The Island Castle

Australia is an island, which is to say a castle. Two defence policies are on the table; this memo measures both against what has actually decided sieges, from Troy to the Berlin Airlift.

1. The numbers

• $11 billion spent or committed on the submarine pathway, with no boat in the water.
• $480–650 billion: where comparable builds put the trajectory of the $368 billion headline.
• Eight boats planned; the US transfers require an American president to certify they can be spared.
• 400,000 barrels a day produced, more than 94 per cent of it exported.
• Two refineries left, both running imported crude, covering about a fifth of demand.
• 37 days of reserve against a 90-day obligation unmet since 2012. IEA members average over 140.
• Zero Australian-flagged product tankers; 80–90 foreign cargoes a month.
• 10 days of urea in the 2026 crisis, and 99 per cent of heavy trucks cannot run legally without it.
• Manufacturing down from 14 per cent of GDP in 1995 to under 5 per cent.
• Three sea lanes closed in three years. No navy reopened any of them.

2. The island is a castle

An island nation is a fortress by geography. The ocean is the moat and the wall at once: no army walks here, and an amphibious landing against a continent is the hardest operation in warfare. On the measure that decided sieges for a thousand years — can the enemy get in — Australia is close to unassailable.

That is exactly why an enemy would not try. A besieger who cannot storm the walls does the other thing: shuts the gate and waits. Sieges were won that way far more often than by assault, and the modern version needs no soldiers at all. Missiles and satellites make a sea lane too dangerous to sail; insurers withdraw; shipping lines stop coming. The gate closes without a shot fired at the walls.

A castle is therefore judged by what it holds inside, not by how thick the stone is. The garrison's swords matter least of all. What matters is the well, the granary, the forge and the stores — and whether they run out before the besieging army does.

The full evidence for that mechanism, drawn from three closed seas in three years, is in The Proof of the Hormuz Pudding. This memo takes it as established. The chapters that follow test it against the record — land sieges, island sieges, and the one island that survived a modern blockade — before asking the narrower question: which of the two defence policies on offer builds the castle.

3. Sieges were decided by supply, not walls

The longest siege in British history ran seven years, and the castle held because a stairway ran down the cliff to the sea. Three cases carry the whole mechanism.

Troy: unbreakable walls did not end the siege

The oldest siege story in Western literature runs ten years and the walls never fall. Troy is taken by a trick, not a breach. Two things in it are worth keeping, and the first is that impregnable walls did not end the siege — they simply moved the contest to time and supply.

The second is usually skipped. The Greek army spent those ten years camped on a foreign shore, raiding the surrounding country to feed itself. A besieging force has a supply problem of its own, and most sieges in history ended because the besieger ran out before the besieged did. That is the whole game: two stores, and whichever empties first loses. The account is legend rather than record, and it is used here as the framing all later siege-craft assumes, not as evidence.

Masada: years of stores forced Rome to build a ramp

Masada was a desert plateau with cisterns cut into the rock and storerooms of grain, oil and wine — years of water and food inside the walls. Rome did not starve the defenders out, because it could not. It built a ramp.

The lesson is not that stores make a place safe. It is that stores take the cheap option away from an attacker and force him into the costly one, in full view, for as long as it takes. A place with no stores is taken by waiting, which costs nothing.

Harlech: seven years, held by a sea route

Harlech was the best fortification Edward I built — walls within walls on a sea cliff. It was held against Edward IV from 1461 and surrendered only on 14 August 1468, the longest siege in British history.

It held for one reason. The sea came close to the rock then, and a water-gate with a long flight of steps ran down to the shore, allowing the castle to be resupplied by sea throughout. The garrison also raided outward for supplies and used the castle as a base. Its defence was not its stone. Its defence was a route.

That is the hinge of this memo. When the route is the defence, the siege is decided by whoever controls the route — and if the route runs over water, the castle's fate is settled somewhere else, by someone else's ships.

4. Three islands that lost their sea routes

Three cases, and in all three the walls were never the issue.

Athens: one lost fleet, and the grain stopped

Athens chose this position deliberately. Pericles built the Long Walls to the port at Piraeus so the city could ignore any army on land and live on grain shipped from the Black Sea. For a generation it worked.

Sparta's answer was not to assault the walls. Lysander planned to starve Athens into surrender by cutting the grain route through the Dardanelles. It took one afternoon: at Aegospotami in 405 BC the last Athenian fleet was destroyed almost entire, with only ten of one hundred and eighty ships escaping.

Then the blockade of Piraeus. Through the winter of 405–404 famine gripped the city and the inhabitants ate hides and bark. Athens surrendered in March 404 BC. The walls had never been breached; they were demolished as a term of the surrender. A city that lived by an imported staple lost the moment it lost the route.

Japan: the merchant fleet decided it, not the warships

Japan is the case that should end the argument that a strong fleet is the answer. It went to war over supply and lost the war over supply.

The 1941 asset freeze removed 80 per cent of Japan's oil imports with the stroke of a pen, leaving a navy that owned 88 per cent of the nation's refined stocks burning through a 43-million-barrel reserve at a million barrels a month. Asked afterwards why Japan attacked the United States, the chief of its Naval Affairs Bureau answered: the stoppage of oil imports.

The end came the same way. Japan began the war with about 6.5 million tons of merchant shipping and finished with 1.47 million. Imports of sixteen key materials fell from 20 million tons in 1941 to 2.7 million in the first half of 1945. The mining campaign closed the rest: 12,135 mines, 670 ships sunk or damaged, 35 of 47 essential convoy routes abandoned, and traffic through the Shimonoseki Straits down from 70,000 tons a day in March 1945 to 1,500 tons in August.

The warships were not the ships that mattered. The merchant fleet was.

Malta: a fortress that set its own surrender date

Malta was the most heavily defended rock in the Mediterranean and it very nearly fell to arithmetic. By the end of July 1942 stocks of food, fuel, ammunition and medical supplies were predicted to run out within weeks, and the authorities on the island set a target date for surrender: 7 September 1942.

Operation Pedestal broke through at ruinous cost — only five of fourteen cargo ships survived — landing about 32,000 tons, enough for roughly ten more weeks, with rations still near 1,500 calories a day. The garrison did not save Malta. A convoy did, and a tanker in particular.

5. How an island survived: Britain, 1939–45

Britain is the one island that was besieged by blockade in the modern era and kept fighting. Six things made that possible, and one it never solved.

Food: Britain cut the import task instead of meeting it

Britain imported roughly two-thirds of its food in 1939, about 20 million tons of shipping a year. It never became self-sufficient. It ploughed up pasture, took allotments from 815,000 in 1939 to 1.4 million by 1943 producing some 2 million tons of food, and rationed the rest so it went further. The import task was cut down to what the convoys could actually carry.

Weapons: 44 ordnance factories, begun in 1936

By 1945, 44 Royal Ordnance Factories were operating. They employed about 300,000 workers by 1942 and produced two-thirds of Britain's gun and ammunition output. They were built before the war and sited inland, away from the bombing.

Manufacturing: capacity built in peacetime and dispersed

The shadow factory scheme took military aircraft output from 893 in 1935 to 7,940 by 1939, grew to 175 dispersed sites by 1944, delivered 45 per cent of heavy bombers and 168,040 Merlin engines, and put engine production underground in quarries at Corsham and Drakelow. In the Battle of Britain it lifted fighter output from 446 in June 1940 to 972 a month across July and August. Despite sustained Luftwaffe attacks on the Midlands, dispersal meant the bombing caused no serious interruption to output.

