Building the Sub to Nowhere

Two fleets in one plan: three second-hand submarines bought from the United States, and five built at home to a new, unproven, three-nation design — enlarged beyond its parents, in a yard that does not yet exist, by a nation that has never built a nuclear submarine, while one partner cannot build the boats it owes and the other cannot keep its own fleet at sea. The engineering and build plan, and everything it asks to go right.

1. What problem is this meant to solve?

Before asking whether the AUKUS submarines can be built, it is worth asking what they are for — and whether they answer the threat that should worry Australia most.

The government's case is deterrence at distance: nuclear-powered boats that can range far from home, stay hidden for months, and hold a major adversary's approaches at risk far out into the Indo-Pacific. Taken on its own terms that is a serious argument, and this memo does not wave it away.

But a submarine — nuclear or conventional, eight of them or eighty — does not put a single litre of fuel in an Australian tank, and that is the country's most immediate strategic weakness. Australia imports around 90 per cent of its refined fuel; it is the only member of the International Energy Agency that fails to hold the 90 days of reserve it is obliged to, keeping only weeks of supply; and the supply line runs through the Strait of Hormuz, the Malacca Strait and the South China Sea — the very waters a conflict would close first. Cutting that supply takes no shot fired at Australia. It takes a closed chokepoint a long way away, and within weeks the trucks, the tractors, the mines and the military itself begin to stop. The companion memo, The True Cost of "Cheap" Imported Fuel, sets this out in full.

And there is a harder point still, about what a submarine can and cannot do. A submarine is a hunter: it can sink an adversary's ships and stay hidden, and in that lies real deterrent power. What it cannot do is escort a convoy, sweep a minefield, or hold a strait open. A submarine cannot protect a sea lane, and it cannot re-open one that has been closed — yet protecting the sea lanes to Australia is the very mission this fleet is sold to perform. The boats are offered as the guarantee of the trade and fuel routes the country lives on, but the platform chosen cannot guarantee a route at all. The dependence they are meant to secure is drawn plainly below: every major economy of the Asia-Pacific lives by the same sea lanes, and no submarine keeps those lanes open.

This is the question that hangs over the whole submarine program, and it is one of priority. Deterrence and fuel security are different problems, and a serious defence policy might want both. But the plan commits the largest sum in the nation's history, over three decades, to a platform that leaves the most acute and immediate vulnerability the country has entirely untouched — a vulnerability the same money, spent on fuel, refining, storage and the infrastructure of self-reliance, could close. A nation can own eight nuclear submarines and still grind to a halt when the tankers stop. Real defence is the capacity to keep functioning, and the submarines do not build an ounce of it. What the money could do instead is counted in the companion memo, The AUKUS Cost Blowout.

The rest of this memo sets that larger question aside and takes the plan on its own ground: can these boats even be built, on this plan — and are they the right bet? The answer, chapter by chapter, is troubling enough on its own. But the reader should keep the prior question in view throughout: even if every risk in these pages were overcome, the submarines would still not have fixed the thing most likely to bring the country to a standstill.

2. We have done this before

Before the plan, one fact should sit in front of the reader, because it changes how everything after it should be read: Australia has built its own submarines once before, and it was one of the hardest and most troubled things the country has ever attempted.

The Collins class — six conventional boats, delivered between 1996 and 2003 — was meant to be a straightforward national success. Instead it ran to roughly three times its contracted cost, arrived more than a decade late to reliability, was branded "dud subs" in the press, needed a foreign expert brought in to rescue it, and is troubled to this day. (Section 15 sets out that record in full.) Out of it came a set of hard-won lessons — the Collins lessons — that the AUKUS plan now proposes to test again, all at once:

• that Australia's procurement disasters begin with rejecting a proven, off-the-shelf option in order to build something bespoke;
• that integrating a complex combat and propulsion system is where the years and the billions disappear;
• that first-of-class schedules slip and budgets blow out, every time;
• that when it goes wrong, the remedy is a foreign rescue and a decade of catching up;
• and that the sustainment bill never stops, long after the boats are built.

And here is the comparison that matters. Collins was one design — a single, settled conventional boat — and it still delivered every one of those problems. The AUKUS plan does not build one settled design. It takes two different nations' submarines, merges them into a third that exists nowhere yet, makes it a third larger than either parent, powers it with a nuclear reactor, builds it in a yard that does not yet exist, and runs a second foreign-built fleet alongside it. Every factor that made Collins hard is present again — and multiplied.

The rest of this memo lays out that plan and everything it asks to go right. But the question underneath all of it is the one Collins already answered once, in cost, delay and years of trouble: we have done a version of this before, and we know what it cost. Do we want to do it again — harder, larger, and nuclear?

3. What the plan actually is

Strip away the announcements and the plan is a specific set of engineering decisions. Australia is to acquire a fleet of about eight nuclear-powered attack submarines (SSNs) in two parts. First, from the early 2030s, it buys three Virginia-class boats from the United States — with an option on two more — as an interim capability. Second, and enduringly, it builds five submarines of a wholly new class, SSN-AUKUS, at Osborne in South Australia.

SSN-AUKUS is not an existing submarine. It is the United Kingdom's next-generation design — the Astute replacement, previously called the Submersible Ship Nuclear Replacement (SSNR) — into which United States technology was inserted when Australia joined the programme in 2023: US propulsion-plant components, a common vertical launch system, common weapons, and a jointly developed combat system. The two navies are to share a single design. It will be powered by a Rolls-Royce PWR3+ reactor — an enhanced variant of the PWR3 built for Britain's Dreadnought ballistic-missile boats — manufactured at the expanded Rolls-Royce site at Raynesway in Derby and shipped to Australia. The reactor core is designed to last the life of the boat. The submarine is expected to displace more than 10,000 tonnes submerged — roughly a third larger than the Astute it is derived from.

The build is split across two hemispheres. The United Kingdom builds up to twelve at Barrow-in-Furness, beginning construction in the late 2020s and entering Royal Navy service in the late 2030s. Australia builds five at a new Submarine Construction Yard at Osborne, run by a joint venture of ASC — which built and maintains the Collins class — and BAE Systems. On the governments' own timeline, Australia is to begin building by the end of the decade and deliver its first Australian-built boat in the early 2040s, then produce one roughly every three years.

The Australian Submarine Agency describes the result in confident terms: a state-of-the-art boat combining the best submarine technology of all three partners, built in a purpose-built yard more modern than Barrow, using a manufacturing system "identical" to Barrow's, with Australia becoming a "fourth production line" in the AUKUS enterprise alongside the two American yards and the British one. That is the plan as its sponsors present it. What follows is what the plan requires to be true.

4. Why a submarine is the hardest thing a nation builds

Before a single nuclear consideration enters, a submarine is already among the most demanding things any nation can manufacture. Very few nations can build one at all, and the reasons are physical, not political.

