What a Nuclear Submarine Needs Ashore

Every facility, every maintenance tier and every waste stream a nuclear-powered submarine requires. What Britain and America actually run. What Australia has announced.

1. Neither partner does this from one place

The first thing the comparison shows is that the functions are separated deliberately. An operating base is not a maintenance yard. A maintenance yard is not a weapons depot. A weapons depot is deliberately somewhere else again: Coulport sits three kilometres overland from Faslane and nineteen by sea, and the American equivalents are underground stores on separate bases with their own military guard forces.

Two figures put the scale beyond argument. The United States uses two private and four public shipyards, with approximately seventy thousand employees between them, to overhaul, repair, refuel, inactivate and dispose of its nuclear-powered ships, supported by more than eight hundred industrial contractors; the Department of Energy describes these as unique industrial assets with capabilities found nowhere else in the United States. And Britain, with an active fleet of ten boats, has just two sites in the country able to take a submarine out of the water for work requiring full access to the hull, and does not have enough dock capacity for the fleet it has.

2. What each site is for

Australia has three sites under way. A fourth was announced in 2022 and has no site; the functions of a weapons depot and a waste site have no location at all; and a seventh requirement is not a site but the route between them. How many separate establishments Australia needs is a design question, and eight hulls do not automatically require a copy of an American map built for ninety-seven reactors. What both partners do is separate these functions, and what Australia has published does not yet say where two of them go. Henderson is the one carrying the most: it must deliver six separate programmes concurrently, of which the nuclear work is only two.

1. HMAS Stirling — Garden Island, Western Australia

The operating base.

The Submarine Agency has separately valued the Stirling upgrades at $5.2 billion. The two figures have not been reconciled.

2. Henderson — Australian Marine Complex, Western Australia

The maintenance site — and five other things at once.

Six separate programmes concurrently, of which the nuclear work is two. Site and boundary announced August 2026; the master plan has no industry partner; the dry dock is at preliminary design.

3. Osborne — Lefevre Peninsula, South Australia

The construction yard.

Not a maintenance site. No graving dock is embedded in the design, so overhauls must go to Henderson.

4. An east coast base — site not named

A second operating base.

Announced for four and a half years without a site.

5. A weapons depot — not announced

Storage and issue of submarine weapons between loadings.

No site, no policy on custody or handling, and no cost. Both partners treat this as a separate establishment, not a function of an operating base.

6. A waste and disposal site — not announced

Storage and permanent disposal.

No site, no facility, no cost and no route. Four attempts to site a national low-level facility have failed.

The east coast base has been announced for four and a half years without a site. The weapons depot and the waste site have never been announced at all, and both partners treat each as a separate establishment rather than a function of an operating base.

7. Getting the reactor to the site

The step between the yard and the storage, which has no Australian equivalent.

A retired reactor compartment is not freight. It is a single sealed monolith containing the activated pressure vessel and its internals, weighing well over a thousand tonnes, which cannot be divided and must be isolated for centuries. Australia’s own documents describe the two streams that come out of a boat: spent fuel roughly the size of a small hatchback, which is high-level waste, and a reactor compartment roughly the size of a four-wheel drive, which is intermediate-level.

The United States has done this 131 times since 1986, and the route is the point.

The American route is a water route. A barge from the yard to a river port, then twenty-five miles of road inside a closed federal reservation. That is what makes a thousand-tonne indivisible load movable.

Both ends of the chain have been chosen on opposite principles. The yards are metropolitan: Osborne is fifteen kilometres from the centre of Adelaide, Henderson twenty-three south of Perth, Stirling five off Rockingham. The disposal site is to be on remote defence land, which in Australia means inland. Nothing published joins them.

3. What is documented, and where the money is visible

Australia has a standing mechanism for exactly this. The Parliamentary Standing Committee on Public Works examines significant public works before the money is spent: Defence projects above $75 million are generally referred to it, and it reports publicly on scope, cost, risk and delivery arrangements. Where that process has been used, the documents exist and anyone can read them. Where it has not, there is nothing to read.

The asymmetry is not a matter of interpretation. Two sites deliver the same programme, one is examined by a parliamentary committee before it spends and one is not, and the difference rests on an exemption written for a company that no longer exists.