Materials: coal and steel came from under the island

The forges ran on what was under the island. Materials were the one input Britain never had to convoy.

Medicines: Britain invented penicillin and could not make it

Penicillin was discovered and developed in Britain. It was mass-produced in the United States, because British industry was already committed to full wartime output and could not scale it. Invention without manufacturing capacity becomes another country's industry.

Shipping: Britain owned the merchant fleet it escorted

Britain had the largest merchant fleet in the world and the escorts to convoy it. It could lose ships and keep sailing, because the ships were its own and the men crewing them answered to it.

Oil: the one input Britain never solved

Britain produced none. It convoyed, rationed and in the end depended on an ally with vast spare capacity, and the Atlantic came close to deciding the war on that single input.

The counterexample completes the picture. Germany, blockaded but land-bordered, substituted its way through: synthetic fuel made from its own coal, iron ore railed from Sweden, food taken from occupied Europe. The continent could pipe and rail what it could not grow. The island had to import it or make it.

Every British advantage above was built before 1939. A Royal Ordnance Factory cannot be commissioned during a siege, and a pasture cannot be ploughed the week the convoys stop. The decisions that saved Britain were all made in peacetime, which is the only time Australia still has.

6. The supply list, and why an island cannot improvise

A castle held nine things: water from a well or cistern, grain and salt meat, salt itself for preserving, fodder for the horses, firewood and charcoal, a forge with iron and a smith, timber and lime to repair breaches, arrows and powder, and — the one that killed more garrisons than assault ever did — sanitation and medicine.

Each has a modern equivalent, and each can be measured.

Fuel, reserve and shipping figures are set out with sources in Australia’s Fuel Security — Part 1: The Position and the Government’s Response. Pharmaceutical figures are from Australian parliamentary and industry submissions; manufacturing share from ABS national accounts.

Land borders leak; an island has no second route

A land border leaks by design. It is a line thousands of kilometres long crossed by roads, rail, pipelines and people, and nobody can inspect every truck. To blockade a nation with land neighbours you must occupy or coerce all of them, which is why sanctions regimes leak for decades.

To blockade an island you need none of that. You raise the risk until insurers withdraw, and the ships stop coming on their own. Three sea lanes have been closed to commerce in three years by exactly that method, and no navy reopened any of them.

The clearest modern proof is China's answer to its own version of this problem. Faced with dependence on the Strait of Malacca, it did not build a bigger fleet to force the strait open. It built pipelines from Russia and Central Asia, a corridor through Myanmar, rail west, and reserves — and it has fourteen land neighbours no navy can interdict. Australia has none.

The modern era makes the asymmetry worse, not better. Satellites make the ocean surface transparent, cheap missiles put any hull at risk from a thousand kilometres, and war-risk cover withdraws before a shot lands. The sea has become more interdictable over twenty years. Land routes have not.

The largest airlift in history moved 16 days of Australian fuel

The Berlin Airlift is the largest sustained air resupply in history. Over fifteen months it delivered 2,334,374 tons on 278,228 flights, nearly two-thirds of it coal, with daily deliveries rising from 6,729 to 8,893 tons and a record of 12,941 tons in a single day, to sustain a city of two million.

Australia consumes roughly 160 million litres of liquid fuel a day, about 130,000 tonnes. On those figures the entire fifteen-month airlift equals about sixteen days of Australian fuel, and its record day covers about two hours. Fuel is only one item on the list above.

No country's fuel has ever been flown in, and none can be. What arrives, arrives by sea.

7. Fuel decides wars, and Australia does not make it

Water and food set the floor: days without one, weeks without the other. In a modern nation both sit downstream of fuel — food is only food if it can be harvested, freighted and chilled, and water is only water if the pumps and treatment plants run. Fuel is the input that decides whether any of the others move.

Three cases show what happens when it stops, and none of them involve a shortage of equipment.

Germany: fighter production peaked while the fuel ran out

German monthly aviation fuel output fell from 181,000 tons in March 1944 to 53,000 tons by June and 11,000 tons by January 1945. The effect ran past grounded aircraft into the pilots themselves: average training hours fell from 240 in 1942 to 110 in 1944, and new pilots were reaching combat units with under 150 hours against 450 for the average British pilot and 600 for an American.

German fighter production peaked in 1944. They had the aeroplanes. Production is not capability without fuel.

Rommel: full tanks at El Alamein, empty tankers at sea

At El Alamein the German panzer divisions held only 20 per cent of their necessary fuel supplies, and by the morning of 2 September 1942 Rommel's offensive had been paralysed by shortage of petrol. The tanks were present and crewed. The tankers were on the bottom of the Mediterranean, sunk by aircraft and submarines operating from Malta.

Japan: a battleship sent out with one-way fuel

By 1945 the Imperial Navy's ships were fuel-bound. Operation Ten-Go sent the battleship Yamato to Okinawa with barely enough fuel to get there and none to return, and kamikaze attacks displaced conventional tactics in part because a one-way mission needs half the fuel. A navy that cannot afford the trip home has already lost.

Defence holds $61.6 million of fuel, bought on the same imports

Modern war has the same shape with a shorter fuse. Both sides in the current European war strike refineries in preference to armies, because that is the decisive target. In 2026 the United States expended roughly half its Patriot inventory and more than half its THAAD stock in 39 days.

Australia's own position is stark. Fuel is Defence's largest single commodity expenditure, but Defence reported fuel inventories of $61.6 million at 30 June 2022, against $5.3 billion of explosive ordnance and $2.6 billion of general stores — under one per cent of its inventory, and a fraction of one day of national consumption. The national auditor found no mandated requirement in Defence's own manuals to set stock holding policies at all.

More to the point, it is not a separate supply. Defence buys its fuel on commercial contracts, off the same imported cargoes, through the same ports, refined by the same two plants. The ADF's fuel security is the nation's fuel security, and since 2012 the country has lost six of its eight refineries.

8. A military fuel reserve does not survive a hungry country

The obvious answer to all of the above is a military reserve: let Defence hold its own stocks and let the civil economy take its chances. Two cases show why that is not a defence.

In 1918 the German army was still in the field on French soil when the home front broke under blockade. Hunger produced mutiny at Kiel and then revolution, and the state collapsed behind an army that had not been driven out of the territory it held. What ended that war was not the stores at the front.

The modern version took three months. In 2022 Sri Lanka stopped receiving fuel shipments; schools closed, petrol and diesel were rationed for essential services, and inflation reached 54.6 per cent. On 9 July tens of thousands stormed the presidential palace and set fire to the prime minister's residence. Police tear gas failed to disperse them. The president fled the country days later aboard an air force plane.

No invasion, no enemy fleet. A fuel shortage, a hungry population, and a government physically overrun by its own citizens while its security forces stood in front of the buildings.

This is the oldest and ugliest part of siege history. Besieged towns expelled their “useless mouths” to make the stores last, and garrisons that ate while the town starved were dealt with by the townspeople before the enemy arrived. A stocked base inside a starving country is not a fortress. It is the last full tank in a hungry nation, and the people who come for it are Australians.

There is therefore no military-only answer to fuel. National production is the defence plan — a conclusion reached from inside the military's own position, not from outside it.

9. No navy has reopened a closed lane

Three sea lanes have been closed to commerce in three years. In none of them did the presence of a navy restore the traffic. That record is the direct test of the alternative to the castle: a force that operates outside the walls to keep the route open.

The Atlantic: the lane was held, at a price we cannot pay

The Atlantic is the case for the defence, and it should be stated at full strength. Convoying worked. Across the campaign only about one convoy in ten was attacked, and of those about one in ten lost ships; more than 99 per cent of vessels sailing to and from the British Isles arrived.