A submarine is a pressure vessel that must survive being crushed. Its hull has to hold its shape against tonnes of force on every square metre, dive after dive, for decades, with welds that approach perfection because a single bad one is a catastrophic failure at depth. It must also be silent: a warship whose entire purpose is to not be found. That means every pump, every valve, every rotating part is mounted, isolated, and quietened so its vibration does not reach the water — an acoustic engineering problem that runs through the whole boat and cannot be inspected from outside once it is sealed. Into that sealed, curved, cramped volume go tens of thousands of components — propulsion, weapons, sensors, life support, power, control — each of which must work, fit, and not interfere with the others, integrated to tolerances measured in fractions of a millimetre.

The boats are built in ring-shaped hull sections, packed with machinery and then joined — a method that only works if the design is complete and correct before the steel is cut, because rework inside a welded pressure hull is brutally expensive. And the capability lives in people, not just drawings. The welders, the nuclear technicians, the systems integrators, the naval architects take years to train and years of continuous work to keep sharp. Stop building for too long and the skills evaporate — the industry's grim term for it is the "valley of death." A nation that lets its submarine workforce lapse does not simply pause; it has to rebuild the capability almost from the beginning. This is the ordinary difficulty of submarines, the part every builder faces. Australia is proposing to take it on having never built a nuclear submarine, and to do it with a design and a reactor supplied by others.

5. Then you add the reactor

Nuclear propulsion does not add a component to a submarine. It adds a second, harder machine — a power station — inside the first, and a regulatory and safety apparatus around both.

A naval reactor is engineered to run for the life of the boat without refuelling, to operate in near silence, to withstand the shock of combat, and to never fail in a sealed hull hundreds of metres down. Australia will not build these reactors; they are made abroad and shipped out. But operating and maintaining them still requires a nuclear-qualified workforce the country does not have, trained over years, and an entire safety and regulatory system built from nothing. Australia has had to stand up a new independent regulator, the Australian Naval Nuclear Power Safety Regulator, to certify the safety of a technology the nation has never operated.

There is a tail to nuclear ownership that rarely features in the announcements: what happens at the end. A retired nuclear submarine is not scrapped like a ship; its reactor makes it a long-term radiological liability that must be stored, defuelled, and eventually dismantled at specialised facilities. Britain, which has operated nuclear submarines since the 1960s, still has more than twenty retired boats afloat awaiting dismantling — its very first, decommissioned in 1980, among them — with the stored-liability cost running into the billions. Australia is signing up to the front of that commitment as a non-nuclear-weapon state that has never run a nuclear reactor of any kind, and will own the whole of it, cradle to grave.

6. The design is not finished — and it is a three-nation graft

The specific engineering risk of AUKUS is not that submarines are hard. It is that this submarine multiplies every ordinary difficulty by building something that has never existed, in a way that has never been tried.

SSN-AUKUS is a first-of-class boat — and first-of-class boats always surface problems, because no amount of design foresees everything until the thing is in the water. But this is a first-of-class with three complications stacked on top. It is a graft of two nations' engineering: a British hull and design philosophy fused with American propulsion components, launch systems, weapons and combat systems. Merging two mature but different submarine traditions — different standards, different tolerances, different supply chains, different ways of doing the same job — is itself a vast integration task, and the interfaces between them are exactly where first-of-class failures hide.

Put in plainer terms, it is less like modifying one car than like taking a Rolls-Royce and an Oldsmobile — two different makes, built by two different companies to two different standards — and setting out to build a new third car from the best of both. Every major system has to come from the maker who understands it, and then be made to work with a system it was never designed to meet. The size compounds it: SSN-AUKUS is to displace more than 10,000 tonnes submerged — roughly a third larger than the British Astute it is drawn from, and nearly three times the Collins-class boats Australia operates today. A submarine is not scaled like a drawing — enlarge it and the buoyancy, the acoustic signature, the structural loads and the internal arrangement all have to be solved again. And the single design must then satisfy two navies with different requirements.

None of it is proven, because none of it is finished. The United Kingdom does not begin building its first SSN-AUKUS until the late 2020s and does not expect it in Royal Navy service until the late 2030s. Australia is committing to build a class of boats to a design that will still be maturing as the yard rises around it — and, on the current plan, its own build runs in parallel with Britain's rather than following a proven boat into production. The country is not buying a submarine that works. It is buying into the design of one.

It is worth naming the engineering principle the plan set aside. What de-risks a submarine build above all else is a mature, settled design built to a fixed pattern — commonality is the single greatest protection against the delays and cost blowouts every program in this memo records. Merging two different boats into a bespoke third design does the opposite: it maximises the integration risk at exactly the point a first-time builder can least afford it. A single proven design would have been the lower-risk road. This is the Collins lesson exactly: the trouble there began with the choice to build bespoke rather than buy proven, and the plan has made that choice again. The government's case for the harder one is that it wanted the best of all three nations' technology and a sovereign build at the end — real goals, honestly held — but they were bought at the price of designing and building something that has never existed, by a nation attempting it for the first time. Whether that trade was worth making is one more question the country was never squarely asked.

It is worth adding that the world knows how to build submarines across nations — and that AUKUS is not how it is usually done. The established collaborative models are built to avoid exactly this risk. The United States and United Kingdom, building their Columbia- and Dreadnought-class boats right now, share a common missile compartment and common reactor and machinery technology, but each country builds its own hull around that shared middle: the boats are essentially identical in the centre and national at each end. Germany, Norway and Italy are collaborating on conventional boats by all adopting a single proven design, the Type 212, with national modifications — not by blending two. Even the American Virginia class, built by two rival yards at once, is two builders working to one design. The disciplined way to collaborate is to share a proven module or adopt a whole proven design; it is not to fuse two different nations' submarines into a novel third. On their most important boats, in the very same alliance, the US and UK chose the module-sharing road. AUKUS took the one they avoided.

And the graft runs deeper than swapping systems in and out, because the hull itself is new. SSN-AUKUS is to displace more than 10,000 tonnes against the Astute's roughly 7,400 — about a third more — and a third more displacement is not achieved by adding sections to an existing boat. It is a new, larger pressure hull: a different size and structure to survive depth, with the buoyancy, trim, acoustic signature and internal arrangement all resolved afresh. That is what defeats the reassurance that this is merely "modular" — a matter of dropping American systems into a proven British boat. Modularity only de-risks a build when there is a settled, unchanged hull to drop the modules into. Here there is not: the hull is new, the systems are a two-nation graft, the reactor is bought from an overstretched supplier, and the builder has never built a nuclear submarine. Every layer that borrowing a proven design would have made safe has instead been made new at once.

