01 · The wound is legible

Jet fuel needs no explaining.

A rare-earth magnet takes a paragraph to justify. Fuel security takes a sentence: the tankers stop, and within weeks the country stops. Australia holds around 50 days of diesel-and-jet cover — shorthand for how many days of normal demand a stockpile alone could meet — it is now racing to build, refines fuel at just two remaining plants, and depends on sea lanes running past the Malacca and Hormuz straits. New Zealand is starker still — no refinery, one import terminal, a single pipeline feeding Auckland Airport whose 2017 rupture grounded flights nationwide. This is not a modelled risk. As of 28 February 2026, the Strait of Hormuz — roughly a quarter of the world's seaborne oil — has been effectively closed, and Australia has responded with a A$3.2 billion Fuel Security Reserve (Australian dollars; a government fund built to stockpile emergency fuel) inside a fuel-security package of up to A$11.9 billion.

In plain terms

Australia and New Zealand don't make most of their own jet fuel — they ship it in past places a hostile navy could block. Right now, one of those places is blocked. So the government is spending billions filling tanks. The question this page answers: would we be safer spending that money building fuel factories instead?

That live crisis is the exhibit. It also happens to be the cleanest natural test of the comparator below — because it shows us what a government actually does when the strait closes. It does not break ground on synthetic-fuel plants. It buys tanks. The model explains why that is the right first move — and exactly where it stops being enough.

02 · Stock versus flow

The tank wins, and it isn't close.

The comparator prices two ways to survive a disruption of length D days. A stockpile is a stock: buy N days of cover once, deploy it to bridge the gap, then it's dry — cost roughly linear in days. Domestic production is a flow: each 100 kt methanol-to-jet line supplies a fixed ~0.55% of daily demand, forever — so covering the nation's jet flow takes about 179 lines, regardless of how long the disruption lasts. Forcing the two into one unit flatters neither; the model instead minimises annualised cost to hold a coverage target through D days, choosing the cheapest blend.

The decisive variable turns out to be how much of a plant's cost its peacetime fuel sales already cover. Call it commercial credit. At credit zero — a plant built purely for security — the crossover where production finally beats the tank is decades away:

Commercial creditCrossover D*Reading
0.0 · security-funded31.5 yrNever, on any realistic contest
0.5 · half-commercial15.7 yrNever
0.9 · mostly commercial3.1 yrOnly a multi-year contest
1.0 · fully commercial0 yrResilience is free — always build

Exhibit · kata/sovq/comparator.py on locked ledger params · stockpile-carry band moves D* to 20.1–67.4 yr at credit 0 — direction invariant. Falsification (the test that could prove this wrong): the model is a linear corner-solution; it omits rationing, demand-destruction and plant learning-rates. These move magnitudes, not the sign — 179 lines for full flow is arithmetic.

In plain terms

If you build fuel plants only for security, filling tanks is cheaper even through a thirty-year blockade. The maths only flips when the plants already pay for themselves selling clean jet in normal times — then the security comes free. So sovereignty isn't a reason to build; a profitable clean-fuel industry is, and the security rides along.

03 · Where the flow earns its keep

Two places the factory beats the tank.

Production is not useless — it does one thing a stockpile physically cannot: it keeps supplying after the tank runs dry. That tail is worth paying for in exactly two structures, and the model names both.

  1. A commercially self-funding civil industry. Backed by the synthetic-aviation-fuel premium (10–13× fossil jet under emerging mandates), a plant that pays for itself makes resilience a co-benefit you didn't pay extra for. This is the only path to scale. CM-6 finance · CM-9 siting
  2. A small defence-anchored physical core. Anchored here means funded and guaranteed by defence spending, not sold on the open market. Military jet demand is only 1–3% of civil volume — a rounding error — but defence values assured, contested-logistics fuel — supply that keeps moving through blocked or fought-over transport routes — at 100–300× the commodity price. About 3.3 lines ≈ A$644 m/yr covers all defence jet domestically: a policy appropriation, not a market trade, but a defensible one. CM-14 defence
The binding constraint · F-S3
A synthetic-fuel certificate — book-and-claim, meaning a paper claim on green fuel that need not travel with the physical litres — delivers zero litres when the lane closes. Today's airline offtakes (long-term purchase contracts) are certificate-based and delivered overseas: useless for sovereignty. The commercial driver that funds production (the carbon premium) runs on exactly those certificates. So the only coherent sovereign structure is physical delivery plus Guarantee-of-Origin certification, stacked — green-origin paperwork layered on top of the physical barrels, not instead of them — capturing the premium and the molecules. A plant optimised for the certificate is the wrong plant for a blockade. This single design rule is the whole argument.
04 · What the engine outputs

A posture, not a promise.

Inari does not tell a government to build a fuel industry to defend itself; the numbers won't carry it. It outputs an ordering: stockpile the shock (the tank is correct, and it is what governments already do); back a commercially self-funding clean-fuel industry so the flow scales as a co-benefit; fund a small physical defence core on the assured-supply premium; and structure every molecule physical-plus-Guarantee-of-Origin, never book-and-claim. Siting follows curtailment depth against biogenic carbon (carbon from recent plant matter, not fossil sources — the feedstock needed to make synthetic jet fuel) and jet infrastructure — Portland in Victoria, Marsden Point in New Zealand.

In plain terms

Fill the tanks first — that's the cheap, fast safety. Then help a clean-jet industry stand up on its own commercial legs, and the country quietly gains a factory that keeps running when imports stop. Pay for a few plants outright only for the military. And insist the fuel is real barrels made here, not paper certificates.

The one thing the model can't price
A warm production line in a mobilisation is worth more than the litres it makes that year — capability can't be stockpiled, and option-value on a running plant — the worth of simply having it ready to scale up fast, before it's even needed — is genuinely unpriced here. It is the strongest argument for building, and it sits outside the cost model by construction. We flag it rather than smuggle it into the numbers.

燃料主権 — sovereignty measured in days can be bought; capability measured in a warm line cannot. The comparator prices the first honestly and refuses to fake the second.

Discuss the fuel-sovereignty comparator →