The price of that result was six years and 3,500 Allied merchant ships totalling 14.5 million gross tons, 175 warships, and some 72,200 naval and merchant seamen dead. It required hundreds of escorts, very-long-range maritime aircraft to close the mid-Atlantic air gap, sustained code-breaking, and an ally that completed 14 million tons of new shipping in 1943 alone. In the spring of 1941 Britain was losing ships fast enough to lose a quarter of its fleet within the year.

Two conditions made it possible, and Australia has neither. Britain owned the merchant fleet it was escorting, and its ally could replace losses faster than they were inflicted. Australia has no merchant fleet to escort and no shipbuilding to replace one, and eight submarines arriving from the 2040s are not an escort force.

The Atlantic is not a counterexample to this memo. It is the price list.

A loaded tanker cannot hide, outrun or absorb a hit

The physics have moved against the ship. A loaded very large crude carrier runs about 330 metres long on a 20-metre draft at some 15 knots. It cannot accelerate away, cannot manoeuvre, cannot be armoured, and cannot be hidden: it is radar-bright, thermally obvious, satellite-visible, and in normal operation it broadcasts its own identity and position continuously.

Against that, the modern surveillance picture is close to complete. Satellites, over-the-horizon radar, high-altitude sensors and drones have made the ocean surface largely transparent, and long-range anti-ship missiles hold any surface vessel at risk from more than a thousand kilometres. This is not 1942, when a darkened ship had a real chance of slipping past. A vessel that runs a contested lane today is found, tracked and targeted, and one hit ends the voyage because the cargo is the hazard.

The economics are worse than the physics. A drone or missile costing tens or hundreds of thousands of dollars threatens a hull and cargo worth hundreds of millions, and the interceptor that stops it costs more than the weapon it stops. No navy wins that exchange for long: in 2026 the United States expended roughly half its Patriot inventory and more than half its THAAD stock in 39 days.

Insurers close sea lanes, and a few hits are enough

The mechanism that closes a sea lane is not attrition. It is underwriting. War-risk cover is not optional in practice: without it a ship cannot be chartered, cannot enter most ports, and cannot satisfy the bank that financed it. When cover is withdrawn or repriced past the value of the voyage, the ship stays alongside, and that decision is taken in an underwriting room rather than at sea.

The arithmetic is public. Additional war-risk premium for a Hormuz transit ran at 0.15 to 0.25 per cent of hull value before the crisis — roughly $150,000 to $250,000 for a $100 million very large crude carrier. By March 2026 quotes ran at 1 to 5 per cent, and in some cases 5 to 10 per cent depending on flag, ownership and routing.

At 5 per cent, insuring that same $100 million tanker costs about $5 million for a single passage; for hulls valued between $200 and $300 million the figure passes $7.5 to $15 million. That is a premium larger than the voyage earns. At that point nothing further needs to be sunk, because the trade has stopped paying.

The market also picks sides without being asked. In 2026 vessels associated with American, British or Israeli interests were quoted around three times the rate applied to other flags — a Greek-flagged carrier near 1.5 per cent where an American-flagged one paid 5 per cent or more, if cover could be obtained at all. An adversary wanting to close a lane to one nation in particular does not need to find that nation's ships. The underwriters will do it.

Hormuz 2026: traffic fell 80 per cent before a mine was laid

Hormuz is the case that proves the mechanism, and the sequence is the part to hold onto.

Within 48 hours of the strikes on Iran on 28 February 2026, war-risk premiums rose fivefold, major marine insurers terminated existing cover and offered replacements at around sixty times pre-crisis rates, and Lloyd's Joint War Committee redesignated the entire Arabian Gulf as a conflict zone. Tanker traffic collapsed by more than 80 per cent.

The mines and the attacks on tankers came afterwards. At least eight vessels were eventually struck, and that deepened and sustained the closure — but the commercial shutdown preceded the physical blockade. The strait was effectively shut before a single mine was laid.

It then stayed shut on its own terms. Analysts concluded that rates would normalise only once sustained incident-free transits had rebuilt actuarial confidence — a process measured in months of quiet, and one no navy can accelerate by being present. United States analysis of the episode noted that threats alone can produce closure-like conditions when shippers judge the cost of transit to exceed the benefit, whether or not every vessel is attacked.

The Red Sea had already run the same experiment at a chokepoint carrying 12 to 15 per cent of world trade. Under a United States carrier strike group and allied escorts, container transits through the Suez Canal fell about 90 per cent between December 2023 and March 2024.

A hundred days after the last attack on a ship, and after the attackers had declared an end to their campaign, Suez traffic was still running around 60 per cent below pre-crisis levels. The decision to sail belongs to owners and insurers, and they are slower to return than a warship is to arrive.

The Black Sea: two navies, neither able to protect a ship

The Black Sea is the clearest demonstration available, because both belligerents want their own trade moving and neither can achieve it.

Ukraine began the war effectively without a navy. Using anti-ship missiles and uncrewed surface drones it has destroyed or seriously damaged around 30 per cent of Russia's Black Sea Fleet, driven the surviving surface ships out of Sevastopol, and then reached them again at Novorossiysk, sinking a minesweeper there in March 2026 and disabling the fleet flagship in April. A fleet that cannot keep itself safe in its own harbour cannot hold a sea open for anyone else.

Turned on commerce, the same tools closed the trade in both directions. Ukraine has struck tankers carrying Russian crude and fuel, with more than forty vessels attacked in the Sea of Azov by mid-July 2026. Russia has struck merchant ships in Ukrainian ports, hitting 28 civilian vessels and killing 21 people between 20 June and 20 July.

The commercial closure followed within weeks. War-risk premiums for a Ukrainian port call rose from about 0.2–0.3 per cent of hull value to around 3.5 per cent, and on 22 July the shipowners suspended arrivals altogether — with the agriculture minister stating plainly that the government had imposed no restriction, the owners had simply stopped coming.

On the Russian side insurers refused Azov cover, Don River grain exports halted, and Novorossiysk went under a night-time ban on vessel movements. The full account is in The Proof of the Hormuz Pudding.

10. Submarines impose blockades; they do not break them

The instrument at the centre of the submarine programme does the opposite of what the programme is sold as doing, and the historical record on that is not ambiguous.

The attack submarine is the premier commerce-destruction weapon

The nuclear attack submarine is the finest commerce-destruction instrument ever built. American submarines destroyed Japan's merchant fleet; German U-boats came closer than anything else to starving Britain out of the war. Both campaigns are covered in the sections above, and both were prosecuted by submarines against merchant shipping.

Submarines do not escort convoys, do not clear minefields, do not restore a port's insurance rating, and cannot make an underwriter write a policy. They close lanes. They are the weapon of the besieger, not of the besieged.

Australia is buying eight of them, for delivery from the 2040s, drawable with a foreign government's consent, as its answer to the risk of being strangled. It is buying the instrument that strangled Japan and calling it a defence against strangulation.

A fleet outside the walls is a sortie, not a substitute for stores

A force that operates outside the walls is a sortie. Sorties are useful, and every well-run castle made them — to raid the siege lines, break up an assault, and buy time.

No castle in history was defended by sorties alone, and no garrison ever sortied instead of filling its granary. That is the finding, and it is a question of order rather than of either-or.

Australia needs a navy, and needs a real one: mines, coastal denial, uncrewed systems, and the capacity to escort what must still arrive by sea. What it cannot do is buy a fleet in place of the stores, the fuel and the forge, and then rely on that fleet to hold a route the evidence says no fleet holds.