The government is aware of this precedent — and its own words show how partial its answer is. Setting out what he called the critical lessons of the Collins program, the Minister for Defence Industry, Pat Conroy, named the danger of "combining a new hull design with new combat and weapons systems," alongside being a lone "parent" navy and neglecting through-life support. The mitigations he points to are real as far as they go: the SSN-AUKUS combat system is to be an evolved version of the existing AN/BYG-1 rather than a clean-sheet design, the boat is shared with the United Kingdom so Australia is not a lone parent navy, and sustainment, he says, is considered from the outset. But note what the reassurance leaves out. By the minister's own account the hull is new — and to that new hull the plan still adds a co-developed combat system, a vertical launch system that neither the Collins nor the Astute ever carried, American weapons and launch integration, a British reactor new to Australian hands, and a first-ever nuclear build. The single Collins lesson the government claims to have banked — an evolved rather than a wholly new combat system — is one strand of the very combination it warns against. On the hull, the systems integration, the launch system and the reactor, the plan does precisely what its own minister identified as the danger — and at a scale Collins never approached.

7. Two reactors, two fleets

The fleet the plan produces is not one navy's worth of submarines. It is two.

Australia is to operate three American Virginia-class boats and five British-derived SSN-AUKUS boats — two entirely different classes at once. Different hulls, different combat systems, different spare parts, different training pipelines, different maintenance regimes, different supply chains reaching back to two different countries. Every navy that runs submarines works to reduce the number of classes it operates, because each class it adds multiplies the burden of sustaining it. This plan does the opposite, by design, for a fleet of only eight boats.

The sharpest form of it is the reactor. The three Virginia boats are powered by the American S9G reactor, built by the United States' naval nuclear enterprise and fuelled through its supply chain. The five SSN-AUKUS boats are powered by the British Rolls-Royce PWR3+, built in Derby. The two are related — the PWR3 draws on the American design lineage — but they are not the same reactor: a different manufacturer, a different fuel arrangement, a different safety case, a different maintenance regime, a different set of specialist skills. A nation that has never operated a single naval reactor is proposing to stand up the support for two different ones at the same time. There is no commonality to bank. The new regulator must certify two designs; the workforce must be trained on two; the country must sustain two through their whole lives and, at the end, dispose of two.

That is a first-order question of coherence, and it is one the plan's authors have never put plainly to the public: why is a nation operating its first-ever nuclear reactors committing to run two different national types at once — and what does that double burden cost in money, people and time?

8. Building it where it has never been built

Australia has built submarines before — the six Collins-class boats, conventionally powered, at Osborne in the 1990s. It has never built a nuclear one, and the yard to do it does not yet exist.

In February 2026 the government committed about A$3.9 billion as a down payment to begin full-scale construction of a new Submarine Construction Yard at Osborne, with environmental approval granted days later. The yard is to be roughly three times the size of the one planned for the cancelled Attack-class program. Building the yard itself is expected to occupy around 4,000 workers; building the submarines in it, another four to five thousand at peak, with up to a thousand apprentices trained on site each year. The plan is explicitly to replicate the Barrow methodology — British and American experts are already advising — so that Osborne operates to the same production logic as the yard in Cumbria.

Before Australia can build these submarines it has to create almost all of it from nothing, while the boat is still being designed: the yard, the supply chain, a nuclear-qualified workforce, a regulator, and the skilled trades to do the work. It is the position the United Kingdom was in when it began Astute — building from a low industrial base — and Australia's defence analysts have said so directly, warning that with BAE Systems as the contractor in both cases, and Australia starting from a thinner base than Britain had, the country is at risk of repeating the Astute experience unless the workforce problems that wrecked that program are deliberately avoided.

And there is a question folded inside the workforce numbers the plan advertises but does not answer. The headline figures are large and Australian: thousands to build the yard, thousands more to build the boats, up to a thousand apprentices a year. But the expertise to design, integrate, certify and lead a nuclear-submarine build does not yet exist in Australia and cannot be trained overnight. It has to be imported — from BAE Systems and Rolls-Royce in Britain, from the United States naval nuclear program through which Australian officers are now being sent, and, as the Astute rescue showed, potentially from America's General Dynamics Electric Boat, which was brought in to run the British yard when it faltered. The Australian jobs are real, but many are trained into existence across a decade or more, while the high-end skills are foreign at exactly the point they are most needed. The plan publishes the number of Australians it hopes to employ. It does not publish how many foreign specialists it must import to deliver them, in what ratio, or for how long. What that imported expertise has had to do elsewhere, the next two sections show.

9. The United States cannot build the boats it already owes

The interim capability — the three Virginia-class boats Australia is to buy from the early 2030s — rests on an assumption that the United States can supply them. On the current evidence, it cannot.

The US Navy has procured Virginia-class submarines since 1998, and has funded them at two boats a year since 2011. The actual production rate has never reached two. Since 2022 it has fallen to about 1.1 to 1.2 boats a year, held down by workforce and supply-chain constraints, leaving a growing backlog of submarines that Congress has paid for but the yards have not built. To sustain its own fleet and supply Australia at the same time, the Navy says it needs to build about 2.33 a year — roughly double the current rate. The date by which it hopes merely to reach two a year has already slipped, from 2028 to 2032.

The gap is not for want of money. Congress added $1.9 billion to the program in the 2026 appropriations to head off a funding gap that would have forced a stop-work order, on top of years of multi-billion-dollar industrial-base investment — and Australia is itself paying billions to expand American shipyards. The spending has already happened; the boats have not appeared. Meanwhile the US attack-submarine fleet sits at around 46 boats against a stated requirement of 66 — seventeen below the line — with roughly 40 per cent of the fleet at times unavailable because of a maintenance backlog. Each new boat costs about US$5 billion.

The consequence for Australia arrived on 30 May 2026, when the three governments revised the plan so that Australia's first submarines will come from the existing US fleet rather than from new construction — because the new boats meant to backfill the sale do not exist and will not be ready in time. And there is a sting in that revision. The Virginia is designed for a 33-year life on a single reactor core — but a boat drawn from the existing fleet is not new. It arrives with a large share of that life already spent, for a price reported at around five billion US dollars, on the delivery dates now set for 2032, 2035 and 2038. Australia is to take three part-worn submarines, from a fleet already below its own minimum strength, that will need replacing sooner than new hulls would — to bridge to a domestic boat that does not arrive until the early 2040s. That is not a criticism of American shipbuilders; it is arithmetic that the United States' own Congressional Research Service, its Naval Institute and its defence press have documented in detail.

10. The United Kingdom is the cautionary tale — and it is happening now

If the United States is the supplier that cannot keep up, the United Kingdom is the template — and the template is a warning, both in its history and in its present.

The Astute class, the boat SSN-AUKUS is derived from, began as design studies in 1986. The first boat was laid down in 2001, launched in 2007, commissioned in 2010, and not declared fully operational until 2014. In 2009 a House of Commons committee found the first three boats running 57 months late and 53 per cent — about £1.35 billion — over budget. The programme's valuation climbed from a projected £4.3 billion in 2001 to £6.6 billion in 2009, £9.6 billion in 2015, and £12.2 billion by 2025. The fourth boat alone ran some seventeen months late.

