uninterrupted supply? The gap between dependency-chain risk and current pricing is the signal. For TAO/decentralised compute: GPU shortage increases demand for alternative compute supply — a contradictory signal worth monitoring.Pro tipWatch helium supply news as a leading indicator for semiconductor and AI infrastructure positioning — it will move before GPU price signals do.
30% of global helium supply routes through a Saudi/Iran-straddling gas field via the Strait of Hormuz. South Korea sources 65% of its helium from this region and manufactures two-thirds of the world's memory chips. A Strait disruption triggers a helium shortage; expert Phil Cornblutch estimates a 2-3 month minimum shutdown and 6 months to normalise. With no substitute in semiconductor lithography and no stockpiling possible, global memory chip production drops proportionally. GPU manufacturing, dependent on memory chips, is directly constrained.
A single attack on Saudi Arabian LNG infrastructure destroyed 2 of 14 liquefaction units. One quarter of global liquid natural gas routes through the Strait; the attack destroyed one tenth of that capacity. Only 5 companies on the planet can rebuild the units, with a 5-year timeline. The result: 2.5% of global energy supply gone for 5 years from a single event.
20-30% of global fertilizer supply routes through the Strait of Hormuz, produced via the Haber-Bosch process fed by Gulf gas fields. A Strait closure cuts global fertilizer supply by 20%. Without fertilizer, the planet's sustainable human population drops from 8 billion to 1-2 billion. Keen's 2-3 month timeline: India runs out of fertilizer and triggers a famine cascade.
Keen's supply chain fragility framework emerges from his broader heterodox economics approach: standard models treat resource inputs as substitutable and markets as self-correcting. Keen argues both assumptions fail when physical infrastructure is the constraint. His 40-year energy-consumption-to-GDP chart is the empirical backbone — it demonstrates that energy is not a marginal input but a foundational correlate of economic output.
The helium-semiconductor link was surfaced in the episode via expert Phil Cornblutch's analysis, which Keen incorporated into his geopolitical risk assessment. The framework's value is not in Keen's military predictions (explicitly outside his expertise) but in the dependency-mapping methodology: identify the single physical chokepoint, trace all commodity flows through it, and ask which flows are both non-substitutable and non-stockpilable.