The walls come first, because the walls are the only part of a castle that does not depend on anyone else's decision.

11. China can find and hit ships; we cannot outlast a blockade

The People's Liberation Army Navy is the largest navy in the world by hull count, with a battle force of over 370 platforms, and that is the least important number in this section. What decides the contest is the sensor and missile architecture built around it, and the endurance of the country behind it.

Satellites and radar find the ships; missiles reach 4,000 km

China's maritime targeting picture is assembled from several layers: Yaogan-30 electronic and signals intelligence satellites flying in threes, Yaogan-31 synthetic aperture radar satellites imaging through weather, and shore-based over-the-horizon backscatter radar cueing on major surface contacts at ranges beyond 3,000 kilometres.

The shooters behind that picture reach across the theatre. The DF-21D anti-ship ballistic missile is credited with a maximum range around 2,150 kilometres, and the DF-26 with about 4,000 kilometres, described by the United States Department of Defense as capable of precision anti-ship strikes across the Western Pacific, the Indian Ocean and the South China Sea from mainland China. Hypersonic anti-ship weapons sit alongside them.

The operative point is not that these weapons are unstoppable. It is what they do to decisions. A missile of this kind does not need to sink a warship, or even hit one; it needs to create enough doubt that a commander will not place a thirteen-billion-dollar ship and five thousand sailors inside its range. That is the same mechanism as the insurance market, applied to admirals instead of underwriters, and it is why sea lanes close.

Then apply it to the ships this memo is actually about. If that architecture is enough to move a carrier strike group out to arm's length, an unarmed, unarmoured tanker running at fifteen knots with its transponder on is not a difficult problem. Warships can at least shoot back and manoeuvre. The merchant fleet that carries Australia's fuel can do neither.

China can absorb a blockade for months; we last weeks

The strategic logic of Australia's submarines is a distant blockade: attack submarines interdicting Chinese trade at the Malacca, Sunda and Lombok Straits, imposing on China the same strangulation this memo has spent nine sections describing.

China is genuinely exposed there. About 80 per cent of its crude imports transit the Strait of Malacca, and analysts broadly agree that no combination of pipelines and overland corridors replaces that volume in the near term. The point is not that China is invulnerable. It is that China has spent two decades buying time, and Australia has spent the same two decades losing it.

China holds a strategic petroleum reserve variously estimated between 900 million and 1.4 billion barrels, on the order of 80 days of imports or more. It has overland pipelines from Russia, Kazakhstan and Myanmar with a combined capacity near 1.5 million barrels a day.

It has its own domestic production, the largest refining industry on earth, a coal-to-liquids industry, the world's largest electric vehicle fleet displacing petrol demand, rail freight west across Eurasia, and fourteen land neighbours no submarine can reach.

Australia holds about 30 days of diesel, imports roughly 90 per cent of its refined fuel, refines almost none of its own crude, and has no land neighbour at all.

So a blockade of China is a strategy measured in years, uncertain in outcome, and requiring the blockading side to outlast the blockaded one. Australia's contribution to it would be eight boats arriving from the 2040s, drawable with a foreign government's consent. The weapons and sensors pointed the other way are deployed now, and Australia's endurance in the same contest is measured in weeks.

That asymmetry is the whole finding. It is set out at length in The Prize: A Unified AsiaPac, which argues that logistics rather than firepower decides a war between interdependent economies, and in How China Held the World Up, which shows what two decades of built resilience did when Hormuz closed in 2026.

Our fuel comes from refineries inside the war zone

There is a further step that removes the need for anyone to target Australia at all. Australia's refined fuel comes from the refineries of Singapore, South Korea, Japan, Malaysia and China. Those refineries are inside the theatre.

In a regional war they would be prioritising their own nations, their own militaries and their own populations, and no hostile act toward Australia would be required for the cargoes to stop. It has already happened without a war: in March 2026 the Energy Minister confirmed that six fuel cargoes bound for Australia had been turned back or deferred, amid concern that suppliers including Malaysia and South Korea would hold fuel for their own needs.

An island that buys its fuel from the war zone does not need to be besieged. It only needs its suppliers to be busy.

Where this sits in the MMA series

None of the above is new to this programme; it is the conclusion several memos reach by different routes. America First, Australia Loses runs the net assessment of the war plan. The Proof of the Hormuz Pudding assembles the three-seas evidence. The AUKUS Cost Blowout counts the bill and the sovereignty ceded to pay it.

The AsiaPac Predicament sets out the flashpoint impartially, and The AsiaPac Solution proposes the way through it. The whole case is summarised on the Defence pillar page, and the programme that replaces the submarine pathway is set out on the AUASIA page.

12. Energy infrastructure is the first target, and ours is two plants and a pipeline

Everything above assumes the fuel is cut off at sea. The shorter route is to destroy the places that make it, and that is now where modern wars start rather than where they end.

2026: energy for energy, with the world's largest gas field as the target

On 18 March 2026 Israel struck the Asaluyeh processing hub and the South Pars field — the world's largest natural gas field, supplying up to 70 per cent of Iran's gas. The strike damaged about 12 per cent of Iran's total gas production, halted output at two refineries and forced Iran to cut gas supply to Iraq, with knock-on effects on Iraqi electricity.

Iran's answer was symmetrical. It struck Qatar's Ras Laffan Industrial City, which processes all the gas from the North Dome, causing what QatarEnergy called extensive damage; it hit the Haifa refinery complex in Israel; and its state media named Saudi Arabia's SAMREF refinery and Jubail petrochemical complex and the United Arab Emirates' Al Hosn gas field as further targets. The American president publicly threatened to destroy South Pars outright.

The Middle East Council's assessment of that week names the pattern: the conflict has moved into an energy-for-energy escalation dynamic in which critical infrastructure becomes the primary target. It was not the first instance. Israel struck the Fajr Jam refinery and a South Pars processing unit with drones in June 2025, and the United States struck the Kharg Island export hub.

None of those facilities were military. All of them were chosen because they are what a modern country actually runs on.

Gas plants cannot be hidden, hardened or replaced quickly

An LNG or gas processing plant is the easiest strategic target in existence. It is fixed, it is enormous, it sits on the coast, it is visible from orbit, its layout is published in engineering documents, and it is full of pressurised hydrocarbons that do most of the work once a warhead arrives.

It also cannot be repaired at speed. Turbines, compressors and liquefaction trains are built to order by a handful of manufacturers with multi-year queues. A hit on a processing hub is not a week of disruption; it is a structural loss of national capacity for as long as the replacement takes to build and ship.

Australia has already run the experiment twice, by accident

Western Australia's domestic gas comes from essentially two sources: the North West Shelf at Karratha at around 65 per cent, and, before 2008, Varanus Island at around 30 per cent. It travels to Perth down one 1,600-kilometre pipeline.

On 3 June 2008 the wall of a pipe at Varanus Island, corroded from 11 millimetres to 1.5, ruptured and exploded. Western Australia lost 30 per cent of its domestic gas supply for two months. Supply to Pilbara industry fell 45 per cent.

Losses were put at up to $3 billion by the state's own inquiry and $6.7 billion by the Chamber of Commerce and Industry. Full production took more than a year to restore, and the Premier addressed the state on television and radio about household supply.

There was no fallback. The Collie coal-fired station was out with damaged turbine blades and a 120-megawatt plant at Kwinana was down for maintenance, so the state met a gas emergency with its reserve generation already offline.

Victoria had run the same test ten years earlier. The Longford fire of 25 September 1998 killed two workers and left the state without gas for twenty days, with property damage of about US$443 million.