The causes matter, because they are the ones that map directly onto Australia. A seventeen-year gap in submarine work at Barrow — a valley of death — had scattered the experienced workforce, and a philosophy of cutting government oversight had reduced the Ministry of Defence's presence at the yard from around fifty people to fewer than half a dozen, leaving no one with a clear view of the trouble building. Design drawings were incomplete when construction began. It was the first British submarine designed entirely in 3D software, and problems with that software added further delay. In the end the Ministry had to bring in America's General Dynamics Electric Boat to rescue the yard's management and workflow — a US firm ran the British submarine program. It is the same foreign rescue the Collins class required, and the same one the Australian build may yet need. The technical results of all this reached the boats: the lead submarine grounded, could not reach its intended top speed, took on water during a dive, and suffered corrosion that a confidential Ministry memo attributed to cost-cutting and called a cause for major concern.

The boat is also the shorter-lived of the two Australia will run: the Astute's reactor core is rated for about 25 years against the Virginia's 33. The design Australia is building from has the shorter life and, on the record above, the more troubled history. (SSN-AUKUS moves to the newer PWR3+ core, which outlasts the Astute's — one of the few places the plan buys itself margin.)

And the present is no reassurance. In June 2026, every one of the Royal Navy's five operational Astute-class submarines was in port at once — an availability rate, for the country's front-line attack fleet, of zero — after defects were found on HMS Anson following a joint exercise with Australia. Britain can perform nuclear submarine maintenance at only one site with the right radiological facilities, so when work backs up, the whole fleet queues. Barrow, meanwhile, is trying to finish the last Astutes, build the Dreadnought ballistic-missile boats, and design SSN-AUKUS all at once, its workforce stretched across three strategic programs. This is the industrial base whose methods Australia intends to replicate, and the partner whose design it intends to build — living, at this moment, through exactly the failure modes the Australian plan would have to avoid.

The same doubt hangs over the one part of the boat Australia cannot build for itself: the reactor. Every SSN-AUKUS reactor — for the Royal Navy's boats and Australia's alike — is made at a single Rolls-Royce plant at Raynesway in Derby, and Australia has already paid into it: about A$4.6 billion (£2.4 billion) over ten years to help double the site, plus a further A$310 million in early 2026 for the long-lead components of its first two units. Money up front, for a place in a queue. Rolls-Royce says the work is going well — around seven reactors are in build, the design was frozen at the end of 2025, the plant has doubled, and it speaks of a "drumbeat" of one reactor every eighteen months, falling to twelve. But that is the optimistic account, from the same British nuclear-submarine enterprise whose attack boats ran 57 months late and sit today at zero availability, and independent analysts already judge the drumbeat target too optimistic. And Australia is not first in that queue. The Derby line must feed, in order of Britain's own national priority, the Dreadnought ballistic-missile boats — the untouchable core of the UK nuclear deterrent — then the Royal Navy's twelve SSN-AUKUS, and only then Australia's five. When one plant is stretched across three programs and the tempo slips, as it has slipped in every British submarine program in living memory, it is not Britain's deterrent that waits; it is the customer at the back of the line. Australia has prepaid billions for reactors it cannot make, cannot source anywhere else, and cannot be sure will arrive when the hulls in Adelaide are ready to receive them.

11. The numbers, and the coastline

Suppose every engineering risk in this memo is overcome — the yard rises, the workforce is trained, the two fleets and two reactors are made to work, and the boats are delivered. There is still the question Paul Keating put at the National Press Club in 2023, and it is one of arithmetic and geography.

Australia has the world's third-largest maritime domain and a coastline of nearly sixty thousand kilometres. Against that, the plan fields eight submarines — and eight is never eight. Navies plan on roughly a third of a submarine force being unavailable at any moment, in deep maintenance, refit or trials, and the real figures are often worse: the United States has had around forty per cent of its attack fleet unavailable at times, and Britain's entire Astute fleet, as this is written, is at zero. Australia's own Collins boats, thirty years on, have lately mustered as little as one of six available — the Collins lesson, still unlearned. Take the ordinary rule and eight boats become five or six available; account for the transit time across Australia's vast approaches and the number actually on station, in any one place at any one time, is very small. Keating's phrase for that contribution was "throwing toothpicks at a mountain."

The tempo makes it worse. On the government's own timeline the Australian-built boats arrive one every three years from the early 2040s — so the five-boat sovereign fleet is not complete until the late 2050s, by which point the first of them is already middle-aged. There is no point at which Australia has five new boats at once; there is a slow trickle, and the availability rule bites the whole way along it. Nor is Australia first in line for the design it is co-building: the United Kingdom begins construction in the late 2020s and expects its own SSN-AUKUS boats in Royal Navy service in the late 2030s — years ahead of Australia's first. On the reactor, on the design, and on the water, the partner's boats come first, and the Australian ones wait.

His alternative was mass. For the money AUKUS will cost, he argued, a far larger number of smaller, cheaper conventional submarines could be fielded — he put the figure at around forty-five Collins-class boats against the price of a handful of nuclear ones. The exact multiple is his, and rhetorical; the direction is not seriously disputed. A conventional submarine costs a fraction of a nuclear one, and the original program AUKUS replaced had proposed twenty of them; respected strategists such as Hugh White have argued for a force of twenty-four to thirty-six. Numbers matter when the task is to cover a continent's approaches. It is an old lesson of undersea warfare — the reason Germany built its submarine fleet in the dozens rather than the handful — that a boat, however capable, cannot be in two places at once, and a coast this size asks for presence in many. (The analogy is about numbers and dispersal, not the mission; mass alone has never been a guarantee — but a fleet too small to be present is a different kind of failure.)

The size of the boat sharpens the same point. At more than 10,000 tonnes — roughly a third larger than the Astute — SSN-AUKUS is built for deep, blue-water, long-range work. Much of the water Australia would actually have to defend — the shallow shelf seas and the archipelagic approaches to the north — is confined and shallow, where a very large nuclear boat is less suited and a smaller conventional one has long been considered the better fit. Keating's blunt version was that these boats are "too large to operate in the shallow waters on our continental shelf," and are better explained as forward defence for the United States than as the defence of Australia.

Honesty requires the other side, because it is real. A nuclear boat is a genuine leap over a conventional one. It has almost unlimited range; it does not have to surface every few days to recharge and betray its position, as the Collins must; it can stay submerged for months and strike from far off. For deterrence at distance — holding an adversary's approaches at risk far from home — the nuclear submarine is the superior instrument, and the government's strategists argue, not unreasonably, that this is the capability the times demand. The disagreement is not about whether a nuclear submarine is a better submarine. It is whether eight of them — five, realistically, at sea, first arriving in the 2040s — are the right answer to the defence of this coastline, at this price, on this timeline, when the same money could put far more hulls in the water far sooner. That is a judgement the country is entitled to make with its eyes open, and it turns on numbers the plan prefers not to dwell on.