The funding, as announced

These are not a budget. They are a map of what has been said. The rows mix ministerial announcements, decade totals described as “indicated” by planning advice, figures that have not been reconciled with each other, and a programme envelope of $268 to $368 billion whose composition has never been published. They should be read as the public record of statements, not as an account of committed money.

Two of those rows carry a cross-check from outside Australia. The House of Commons Defence Committee, reporting in April 2026, put the Stirling investment at about £4 billion and the Henderson precinct at about £12.5 billion. A British parliamentary committee has therefore published an estimate of Australian infrastructure spending in a document of a kind Australia has not produced.

Four of the eight rows above are zero: the second operating base, the weapons depot, the waste site, and the route between the yards and the site. Between them they are the four requirements with no published cost, no site and no document.

The programme beyond the sites

Shore infrastructure is one part of the bill. These are the other announced commitments, for context, because the infrastructure figures are often quoted as though they were the programme.

Two of the largest single commitments buy capacity in someone else’s country rather than a capability in this one: about $3 billion to American yards and about $4.6 billion to British industry, neither of which delivers a submarine.

4. The list

What follows is every category of shore facility and system a nuclear-powered submarine requires. It is grouped by function rather than by site, because the same function may be delivered in different places.

A note on what the third column means. “Not announced” is not the same as “does not exist”. Some of this work will be classified, some will be phased, and some will sit in business cases that have not been released. The register and this memo are entitled to ask for a public specification; neither is entitled to treat silence as proof of a vacuum. Where the silence has run for five years and concerns a site, a custody arrangement or a route — weapons storage, spent fuel, the reactor compartments, and the road or water between the yards and the disposal site — it has stopped being a planning lag and become the thing to explain.

5. The workforce

Every facility above is operated by people, and the qualifications are the long-lead item. A graving dock can be built in seven to ten years. A nuclear-qualified welder takes at least ten to produce, and the assessment of a retired submarine force commander is that the work demands a high proportion of master-level trades with decades of precision manufacturing experience.

The Australian figures are projections against a labour market that is already short. Published analysis puts the national shortfall at about 110,000 skilled trades workers, and states plainly that there is no surplus workforce waiting to be mobilised. Osborne and Henderson must recruit concurrently, in two states at historically low unemployment, against the resources, infrastructure and energy sectors. At Henderson the competition is mining, liquefied natural gas and offshore resources directly.

What has to be produced, and how long each takes

The crews

The industrial workforce builds and maintains the boats. A separate workforce takes them to sea, and that requirement has never been published either.

The transition is already visible. Six or more years before Australia receives its first nuclear submarine, the Chief of Navy has stated that more sailors are entering the nuclear submarine pipeline than the Collins pipeline. Published analysis draws the consequence: the need to siphon personnel into the nuclear pipeline means the retention difficulties of the Collins era could return to the Collins class itself, at the moment those boats are meant to be bridging the gap.

No Australian document states the total crewing requirement for a fleet of eight nuclear submarines, the number of qualified submariners available today, or the trajectory between them. The trades figure and the crew figure are both absent, and they are different problems: one is produced by an academy, the other by a decade of sea time.

Where the people are expected to come from

Two authorities have already answered this, and neither answer is that they will be Australian.

The House of Commons Defence Committee, reporting in April 2026, recommended that the three governments design a dedicated AUKUS visa, with pre-approved security clearances and mutual recognition of qualifications, on the finding that existing immigration channels are too slow to support joint construction and maintenance. Australia’s Skills in Demand visa carries waiting times of up to seven months; the United States E-3 carries academic requirements.

And the Premier of the state building the submarines told a defence conference in 2026: AUKUS is housing policy. AUKUS is education and skills policy. AUKUS is migration policy.

No document states how many of the 20,000 are expected to be recruited from overseas, on what visa, for how long, or what proportion of the workforce is expected to be Australian at any point in the programme. Nor has any Australian industry content requirement been published for the submarines — unlike the Hunter class frigates, built at the same yard by the same company, which carry a contractual minimum of 58 per cent.

6. Liability, and who pays

Every facility above carries a risk, and every nuclear operator in the world sits behind a liability regime. Australia has accepted the risk and has not built the regime.