Both were accidents. One was a corroded pipe. Neither involved an enemy, a warhead, or any intent at all.

Two plants and a pipeline is the most brittle arrangement available

Set the 2026 targeting pattern against that structure and the conclusion is arithmetic rather than speculation. A single cruise missile from a submarine two thousand kilometres offshore does not have to be lucky to do what a corroded pipe did, and no Australian gas facility is defended against one.

The same concentration runs through the rest of the system: two refineries, both on the coast, both running imported crude; a handful of fuel import terminals; a transmission network with a small number of critical nodes. A country that has centralised its energy into a few large plants has built a target set, not a system.

This cuts at the AUASIA programme too, and the memo should say so. Corridor landing points, liquefaction trains and substations are targets like any others. The answer is not that the build is invulnerable; it is that dispersal is the only real defence available. Many generation sites, several refineries, distributed storage and redundant corridor routes survive a strike that two plants and a pipeline do not. Concentration is efficient in peacetime and fatal in war.

Which is the argument for not having the war

There is a conclusion in the 2026 record that goes beyond Australia. When both sides can reach each other's energy systems, the exchange has no winner. Iran lost 12 per cent of its gas production and its supply to Iraq; Qatar — a third party to the quarrel — had the hub that processes the North Dome shut and damaged; Israel took a hit on Haifa; and the global price of energy rose for everyone who buys it, which is everyone.

That is the shape of a modern war between developed economies. Not a contest of armies, but a mutual dismantling of the systems that keep populations alive, conducted at missile speed and settled by whoever can still refine, generate and move goods at the end of it.

It is the reason the castle argument does not stop at the walls. Self-reliance is what a country needs if the worst happens. Preventing the worst is worth more than surviving it, and that is the case set out in The Prize: A Unified AsiaPac and The AsiaPac Solution — a region wired together by trade, grid and treaty has no venue for this kind of war. Build the stores because the risk is real. Build the union so the stores are never needed.

13. The eight parts, and where Australia stands

A castle has eight functional parts. Each has a modern national equivalent, and each can be measured.

Fuel, reserve and shipping figures are set out with sources in Australia’s Fuel Security — Part 1: The Position and the Government’s Response; manufacturing share from ABS national accounts.

Seven of the eight are weak or contested. The one that is strong is the one no enemy intends to test.

14. AUKUS, measured against the castle

The submarine pathway is a garrison policy. It buys eight swords, and they are good swords: a nuclear attack submarine is among the most capable platforms in existence, and in a shooting war it would matter.

Three things follow from that, and they are not rhetorical.

The boats arrive in the 2040s, after the decade of risk

The first Australian-built boat is expected in the 2040s. A capability that lands two decades after the decade of danger is not a defence of the present.

The first boats need an American president to release them

The earlier Virginia-class transfers depend on a serving American president certifying the boats can be spared, against a US build rate running near half what the American navy needs. The build problem is examined in Building the Sub to Nowhere. A sword that another household must hand you is not in your armoury.

It strengthens the one part no enemy intends to attack

Submarines do not refine fuel, generate power, make fertiliser or repair a broken turbine. They cannot keep a sea lane open either — three closures in three years, and no navy reopened any of them. The programme spends the nation's largest-ever capital sum on the garrison while the gate, the well, the granary and the forge stay as they are.

There is a fourth effect, and it works against the walls rather than for them. A castle that quarters a foreign army fights that army's wars. The full accounting is in The AUKUS Cost Blowout; the net assessment — that even a war the United States wins is a war Australia loses — is in America First, Australia Loses.

15. AUASIA, measured against the castle

AUASIA is not a diplomatic posture. It is a build, and each part of it maps onto a part of the castle.

The well: fuel and power made here

Fuel made here rather than bought abroad: renewable diesel and aviation fuel from a crop already grown, Australian oil produced and refined in modular plants, gas and renewable power converted to liquid fuel. The build-out is in Australia’s Fuel Security — Part 2: Every Method of Domestic Production. Generation and transmission are the same programme, not a separate one.

The granary: 90 days of stores, held onshore

A real reserve at the 90-day obligation rather than the 50-day ceiling currently funded, held onshore, plus sovereign stocks of the small inputs that stop a country before the fuel does — urea first among them.

The forge: the AUKUS industrial base, redirected

The AUKUS industrial foundation redirected into a sovereign manufacturing hub: 95 per cent sovereign content, 58,000 direct jobs, 400–600 new Tier-2 and Tier-3 businesses on a twenty-year order book, set out in The Sovereign Manufacturing Multiplier and mapped stream by stream in The Subsea Industrial Base. Nothing already committed is wasted; the output mix changes.

The gate: fewer cargoes, carried on our own ships

Cargo that cannot be turned back because it never leaves: fuel refined here, freight moved on an electrified continental corridor, goods made here. What must still arrive by sea travels on a fleet that answers to Australia.

The garrison: coastal defence under Australian command

A real defence of the continent under Australian command — uncrewed coastal systems, mines, missiles and the maritime capability the same yards and trades already build — on the doctrine set out in Defence Through Nation Building.

The neighbours: 32,500 km of cable, both ways

Seven subsea corridors, about 32,500 kilometres, carrying power and data both ways to a starter network of 600 million people. Nations wired to each other's grids do not blockade each other. That is the union in hardware, and the case for it is The Prize: A Unified AsiaPac, with the physical programme in The Asia-Pacific Subsea Corridor Network.

The last one is the part a fleet can never buy. Walls decide whether a siege succeeds; neighbours decide whether one is laid.

16. The two policies, side by side

Both columns describe the stated programmes, not worst cases. AUKUS delivers what it promises, on time, and the assessment above still holds.

17. A foreign command is being established on Australian soil

One part of the castle is changing now, and it is the garrison.

HMAS Stirling, on Garden Island off Rockingham, is being expanded at a cost of up to $8 billion to host Submarine Rotational Force – West: from as early as 2027, one British and up to four American nuclear submarines operating from Australian soil. In May 2026 the United States established Naval Support Activity Stirling, a US Navy shore command on the base, and reactivated Submarine Squadron 3 to run the boats.

The US Navy expects about 2,300 Americans in Western Australia by 2030 — sailors, civilian staff, contractors and their families. HMAS Stirling's own Australian complement is roughly 2,300 service members. The foreign presence is being sized to match the Australian one at the country's largest naval base.

Australia's stated protection is the doctrine of full knowledge and concurrence, paired with an undertaking to respect the American policy of neither confirming nor denying what the boats carry. Both cannot be fully honoured. A country that has agreed not to say what is inside the walls cannot demonstrate that it knew, and a parliament that cannot be told cannot concur on the public's behalf.

Distance from the mainland does not protect the base

The usual answer is that Perth is too far away to matter. It is roughly 7,000 kilometres from the Chinese mainland, well beyond the reach of the anti-ship and theatre missiles described in the sections above. On land-based ranges alone, that is true.

It stops being true the moment the launcher moves. China's Type 093B is a guided-missile nuclear attack submarine of about 6,200 tonnes carrying 12 to 24 vertical launch cells, armed with YJ-18 anti-ship missiles and DF-10 land-attack cruise missiles reaching about 2,000 kilometres with a 500-kilogram warhead, guided by satellite, inertial and terrain-following systems. Around 16 hulls are assessed to exist, built at roughly two boats a year since 2022, using pump-jet propulsion rather than an exposed propeller to run quieter.