12. The boats may be obsolete before they float

There is a last risk, and it is the one a planner should worry about most, because it waits at the far end of the timeline, after the money has all been spent. The SSN-AUKUS boats are a thirty-year commitment and will serve into the 2060s. The undersea world they will operate in is not the one being described today.

The direction of travel is already visible — in Australia's own navy. While the crewed program is renegotiated at summits, the cheapest and fastest-moving part of the build-up is already in the water. In April 2026 the Navy stood up a dedicated Maritime Autonomous Systems Unit and began taking delivery of dozens of Ghost Shark extra-large uncrewed submarines — sovereign, built in Sydney, under a A$1.7 billion program a fraction of the AUKUS cost, entering service now rather than in the 2040s. Alongside them come the smaller Speartooth and the surface Bluebottle. The contrast is hard to miss: the uncrewed fleet did not need a treaty or a shipyard that does not yet exist. It needed a factory, and it is here.

These systems are cheap, numerous, and — with no one aboard — able to work in dangerous water where the loss of a vessel costs no lives. Australia's own strategists describe them taking on the persistent, high-risk undersea tasks a crewed boat cannot safely do: watching chokepoints, laying mines, loitering on the seabed awaiting instructions, and forming a distributed sensor-and-strike network across the approaches. The broader current runs the same way. Sensors, artificial intelligence and networked detection are advancing fast, and the ocean is becoming steadily less opaque. The entire premise of a submarine is that it cannot be found; committing to a small number of very large, very expensive crewed boats being reliably hidden in the 2050s and 2060s is a wager against the direction the technology is plainly moving. As cheap, patient, autonomous systems multiply on every side, the danger grows that the most exquisite and expensive platform in the fleet becomes the most perilous place to put a crew.

Honesty demands the counter, and it is a serious one. No one — not the government, not the builder, not the independent analysts — claims uncrewed submarines can replace a nuclear attack boat today. Ghost Shark is slow, cannot manoeuvre or strike like an SSN, and carries a fraction of the range and payload; the official vision is a "mesh fleet" with crewed submarines at the centre of a network of autonomous systems, not one in which the drones take over. For deep-ocean, long-range, high-end work, the crewed SSN remains, for now, in a class of its own.

But "for now" is the whole difficulty, because the SSNs are not a now capability — they are a 2040s-to-2060s one. The plan asks the country to spend $368 billion and three decades on the crewed nuclear submarine being the right centre of gravity deep into a future that undersea autonomy is already reshaping — while, in the same navy, it funds the cheaper and faster alternative it insists is only a complement. Perhaps the crewed boat holds its place; perhaps the mesh fleet is the right model and the SSNs its rightful core. But the plan has never openly argued that case against the technology curve, nor costed the risk that by the time the boats arrive the domain has shifted beneath them. It is one more question the country deserves answered before the money is gone: are we buying, at the greatest expense in our history, the capital ship of an age that is closing?

13. The base is a fault line too

The mismatch between a very large nuclear boat and Australia's waters is not only an open-ocean problem. It shows up at the front door — at the base the plan is spending eight billion dollars to build.

Australia's nuclear submarines will be home-ported at HMAS Stirling, on Garden Island in Cockburn Sound, near Perth. From 2027 the same base becomes the home of Submarine Rotational Force–West: a rotating presence of up to four US Virginia-class boats and one UK Astute-class boat, with several hundred United States personnel stationed to support them. The Commonwealth is investing up to eight billion dollars over a decade to expand it. It is, in effect, the country's principal nuclear-submarine base and a major United States strategic foothold on Australian soil — one analyst, echoing the late Desmond Ball, has called Cockburn Sound the second "suitable piece of real estate" after Pine Gap. Three problems with the site compound one another.

The water. Cockburn Sound is a shallow, enclosed body of water tucked behind Garden Island — calm, which is useful, but shallow enough that the naval approaches are dredged, engineered channels, cut and maintained through the sound and plainly visible from the air, with a causeway tying the island to the mainland that constrains the water further. A submarine built for the deep ocean has to be worked through those narrow, shallow channels every time it sails — the same size-versus-suitability problem from the open sea, reproduced at the front door.

The doorway. And that bears on the one thing a submarine exists to do: disappear. A boat tied to this base cannot slip into the deep from wherever it chooses. It must funnel through the same few dredged channels on every departure and every return, at its shallowest, slowest and least concealed. A hidden submarine is the entire point of a submarine — and one that must announce itself through a known doorway has surrendered that edge before it reaches open water. An adversary does not have to find the boat in the ocean; it need only watch the doorway. Fixed, shallow channels are the easiest water in the world to monitor — the natural place to lay the kind of seabed sensor arrays the United States and the Soviet Union strung across chokepoints throughout the Cold War — and a boat picked up on its way out can be trailed from the moment it leaves. Navies do sweep and screen their approaches, and Australia's would; but a single-port, fixed-channel arrangement makes that a permanent task against a permanent disadvantage — the opposite of the dispersal that keeps submarines survivable.

The target. A port that home-bases nuclear-powered submarines and hosts a rotating force of American boats is not an ordinary naval base; it is one of the highest-value military targets on the continent, precisely the kind of node an adversary would strike first in a serious conflict. Concentrating the sovereign fleet and its allied partners in a single shallow sound turns one location into a strategic bullseye — and unlike a dispersed force of smaller boats, a fleet tied to one specialised, nuclear-capable port cannot simply scatter. (The boats are nuclear-powered but conventionally armed; the point is the value of the target and the reactors it concentrates, not weapons aboard.)

The neighbours. HMAS Stirling is not in a remote corner of the country. It sits within the Perth metropolitan area, beside the city of Rockingham and close to Kwinana and Fremantle — a settled, growing region of hundreds of thousands of people. Onto that the plan layers naval nuclear reactors, the handling and temporary storage of low-level radioactive waste from submarine maintenance, and the accident risk that attends them. A 2024 parliamentary inquiry recorded community uncertainty about how much nuclear waste would be stored at the site, and independent analysts have noted that accident-scenario modelling, licensing submissions and emergency-management arrangements have not been made public. Whatever the true level of risk, the point of principle is plain: the country is concentrating its nuclear-submarine enterprise, and a foreign nuclear-submarine force, in a shallow sound on the edge of a capital city's suburbs — and asking the public to trust the assurances without seeing the workings.

None of this says Australia should have no submarine base. It says the plan has chosen the largest possible boats, tied them to a single shallow-water port that must be dredged to fit them, made that port a first-order target, and set the whole arrangement beside a major population centre — and, as with the rest of the program, has not squarely answered the questions that raises.

14. The end no one costs: the reactors

Every nuclear submarine reaches the end of its life, and when it does it leaves behind the hardest waste there is. This is the part of the program almost no one discusses, and it is the part with no solution.