The indemnity is in the treaty text. Australia shall indemnify both governments against any liability, loss, costs, damage or injury, including third-party claims, arising out of or resulting from nuclear risks connected with the design, manufacture, assembly, transfer or utilisation of any material or equipment transferred under the agreement, including the propulsion plants and their spare parts. A separate provision has Australia indemnifying both governments and their employees for patent infringement, and neither supplier warrants that what it transfers does not infringe someone else’s patent.

The exclusions are narrow: where a supplier has already been indemnified by a third party, and for an in-service Virginia-class boat up to the point of transfer.

The same article assigns Australia the spent fuel and the radioactive waste from operation, including radioactive waste generated through submarine operations, maintenance, decommissioning, and disposal.

Australia has accepted an uncapped liability to two governments, for reactors it did not design and may not inspect, without a statute, a cap, a fund or an insurer standing behind it. No published document states who would pay a third-party claim, what the ceiling is, or whether one exists. Separately, the head of the submarine programme has declined to say whether the almost $5 billion already paid to the United States would be refunded if no boats were delivered.

7. The maintenance is tiered, and only the top tier is difficult

Most public discussion treats submarine maintenance as one activity. It is six, and they have almost nothing in common. Standards for naval reactors are more stringent than for civil ones, because the plant must survive battle shock, hold to radiated noise limits, sit close to the crew, and accept frequent rapid changes in power.

8. The dates are fixed and they are not the delivery dates

The date that binds is not 2032. It is 2031, because that is when the United States decides whether to sell the first hull, and its law requires Australia to have demonstrated by then that it can fully perform every activity necessary to host and operate the boats.

9. What this list is for

It is not an argument that the task is impossible. Both partners do it, and one of them has run reactors at sea since 1948 without an accident that harmed human health.

It is an argument that the task has not been specified. A programme of this size would ordinarily be preceded by a document of exactly this kind, published, priced and open to submissions. Australia produced seven defence white papers between 1976 and 2016 and none since. What exists instead is a set of announcements: a site here, a boundary there, a designer appointed for a preliminary design, a training academy opening in 2028 for a trade that takes ten years.

Against the list above, the honest summary is that Australia has firm arrangements for berthing, has begun on radiological control, is at preliminary design on the first of two docks, and has named no location at all for weapons storage, for spent fuel, for the reactor compartments, or for the route between the yards and wherever those end up.

The inventory does not settle the argument, and it is worth being exact about that. Two different conclusions can be drawn from the same list. One is that the gaps should be closed: publish the specification, site the weapons depot and the waste facility, design the route, legislate the liability regime, and keep the boats. The other is that a pathway which left this much unspecified for five years was ordered backwards, and that the capability should be re-examined rather than repaired. This memo does not choose between them. It establishes what would have to be true for either.

What the list does establish is narrower and harder to argue with. Australia has not yet shown the shore system its own partners run. Most of what is missing is not a problem of nuclear physics. It is siting, custody, transport, liability and trades — political and industrial problems, which compound if they are left until after the next cheque.

That is the case for a review before the next irreversible commitment, rather than after it.

10. Where the rest of the case is set out

The Wrong Defence Plan — the twenty biggest holes in plain terms, and what is being asked for.

The Holes in AUKUS and the Australian Defence Plan — the full register, 184 entries, 368 sources. Category P covers the bases and facilities, category O the waste and stewardship, category L the reactor.

The AUKUS Cost Blowout — what the programme costs, including the costs nobody totals.

The Island Castle — what actually decides a siege.

AUASIA — what replaces the submarine pathway.