The United States Department of Defense states the consequence directly: the Chinese navy will soon have the ability to conduct long-range precision strikes against land targets from its submarines and surface combatants using land-attack cruise missiles. The submarine force stands at about 65 hulls, with nine Jin-class ballistic missile submarines carrying JL-2 and JL-3 above that, and China's longest-range land-based missiles — the DF-41 at 12,000 to 15,000 kilometres and the DF-61 at about 12,100 — already cover every point in Australia from home territory.

So a single boat a couple of thousand kilometres off the West Australian coast holds Stirling, the fuel terminals, the North West Shelf and any future corridor landing point at risk, with conventional weapons, no strategic system and no warning. Finding that boat in the Indian Ocean is the hardest task in naval warfare, and it is not a task Australia's own submarines would be performing if they were operating with the American fleet in the South China Sea.

There is no getting around the symmetry. HMAS Stirling exists to host submarines precisely because a submarine can reach a long way from home. The physics does not care which direction it travels.

In castle terms this is the oldest mistake in the manual. A garrison that answers elsewhere brings its own enemies to the gate, and the household that quartered it discovers the siege was never about them.

18. The money is spent either way

The comparison is not between spending and saving. Australia is committed to roughly $2.7–4.4 trillion over twenty years on the current path across water, energy, transport and defence — the counterfactual costed in Without MMA. The programme that replaces it returns an estimated $264–403 billion a year in defensible monetised value at maturity, with payback in three to four years of mature operation, set out in MMA Return on Investment.

The submarine line is $368 billion to the mid-2050s on the current estimate, and $480–650 billion on the escalation rates of the two most comparable programmes — the British Astute class at about 30 per cent over, and the French Attack class, which Australia itself cancelled after its price had climbed from roughly $50 billion toward $90 billion before a hull was laid down.

The submarines are second-hand, and part of their life is already spent

The boats are not new. All three Virginia-class submarines are now to be transferred second-hand, the new-build option having been dropped in 2026. They are Block IV hulls, first commissioned from 2020 with the last around 2026, to be handed over in 2032, 2035 and 2038 — six to fifteen years of a 33-year design life already consumed before an Australian crew boards one.

The reactor cannot be refuelled. When the core is exhausted the hull is retired, and core life is measured in effective full power hours rather than calendar years, so a boat worked hard by its first owner arrives with less left than its age suggests.

The Australian Submarine Agency told Senate estimates that each will arrive with more than twenty years of life remaining. Independent analysts question that, arguing the United States Navy will not part with premium Block IV hulls while short of its own attack submarines, and that older Block III boats commissioned between 2014 and 2020 are the likelier transfer — which would leave roughly fifteen years.

Either figure makes the same point. This is not a forty-year national asset. It is a partly used platform, arriving in the 2030s, retiring in the 2040s or 2050s, from a seller whose own shortfall is the reason the terms changed. The engineering case against the boat itself — a first-of-class design for a shipyard that does not yet exist, from partners who cannot meet their own build rates — is set out in Building the Sub to Nowhere.

What the money leaves behind is the difference

A refinery, a corridor, a generating station or a factory built with the same money is still working in eighty years. It earns every year in peacetime, and it is the same asset that holds the country up in a siege.

Nothing in the AUKUS bill is available to a hospital, a farm or a freight line on an ordinary Tuesday.

19. The verdict

Measured against the eight parts of a castle, the submarine pathway improves one, weakens two and leaves five untouched, at the largest capital commitment in Australian history. The build improves seven and leaves the eighth — the walls — as strong as geography already made it.

The evidence behind that is not a preference. It is the record. Athens lost with its walls intact. Japan lost with a first-rate navy. Malta set a date for its own surrender while it held the strongest fortifications in the Mediterranean. Britain survived because it made its weapons, grew what food it could, sat on its own coal and owned the ships it escorted — and every one of those advantages was in place before 1939.

The modern additions all run the same way. Three sea lanes have closed in three years and no navy reopened any of them. Hormuz shut on an underwriter's decision before a mine was laid. Germany had the aircraft and no fuel. Sri Lanka's government was overrun by its own citizens three months after the fuel stopped.

Against that record Australia holds about 30 days of diesel, imports nine litres in ten of its refined fuel, owns no tanker, makes almost none of its own medicine or ammunition, and is standing up a foreign command inside its largest naval base.

None of this is an argument that submarines are bad. They are superb, and in a shooting war they would matter. It is an argument about sequence. Stores, a well and a forge keep a castle alive for a year, and they are the part no other government can withhold. Eight excellent swords, arriving in the 2040s and drawable with another household's permission, do not feed anyone.

The decision in front of the country is not whether to spend. The money is committed either way. It is whether that money buys eight hulls that arrive after the decade of risk, or the fuel, power, industry and stores that make the risk survivable — and the regional relationships that make the siege unlikely in the first place.

Build the walls so that a siege cannot succeed. Build the friendships so that one is never laid.