Australia's submarines are fuelled with highly enriched uranium — weapons-grade material, sealed into a reactor built to run for the life of the boat without refuelling. That choice buys endurance, and buys a harder problem at the end. When a boat retires it leaves a spent reactor and spent fuel that stay dangerously radioactive not for years but for hundreds of thousands of them. On the government's own figures, a single Virginia reactor weighs more than a hundred tonnes and holds around two hundred kilograms of weapons-grade uranium; the spent fuel from the first three boats alone runs to hundreds of kilograms of the most tightly guarded material on earth. Released documents describe the reactor compartment as roughly the size of a four-wheel drive and the spent fuel as about the size of a small hatchback — small in volume, lethal for geological time.

Under the treaty it is Australia's alone. Australia has committed to be solely responsible for managing all the waste from its own submarines — operations, maintenance, and, at the end, the reactors and the spent fuel — and it cannot send that waste offshore. It will not take other nations' waste, and the visiting US and UK boats take their reactors home; but every Australian-owned boat, including the three second-hand Virginias, becomes Australia's to deal with, for as long as the material stays dangerous — which is to say, effectively forever.

And here is the state of the plan for it: there is no site, no facility, and no costing. The search for a location began in 2023, to land somewhere on "current or future" Defence land, with the first high-level waste not due for disposal until around the mid-2050s. But a country that has spent seventy years — since its first research reactor — unable to agree where to put even its low-level medical waste is now committed to siting a store for weapons-grade waste that must be guarded for hundreds of thousands of years. The last serious attempt, South Australia's 2016 royal commission into a high-level repository, collapsed when twenty-seven Native Title groups and a majority of its own citizens' jury rejected it. Nor is there a partner whose example reassures: Britain has twenty-three decommissioned submarines still afloat awaiting disposal, decades on; the United States keeps more than a hundred retired reactors in an open trench in Washington state; and Finland, the only nation actually building a permanent deep repository, has taken more than forty years to reach the point of sealing it.

The honest version is not that this is technically impossible — scientists note that storing high-level waste is, in engineering terms, relatively well understood. It is that the problem is political and social, and on that measure Australia has a seventy-year record of failure, no site, no price, and a commitment it cannot walk back. The submarines are sold to the public as the capability. The reactors they will leave behind — weapons-grade, effectively permanent, and nobody's yet — are the part of the ledger the program does not read out.

15. We have built submarines before

This memo opened with Collins as a warning. Having now seen the plan in full — the three-nation graft, the two reactors, the greenfield yard, the overstretched partners, the base and the waste — it is worth returning to that warning and setting out, in full, what it actually contains. The United States cannot keep pace and the United Kingdom is a cautionary tale, but Australia does not have to look abroad: its own record is still afloat.

The Collins class — six diesel-electric boats, the country's first home-built submarines — was contracted in 1987 for about $3.9 billion and meant to be a straightforward success. It became one of the most troubled procurements in the nation's history. The boats were dubbed "dud subs" and "noisy as a rock concert." They suffered unreliable engines, cracking propellers, cavitation that gave away their position, periscopes that vibrated and would not focus, and a combat system that a landmark 1999 review found would never work properly and had to be scrapped and replaced for hundreds of millions more. The last of the six was not delivered until 2003, and in June 2011 a national newspaper could report that not one of them was seaworthy.

It was, in the end, brought good — but note how. It took a brutal external review, led by a British naval architect brought in specifically to fix it, to establish that the boats were sound and the system supporting them was not; and it took years more, and sustainment budgets that ran past half a billion dollars a year. Even now the recovery is fragile: in December 2024 five of the six boats were reported out of action, one available, and the government placed the class back on its official "Project of Concern" list — where, one veteran observer wrote, it may well stay until the boats retire. This year's sustainment bill is around three-quarters of a billion dollars, for as little as a single available submarine.

That is the honest content of the reassurance that "Australia built Collins and made it work." It did — at roughly three times the contracted cost, more than a decade late to reliability, with a foreign expert called in to rescue it, and with the fleet still troubled thirty years on. And here is the part that bears on AUKUS: the clearest lesson many drew from Collins was that Australia's procurement disasters begin with the decision to reject a low-risk, proven, off-the-shelf option in favour of building something bespoke — and that Defence made that same choice again and again, each time believing it could do better than the last.

AUKUS is that choice once more, raised to a power. Collins was a conventional boat, of a single settled Swedish design, and it still became a decades-long, multi-billion-dollar ordeal. AUKUS proposes to build a first-of-class nuclear boat, of a design that does not yet exist, merged from two nations' engineering, a third larger than either, in a yard not yet built, with a reactor bought from an overstretched foreign supplier, and a second foreign fleet to run alongside it. Every dial that made Collins hard has been turned to its maximum. The simplest question of all is the one the country has not been squarely asked: we have done a version of this before, and we know what it cost — did we learn anything from it?

16. The sub to nowhere

Set the pieces side by side, because it is the accumulation that makes the point.

The plan is to build a first-of-class submarine whose design is not yet finished; that fuses two nations' engineering traditions into one boat for the first time; that is enlarged beyond either parent; in a yard that does not yet exist; by a workforce, a supply chain, a regulator and a set of nuclear skills that have to be created from a standing start; in a country that has never built a nuclear submarine. The fleet at the end runs two different classes and two different reactors at once. The interim capability meant to cover the wait depends on a United States that cannot build the boats it has already sold, and now supplies them part-worn from a fleet already below its own minimum. The design partner is a United Kingdom whose comparable program ran 57 months late and tripled in cost, and whose entire attack fleet sits idle as this is written. And the first Australian-built boat is not due until the early 2040s — on a plan where every prior program of this kind, in both partner nations, ran years late and billions over.

None of this is an argument that Australians cannot build hard things. But the one time the country built its own submarines, the Collins class, it took roughly three times the budget, more than a decade past its promised reliability and a foreign rescue to get there — and the fleet is troubled still (Section 15). This plan repeats that bespoke gamble and multiplies it. It is an argument about this plan specifically: it takes every category of risk a submarine program can carry — first-of-class, multi-nation design integration, a size increase, two reactors, a greenfield yard, a workforce built from scratch, dependence on two overstretched partners — and runs all of them at once, for a capability that arrives, if it arrives on schedule, two decades from the decision, and that many who have operated both kinds of boat judge the wrong fit for the coastline it is meant to defend.