20. Sources

1. AUKUS programme cost and schedule (§1, §14) — Australian Government AUKUS optimal pathway announcements and Defence portfolio statements, 2023–2026: $368 billion to the mid-2050s including a 50 per cent contingency; approximately $11 billion spent or committed; first Australian-built SSN-AUKUS expected in the 2040s; Virginia-class transfers conditional on a United States presidential certification that the transfer will not degrade American undersea capability. The $480–650 billion trajectory and its precedents are set out in The AUKUS Cost Blowout — Extended.
2. Fuel dependence, reserves and shipping (§1, §6, §13) — Australian Petroleum Statistics (DCCEEW); IEA oil-stock data (approximately 37 days held on the memo's basis against the 90-day obligation, unmet since 2012; member average above 140 days — note that IEA reporting including stock afloat has been cited at up to 61 days, and the two figures use different bases); EIA Australia Country Analysis 2025 (production near 400,000 b/d, over 94 per cent exported); Australian General Shipping Register (no Australian-flagged refined-product tankers); 98 per cent of imports by volume arriving by sea. Full treatment in Australia’s Fuel Security — Part 1: The Position and the Government’s Response.
3. Troy and the besieger's supply problem (§3) — the Homeric tradition, used as the framing later siege-craft assumes rather than as historical record; the ten-year encampment and the Greek foraging of the Troad are features of the tradition itself.
4. Masada (§3) — Josephus, The Jewish War, and site archaeology: rock-cut cisterns and storerooms of grain, oil and wine sufficient for years, and the Roman siege ramp built because starvation was not available as a method.
5. Harlech (§3) — the siege of 1461 to the surrender of 14 August 1468, the longest in British history; the castle's water-gate and stairway to the former shoreline permitted resupply by sea throughout, and the garrison raided outward for supplies (Cadw; English Heritage; standard Wars of the Roses accounts).
6. Athens (§4) — Xenophon, Hellenica, and standard accounts: the Long Walls to Piraeus and Black Sea grain dependence; Lysander's plan to cut the grain route through the Dardanelles; Aegospotami, 405 BC, with ten of one hundred and eighty ships escaping; the blockade of Piraeus and the famine of the winter of 405–404; surrender in March 404 BC and demolition of the walls as a term.
7. Japan (§4, §7) — the July 1941 asset freeze removing about 80 per cent of Japan's oil imports; the navy holding 88 per cent of refined stocks and consuming about one million barrels a month against a 43-million-barrel reserve (S. L. Wolborsky, Choke Hold: The Attack on Japanese Oil in World War II, and standard accounts). Merchant shipping from about 6.5 million tons to 1.47 million tons; imports of sixteen key materials from 20 million tons in 1941 to 2.7 million in the first half of 1945; Operation Starvation, 12,135 mines, 670 ships sunk or damaged, 35 of 47 essential convoy routes abandoned, Shimonoseki Straits traffic from 70,000 tons a day in March 1945 to 1,500 tons in August; Operation Ten-Go and the one-way fuelling of Yamato.
8. Malta (§4) — stocks predicted to run out within weeks by the end of July 1942 and a surrender “target date” of 7 September 1942; Operation Pedestal, five of fourteen cargo ships surviving, approximately 32,000 tons landed giving about ten further weeks, rations near 1,500 calories a day (Imperial War Museums; contemporary operational accounts).
9. Britain (§5) — roughly two-thirds of food imported in 1939, about 20 million tons of shipping a year; allotments from 815,000 in 1939 to 1.4 million by 1943 producing about 2 million tons of food. Royal Ordnance Factories: 44 operating at the end of the war, about 300,000 workers by 1942, two-thirds of British gun and ammunition output. Shadow factories: military aircraft output from 893 in 1935 to 7,940 in 1939; 175 dispersed sites by 1944; 45 per cent of heavy bombers; 168,040 Merlin engines; underground production at Corsham and Drakelow; fighter output from 446 in June 1940 to 972 monthly across July and August 1940. Penicillin developed at Oxford and mass-produced in the United States because British industry was fully committed.
10. Germany's substitution and its collapse (§5, §7) — synthetic fuel from domestic coal, Swedish iron ore by rail, food from occupied Europe; then the oil campaign: monthly aviation fuel output from 181,000 tons in March 1944 to 53,000 tons in June and 11,000 tons by January 1945, with pilot training hours falling from 240 in 1942 to 110 in 1944 and replacement pilots reaching units with under 150 hours (United States Strategic Bombing Survey and derived accounts).
11. Rommel and El Alamein (§7) — panzer divisions holding only 20 per cent of necessary fuel supplies; the offensive paralysed by petrol shortage by the morning of 2 September 1942, following the interdiction of Axis tankers from Malta (British naval staff history of the Malta convoys).
12. Pharmaceutical dependence (§6) — Australia importing about 90 per cent of its medicines, with almost no capacity to manufacture active pharmaceutical ingredients for most items on the World Health Organization essential medicines list (Institute for Integrated Economic Research – Australia, 2020; submissions to the Australian Parliament; Menzies Research Centre, 2024).
13. Manufacturing share of GDP (§6, §13) — ABS national accounts: manufacturing about 14 per cent of GDP in 1995 and below 5 per cent in the current series; productivity context in Productive Growth at Continental Scale.
14. The Berlin Airlift (§6) — 2,334,374 tons delivered on 278,228 flights over fifteen months, nearly two-thirds coal; daily deliveries rising from 6,729 to 8,893 tons; a single-day record of 12,941 tons on 15–16 April 1949 (United States Air Force and Royal Air Force records; National Cold War Exhibition). The comparison with Australian consumption is arithmetic on approximately 160 million litres a day at an average density near 0.8 kg per litre.
15. Sea-lane closures (§6) — the 2026 Strait of Hormuz closure, the Red Sea 2023–25 and the Black Sea 2025–26; evidence that naval presence did not restore commercial traffic in any of the three is assembled in The Proof of the Hormuz Pudding.
16. Munitions depth (§7) — United States expenditure of approximately half its Patriot inventory and more than half its THAAD stock over 39 days in 2026 (Center for Strategic and International Studies analysis, reported April 2026).
17. Defence fuel holdings (§7) — Australian National Audit Office: fuel as Defence's largest single commodity expenditure; Defence inventories at 30 June 2022 of $7.9 billion comprising explosive ordnance $5.3 billion, general stores $2.6 billion and fuel $61.6 million; no mandated requirement in Defence manuals to develop stock holding policies. Australian Strategic Policy Institute, Fuel under fire (2024): six of eight sovereign refineries lost since 2012.
18. Germany 1918 and Sri Lanka 2022 (§8) — the collapse of the German home front under blockade, the Kiel mutiny and the November revolution while the army held ground in France; and Sri Lanka's 2022 fuel crisis, with shipments halted, schools closed, fuel rationed to essential services, inflation at 54.6 per cent, the storming of the presidential palace and the prime minister's residence on 9 July 2022 and the president's flight from the country days later (contemporaneous reporting).
19. HMAS Stirling and SRF-West (§17) — Australian Submarine Agency and Defence: works valued at up to $8 billion; Submarine Rotational Force – West from as early as 2027, one United Kingdom Astute-class and up to four United States Virginia-class submarines. United States Navy: Naval Support Activity Stirling established 30 May 2026 under Navy Region Japan; Submarine Squadron 3 reactivated, having been decommissioned at Pearl Harbor in 2012; Pearl Harbor Naval Shipyard detachment providing intermediate-level maintenance from mid-2026. The 2,300 figure is a United States Navy statement of expected American personnel, civilian staff, contractors and family members in Western Australia by 2030, reported August 2026; HMAS Stirling's Australian complement of approximately 2,300 service members is from base data reported June 2026.
20. Full knowledge and concurrence (§17) — Australian Government statements applying the doctrine to visiting nuclear-powered vessels, together with Australia's undertaking to respect the United States policy of neither confirming nor denying the presence of nuclear weapons; analysis of the resulting verification gap, Nautilus Institute (2026).
21. The AUASIA programme components (§15) — Australia’s Fuel Security — Part 2: Every Method of Domestic Production; The Sovereign Manufacturing Multiplier (95 per cent sovereign content, 58,000 direct jobs, 400–600 new Tier-2 and Tier-3 businesses); The Subsea Industrial Base; The Asia-Pacific Subsea Corridor Network (seven corridors, approximately 32,500 km, a starter network of about 600 million people); Defence Through Nation Building.
22. Programme economics (§18) — Without MMA (approximately $2.7–4.4 trillion committed over twenty years on the current pathway) and MMA Return on Investment (approximately $264–403 billion a year in directly defensible monetised value at network maturity, payback in three to four years of mature operation).
23. The Battle of the Atlantic (§9) — convoy effectiveness: about one convoy in ten attacked and about one in ten of those losing ships, with more than 99 per cent of vessels sailing to and from the British Isles arriving. Losses: 3,500 Allied merchant ships totalling 14.5 million gross tons, 175 Allied warships and some 72,200 naval and merchant seamen killed, 1939–45; 14 million tons of new shipping completed in 1943 alone; the spring 1941 loss rate that threatened a quarter of the British fleet within a year (Imperial War Museums; United States Naval Institute; Admiralty Battle of the Atlantic figures).
24. Tanker vulnerability and the surveillance picture (§9) — very large crude carrier dimensions and service speed from standard shipping references; the transparency of the ocean surface to satellite, over-the-horizon radar and drone surveillance, and long-range anti-ship missile reach, as set out with sources in The Proof of the Hormuz Pudding. Munitions exchange: United States expenditure of roughly half its Patriot inventory and more than half its THAAD stock over 39 days in 2026 (Center for Strategic and International Studies analysis, reported April 2026).
25. The insurance mechanism and Hormuz 2026 (§9) — additional war-risk premium for a Hormuz transit at 0.15–0.25 per cent of hull value before the crisis, rising to 1–5 per cent and in some quotes 5–10 per cent by March 2026: roughly $150,000–250,000 against about $5 million per passage for a $100 million very large crude carrier, and $7.5–15 million for hulls valued at $200–300 million; flag-based differentiation of roughly three times for American, British and Israeli-linked vessels (Lloyd's List; marine insurance market reporting, March–May 2026). Sequence: within 48 hours of the strikes of 28 February 2026, premiums rose fivefold, insurers terminated existing cover and re-quoted at about sixty times pre-crisis rates, Lloyd's Joint War Committee redesignated the Arabian Gulf a conflict zone, and tanker traffic fell by more than 80 per cent, with at least eight vessels struck subsequently — the commercial closure preceding the physical blockade. Kpler assessment that rates normalise only as sustained incident-free transits rebuild actuarial confidence; United States analysis that threats alone can produce closure-like conditions.
26. The Red Sea and the persistence of closure (§9) — container transits through the Suez Canal down about 90 per cent between December 2023 and March 2024 despite Operation Prosperity Guardian, and traffic still around 60 per cent below pre-crisis levels 100 days after the final attack and the attackers' declared suspension (World Bank analysis; BIMCO, January 2026).
27. The Black Sea (§9) — Ukrainian anti-ship missile and uncrewed surface drone operations destroying or seriously damaging around 30 per cent of Russia's Black Sea Fleet, forcing its withdrawal from Sevastopol, with a minesweeper sunk at Novorossiysk in March 2026 and the fleet flagship disabled in April; more than forty vessels attacked in the Sea of Azov by mid-July 2026; 28 civilian vessels struck in Ukrainian ports with 21 killed between 20 June and 20 July 2026; war-risk premiums for a Ukrainian port call rising from 0.2–0.3 per cent to about 3.5 per cent of hull value, and shipowners suspending arrivals on 22 July 2026 with no state restriction in place. Assembled with full sourcing in The Proof of the Hormuz Pudding.
28. Submarines as commerce destruction (§9) — United States submarine operations against the Japanese merchant fleet and the German U-boat campaign against Britain, as covered in the Japan and Atlantic sources above; the distinction between attack submarines and convoy escort is one of platform function rather than of contested fact.
29. Chinese naval scale, sensors and anti-ship missiles (§11) — Congressional Research Service, China Naval Modernization: a People's Liberation Army Navy battle force of over 370 platforms, the largest navy in the world by hull count; United States Department of Defense assessment that the DF-26 is capable of precision land-attack and anti-ship strikes in the Western Pacific, the Indian Ocean and the South China Sea from mainland China. DF-21D maximum range reported at about 2,150 km and DF-26 at about 4,000 km; the Yaogan-30 electronic and signals intelligence triplets, Yaogan-31 synthetic aperture radar satellites and shore-based over-the-horizon backscatter radar cueing beyond 3,000 km, with the deterrent effect on carrier operations arising from imposed risk rather than attrition (open-source defence analyses, 2024–2026).
30. China's Malacca exposure and its hedges (§11) — approximately 80 per cent of Chinese crude imports transiting the Strait of Malacca against total imports above 10 million barrels a day; overland pipelines from Russia, Kazakhstan and Myanmar with combined capacity near 1.5 million barrels a day, the China–Myanmar line at roughly 440,000 barrels a day; a strategic petroleum reserve estimated between 900 million and 1.4 billion barrels, on the order of 80 days of imports. Analysts broadly agree that no combination of overland routes replaces Malacca volumes in the near term — the memo's argument is one of relative endurance, not Chinese invulnerability (Foreign Policy, May 2026; The Diplomat, May and August 2026; chokepoint and energy-security analyses, 2025–2026).
31. Regional refinery dependence and cargo diversion (§11) — Australian refined fuel sourced predominantly from Singapore, South Korea, Japan, Malaysia and China; ministerial confirmation in March 2026 that six fuel cargoes bound for Australia had been turned back or deferred, amid concern that supplying nations would retain fuel for their own requirements.
32. Submarine reach against fixed targets (§17) — Type 093B guided-missile nuclear attack submarine: approximately 6,200 tonnes, 12–24 vertical launch cells, YJ-18 anti-ship and DF-10 land-attack cruise missiles (DF-10 about 2,000 km with a 500 kg payload, satellite, inertial and terrain-following guidance), pump-jet propulsion, roughly 16 hulls assessed and about two boats a year launched at Bohai since 2022; United States Department of Defense assessment that the People's Liberation Army Navy will soon be able to conduct long-range precision strikes against land targets from submarines and surface combatants using land-attack cruise missiles; a submarine force of about 65 hulls and nine Type 094/094A Jin-class ballistic missile submarines carrying JL-2 and JL-3. Land-based intercontinental ranges for context: DF-41 reported at 12,000–15,000 km and DF-61 at about 12,100 km, covering all of Australia from mainland China — noting these are strategic nuclear systems, not anti-ship weapons (Congressional Research Service; CSIS Missile Threat; Nuclear Threat Initiative; Pentagon China Military Power reporting; open-source naval analyses, 2024–2026).
33. Energy infrastructure as a primary target, 2026 (§12) — Israeli strikes of 18 March 2026 on the Asaluyeh processing hub and the South Pars field, the world's largest natural gas field and the source of up to 70 per cent of Iran's gas, damaging about 12 per cent of Iranian gas production, halting output at two refineries and cutting Iranian gas supply to Iraq; Iranian retaliation against Qatar's Ras Laffan Industrial City, which processes all North Dome gas, described by QatarEnergy as extensive damage, and against the Haifa refinery complex; Iranian state media naming Saudi Arabia's SAMREF and Jubail and the United Arab Emirates' Al Hosn as further targets; the United States president's public threat against South Pars. Middle East Council on Global Affairs assessment of an “energy-for-energy” escalation dynamic in which critical infrastructure becomes the primary target. Earlier precedent: Israeli drone strikes on the Fajr Jam refinery and a South Pars processing unit, 14 June 2025, and United States strikes on the Kharg Island export hub.
34. Australian gas concentration and the two domestic failures (§12) — Western Australian domestic gas supplied roughly 65 per cent by the North West Shelf at Karratha and roughly 30 per cent by Varanus Island before 2008, delivered south by the 1,600-kilometre Dampier to Bunbury pipeline. Varanus Island, 3 June 2008: a pipe wall corroded from 11 mm to 1.5 mm ruptured and exploded, cutting 30 per cent of the state's domestic gas for two months and Pilbara industrial supply by 45 per cent; losses assessed at up to A$3 billion by the state inquiry and A$6.7 billion by the WA Chamber of Commerce and Industry; full production restored only after more than a year; the Collie coal-fired station and a 120 MW Kwinana plant were simultaneously offline (WA parliamentary inquiry into the Varanus Island incident; ABS; contemporaneous reporting). Longford, Victoria, 25 September 1998: two workers killed, gas supply to Victoria interrupted for twenty days, property damage about US$443 million (Royal Commission into the Esso Longford fire).
35. Second-hand submarine transfer and remaining life (§18) — Virginia-class design life of 33 years with a reactor core that cannot be refuelled, core life measured in effective full power hours rather than calendar years; the June 2026 revision under which all three boats become second-hand Block IV hulls, the planned new-build Block VI/VII boat having been dropped; Block IV first commissioned 2020 with the final hull commissioning around 2026, against transfers scheduled for 2032, 2035 and 2038, giving six to fifteen years of use already consumed; Australian Submarine Agency evidence to Senate estimates that each boat will have more than twenty years of life remaining; independent analysis that the United States Navy is unlikely to release premium Block IV hulls while short of its own attack submarines and that Block III boats commissioned 2014–2020 are the likelier transfer, leaving approximately fifteen to sixteen years (Australian Defence Magazine; The Conversation / University of Western Australia; Pearls and Irritations; United States Navy programme briefings; Congressional Research Service on Virginia-class build rates of about 1.3 boats a year against the 2.33 required).