These are not quibbles for the shipyard. They are first-order questions the people who committed the country to this plan have not answered in plain terms, and they deserve to be:

• Why is a nation operating its first nuclear reactors committing to run two different national types — the American S9G and the British PWR3+ — at once, and what does that double burden cost?
• Why buy the largest, deepest-water boat for a task that includes so much shallow, confined water?
• How many foreign specialists must be imported to deliver the Australian jobs the plan advertises, and for how long?
• Why take three part-worn submarines from a fleet already below its own minimum, and how soon must they be replaced?
• And how many boats will actually be at sea, on station, in the 2040s — against a coastline this size?
• Why base the whole fleet — and a rotating foreign nuclear-submarine force — in one shallow sound on the edge of a capital’s suburbs, and what is the plan if that single port is lost?
• Where will the spent reactors and weapons-grade fuel go, what will it cost, and why is a nation that has never sited even a low-level waste store committing to guard high-level waste for hundreds of thousands of years?
• Why has Australia prepaid billions for reactors made at a single British plant — behind Britain’s own nuclear deterrent and navy in the queue — from a supplier whose submarines run years late, with no other source to fall back on?
• We built the Collins class once — at roughly three times the cost and a decade late to reliability — so what in this plan, harder in every dimension, suggests it will go better?

The companion memo, The AUKUS Cost Blowout, sets out what all of this costs in money, and what the same money could build instead. The question this memo raises is narrower and prior: a submarine program is a bet that a long chain of hard things will each go right. This one asks more of that chain than any comparable program before it, asks it of a builder attempting the feat for the first time, and — even if the whole chain holds — may deliver too few boats, too late, in the wrong size, for the coast it is meant to guard. That is the sub to nowhere. Not because the destination is wrong, but because the road, as drawn, may never arrive — and may lead somewhere other than the defence of Australia.

And it should be measured against the plainest test of all. Even finished — even if the chain of hard things somehow all held — this program does not put a litre of fuel in the tank the country actually runs on, the vulnerability its companion memos set out and the one most likely to bring Australia to a halt. And it cannot fix that even in principle: a submarine hunts ships and stays hidden — it cannot escort a convoy, clear a mined strait, or hold a sea lane open, so it can neither protect the routes Australia's fuel and trade sail through nor re-open a chokepoint once it is closed. It cannot be built on this plan without extraordinary luck; it may yet be overtaken by the undersea technology of its own era; and, most fundamentally, it is the wrong tool for the mission it is sold to perform — while leaving the nation's most immediate weakness exactly where it found it.

That is why the position of Modern Movement Australia is not a slogan but a request a reasonable country makes of itself: an independent review of the submarine program — its buildability, its true cost, its strategic fit, and its opportunity cost against the vulnerabilities it leaves unaddressed — before another dollar is committed. If the case for these boats is as strong as the government maintains, an honest review will confirm it. If it is not, far better to know now, while the money is still in the country's hands, than in the 2040s when it is spent and the boats have not come. The questions in this memo are not the objections of opponents; they are the questions any such review would have to answer — and they deserve answering before, not after.

References

1. SSN-AUKUS design, basis and characteristics (UK next-generation SSNR design; US propulsion components, common vertical launch system, weapons and joint combat system; shared UK/Australia design; over 10,000 tonnes submerged (roughly a third larger than the Astute); PWR3+ reactor from Rolls-Royce Raynesway, Derby; core lasts the life of the boat) — SSN-AUKUS, Wikipedia; Australian Submarine Agency, "Australia's nuclear-powered submarines."
2. Build split and timeline (UK up to 12 at Barrow, construction from the late 2020s, RN service late 2030s; Australia 5 at Osborne via ASC–BAE Systems joint venture; construction to begin by the end of the decade; first Australian-built boat delivered early 2040s; one boat roughly every three years; interim purchase of 3 Virginia-class, option on 2 more) — Australian Submarine Agency; Navy Lookout; Army Recognition, February 2026.
3. Osborne Submarine Construction Yard (about A$3.9 billion down payment, 15 February 2026; EPBC approval 19 February 2026; yard roughly three times the Attack-class yard; ~4,000 workers to build the yard, ~4,000–5,500 to build the submarines at peak; up to 1,000 apprentices per year; Barrow methodology to be replicated "identically"; Australia as a "fourth production line"; British and American experts advising) — Australian Submarine Agency, Osborne Submarine Construction Yard; Navy Lookout, "Building SSN-AUKUS"; Australian Defence Magazine, February 2026.
4. Australian Naval Nuclear Power Safety Regulator; RAN personnel embedded in US/UK nuclear programs — Australian Submarine Agency, AUKUS agreement.
5. UK retired-submarine dismantling liability (more than 20 retired boats awaiting dismantling; HMS Dreadnought retired 1980 not yet dismantled; storage cost and long-term liability in the billions) — UK National Audit Office (2019), cited in The Economy, June 2026.
6. Two reactor types (Virginia-class: US S9G reactor, General Electric / Knolls Atomic Power Laboratory, ~210 MWt, life-of-ship core; SSN-AUKUS: Rolls-Royce PWR3+, built at Derby, PWR3 "based on a US design but using UK reactor technology") — S9G reactor, Wikipedia and GlobalSecurity; Rolls-Royce PWR, Wikipedia; World Nuclear Association, "Nuclear-Powered Ships."
7. Reactor / hull life (Virginia S9G ~33-year life-of-ship core; Astute PWR2 Core H ~25 years; per-year cost comparison) — Virginia-class submarine, Wikipedia; World Nuclear Association; Navy Lookout, "Nuclear submarines for Australia — what are the options?"
8. US Virginia-class production rate and backlog (funded at two per year since 2011 but never reaching 2.0; about 1.1–1.2 per year since 2022; 2.33 per year needed to sustain the fleet and supply Australia; recovery to 2.0 slipped from 2028 to 2032; fleet ~46 vs a requirement of 66; ~40% at times unavailable; ~US$5 billion per boat; $1.9 billion 2026 plus-up to avert a stop-work order; Australia funding US shipyard expansion) — Congressional Research Service, RL32418 (26 January 2026); USNI News; 19FortyFive, June 2026; CT Mirror, August 2025.
9. Revision of 30 May 2026 and Virginia delivery dates 2032/2035/2038 (Australia's first submarines to come from the existing US fleet because new boats will not be ready in time) — 19FortyFive, June 2026; The Conversation, "Why is Australia buying used submarines?" (Jennifer Parker), 2026.
10. UK Astute-class history and cost growth (studies from 1986; first boat laid down 2001, commissioned 2010, fully operational 2014; House of Commons Defence Committee 2009 — first three boats 57 months late, 53% / £1.35 billion over budget; whole-programme valuation £4.3 billion (2001) → £6.6 billion (2009) → £9.6 billion (2015) → £12.2 billion (2025); boat 4 ~17 months late) — Astute-class submarine, Wikipedia; National Audit Office figures via Defence Eye, May 2025; National Security Journal, July 2025.
11. Astute root causes (17-year "valley of death" at Barrow; MoD staff at the yard cut from ~50 to fewer than 6; incomplete design drawings; first UK submarine designed entirely in 3D CAD; General Dynamics Electric Boat brought in to run shipyard management and workflow) — Astute-class submarine, Wikipedia; ASPI / United States Studies Centre, "Astute lessons for Australia's AUKUS submarine workforce."
12. Astute technical faults (lead-boat grounding; inability to reach top speed; flooding during a dive; corrosion attributed by a confidential MoD memo to cost-cutting and called "a cause for major concern") — Nuclear Information Service; The Guardian reporting cited therein.
13. Astute fleet availability, June 2026 (all five operational boats simultaneously in port — 0% availability — after defects on HMS Anson following an exercise with Australia; UK nuclear maintenance concentrated at a single site; Barrow stretched across Astute, Dreadnought and SSN-AUKUS) — The Telegraph, 7 June 2026, via The Economy and Defence Security Asia.
14. Risk of Australia repeating the Astute experience (BAE Systems as contractor for both; Australia building from a low industrial base) — ASPI / United States Studies Centre.
15. Paul Keating, National Press Club, 15 March 2023 (nuclear boats "too large to operate in the shallow waters on our continental shelf"; ~45 Collins-class for the cost of a handful of nuclear boats; "throwing toothpicks at a mountain"; "the worst deal in all history") — National Press Club address, 15 March 2023, as reported; Lowy Institute, "Deterring at a distance."
16. Conventional-alternative force proposals (original Future Submarine Program ~20 conventional boats; Hugh White 24–36 conventional submarines) — Lowy Institute, "Deterring at a distance"; contemporary reporting.
17. Collins-class characteristics and the SSN advantage case (Collins ~3,400 tonnes submerged; must surface roughly every three days to recharge; SSN range, endurance, submerged months, Tomahawk reach — the strategic case for nuclear over conventional) — Collins-class submarine, Wikipedia; Lowy Institute, "Deterring at a distance."
18. Submarine-force availability rule of thumb (~20% of the force in depot maintenance at any time as best practice; higher in practice) — Congressional Research Service, RL32418.
19. HMAS Stirling / Garden Island / Cockburn Sound (principal nuclear-submarine base near Perth; ~$8 billion upgrade over a decade; Submarine Rotational Force–West from 2027 — up to 4 US Virginia-class and 1 UK Astute-class boat, ~500–700 US personnel; shallow, enclosed sound requiring dredging; likened to a second Pine Gap; 2024 Parliamentary Standing Committee on Public Works inquiry and community concern over nuclear-waste storage, accident modelling and disclosure) — Australian Submarine Agency; Department of Defence, SRF-West; Lowy Institute, "A second suitable piece of real estate"; Nautilus Institute; The Guardian, August 2025.
20. AUKUS nuclear waste (Australia "solely responsible" for all radioactive waste from its own Virginia-class and SSN-AUKUS submarines, including spent fuel, and cannot export it; highly-enriched, weapons-grade fuel; a Virginia reactor over 100 tonnes with ~200 kg of HEU; reactor compartment ~"size of a four-wheel drive," spent fuel ~"size of a small hatchback"; no site and no costings, site search from 2023 on Defence land, first high-level waste ~mid-2050s; Australia's 70-year failure to site even low-level waste; SA 2016 Royal Commission rejected by 27 Native Title groups and a citizens' jury; UK 23 decommissioned submarines awaiting disposal, US 100+ retired reactors in a Washington-state trench, Finland's 40-year repository) — Australian Submarine Agency, "Nuclear stewardship and waste"; The Guardian; The Conversation; Cosmos Magazine; contemporary reporting, 2024–2025.
21. Australian funding of Rolls-Royce and reactor supply (A$4.6 billion / £2.4 billion over ten years to expand Rolls-Royce Submarines at Raynesway, Derby; A$310 million, February 2026, for long-lead nuclear-propulsion items for the first two Australian boats; UK to supply complete welded reactor units; all AUKUS attack-submarine reactors manufactured only at Raynesway; ~7 reactors in build with critical design review completed end 2025 and the plant doubled; "drumbeat" target of one reactor every 18 months falling to 12, judged optimistic by analysts; the Derby line shared across Dreadnought SSBNs, UK SSN-AUKUS and Australian boats; SSN-AUKUS displacement now cited at over 10,000 tonnes vs Astute ~7,400) — Naval News; Australian Defence Magazine; Naval Technology / Army Technology (UDT 2026); Navy Lookout; SSN-AUKUS, Wikipedia; Defence Security Asia.
22. Collins-class submarine program (six diesel-electric boats, Australia's first home-built submarines; contracted 1987 ~$3.9 billion, last boat delivered 2003; "dud subs"/"noisy as a rock concert"; engine, propeller-cavitation, periscope and combat-system faults; McIntosh–Prescott review 1999 found the combat system would never work properly and had to be replaced; June 2011 "not a single submarine seaworthy"; Coles review turnaround led by a UK naval architect; sustainment >$500m/yr, ~$769m in 2024; December 2024 five of six out of action and returned to the Project of Concern List; lesson that procurement failures begin with rejecting off-the-shelf options) — National Security Journal; ASPI, "Nobody wins unless everybody wins" (Coles review); The Australian; Michael West Media; McIntosh–Prescott report, 1999.
23. Collaborative submarine-build models (US Columbia-class and UK Dreadnought-class share a Common Missile Compartment and reactor/machinery technology, with each nation building its own hull around the shared centre; Germany, Norway and Italy adopting the single Type 212/212CD design with national modifications; US Virginia class built by two rival yards to one common design) — Naval News, "New Submarines Compared: Columbia, Dreadnought and SNLE-3G" (2021); USNI News, "New European Attack Submarine Programs" (2021); Congressional Research Service, R41129 (Columbia program).
24. Pat Conroy, Minister for Defence Industry, on the lessons of Collins (National Press Club address, 28 November 2023: the "critical lessons from the Collins build experience" included combining a new hull design with new combat and weapons systems, being a lone "parent" navy, and insufficient through-life support; the SSN-AUKUS design will be new, but the combat system is to be an evolved version of the existing AN/BYG-1 co-developed by the US and Australia; the vertical launch system distinguishes SSN-AUKUS from the Collins and Astute classes; the minister again defended AUKUS as sovereign manufacturing at the National Press Club on 2 July 2026) — Minister for Defence, speeches, 28 November 2023 and 2 July 2026; Naval News, December 2023.
25. Australian uncrewed undersea programs (Ghost Shark XL-AUV, jointly developed by Anduril and the Department of Defence — first prototype revealed April 2024; A$1.7 billion contract for dozens in September 2025; first operational vehicle delivered January 2026; Maritime Autonomous Systems Unit stood up 14 April 2026 under Project SEA 1200, operating Ghost Shark, Speartooth and Bluebottle; uncrewed systems used for persistent surveillance, minelaying, seabed loiter and distributed sensor/strike networks in high-risk waters; official framing as a complement to, not a replacement for, crewed submarines) — Ghost Shark, Wikipedia; ASPI Strategist, "Sharks for filling the moat" (September 2025); Second Line of Defense (April 2026); Army Recognition and Global Defense Corp (April 2026).