11. Sources

1. Department of Defence, Henderson Defence Precinct programme page: land at Henderson largely owned by the Western Australian Government and managed through leases with multiple entities; the 2023 Defence Strategic Review finding that there is not enough work to sustain the number of shipbuilders located there; $127 million committed October 2024 for planning, $12 billion for early works, about $25 billion indicated over the decade; facilities to be established including contingency docking and depot-level maintenance with graving docks; and the statement that Defence will select an industry partner to create the high-level master plan.
2. Defence Ministers, announcement of the selected site and boundary for the Henderson Defence Precinct, 24 August 2026; and the selection of Leidos Gibbs and Cox Australia to lead the preliminary design of a contingency dry dock, 29 July 2026.
3. Australian Submarine Agency, New Defence precinct at Henderson: the consolidated Commonwealth-owned precinct, and activities across Henderson and HMAS Stirling creating more than 10,000 jobs.
4. Naval News, Australian naval shipbuilding at the Henderson Defence Precinct, July 2026: the Australian Marine Complex from 2003 around the government-owned Common User Facility; Austal at the northern end with six production bays; BAE Systems operating the Marine Support Facility; Civmec’s final assembly hall; and the complex holding Australia’s second largest ship-lift at a nominal 8,065 tonnes.
5. United States Department of Energy, Office of Naval Reactors, The United States Naval Nuclear Propulsion Program, 2025: 97 reactors in operation and over 7,600 reactor-years; two private and four public shipyards with approximately 70,000 employees providing the capability to overhaul, repair, refuel, inactivate and dispose of nuclear-powered ships, described as unique industrial assets with capabilities found nowhere else in the United States; over 800 industrial contractors supporting the programme; deployed tenders and support facilities at major bases; naval standards being more rigorous than civil because components must accommodate battle shock, radiated noise limits, crew proximity and rapid power changes; and the programme’s statutory responsibility under 10 U.S.C. § 6201 and 50 U.S.C. § 2406 for all related facilities, radiological controls, environmental safety and health, and the selection, training and assignment of personnel.
6. United States Department of Energy, Powering the Navy: the statement that naval nuclear-powered ships and support facilities have had no discernible effect on the radioactivity of the environment, and the release of three former shipyards for unrestricted future use.
7. Published descriptions of Naval Submarine Base New London, Groton: 1,102 acres with a 687-acre main base, more than 160 major facilities including eleven submarine piers, homeport for fifteen to seventeen attack submarines, 21,000 personnel and more than seventy tenant commands including the Naval Submarine School and the Naval Undersea Medical Institute; and pier reconstruction from 2022 with platforms up to sixty feet wide.
8. Stars and Stripes, September 2025, and published descriptions of Naval Base Kitsap-Bangor: the Strategic Weapons Facility Pacific, a $294 million underground storage facility for ballistic missiles and warheads opened in 2016; Marine Corps Security Forces Battalion Bangor providing security for submarines and strategic weapons; a $23.9 million armored fighting vehicle support facility with five maintenance bays; dedicated protocols for managing low-level radioactive waste to Department of Energy standards; and coordination with the adjacent Puget Sound Naval Shipyard for complex overhauls.
9. Royal Navy, HMNB Clyde, and published descriptions of the base: the Royal Naval Armaments Depot at Coulport, eight miles from Faslane, responsible for the storage, processing, maintenance and issue of key elements of the Trident system and the ammunitioning of all submarine embarked weapons; up to sixteen reinforced concrete bunkers built into the hillside with two shoreline docks for loading; 43 Commando Fleet Protection Group Royal Marines providing security; approximately 3,000 service and 4,000 civilian staff; a shiplift capable of raising vessels up to 16,000 tonnes; and a £34 million submarine escape training facility.
10. Navy Lookout, Trouble in the docks, May 2026: an active fleet of ten submarines against insufficient available dry docks; just two sites in the United Kingdom able to take submarines out of the water for repairs requiring full access to the hull, hydroplanes and propulsors; the Faslane shiplift able to take only a single boat with no other dry docks at the base, and unavailable for at least some of recent years; and Devonport being no longer an operating base but the main site for nuclear submarine maintenance.
11. Navy Lookout, Rebuilding Faslane, May 2026: the Clyde Transformation Programme, comprising a Clyde Infrastructure Programme begun in 2016 of roughly seventeen discrete projects across Faslane and Coulport, at £1.8 billion and scheduled for completion in April 2032, intended to sustain the deterrent to 2067.
12. Scottish Environment Protection Agency, application by HMNB Clyde for approval to dispose of radioactive waste: two separate letters of agreement, one for solid radioactive waste and one for liquid and gaseous, with the Naval Base Commander as site operator under the Nuclear Installations Act 1965.
13. GlobalSecurity.org, HM Naval Base Clyde: Trident missiles not being serviced at Coulport but returned to the United States Strategic Weapons Facility Atlantic at Kings Bay, Georgia, for reprocessing, an arrangement which has meant the United Kingdom has not had to construct its own reprocessing facilities.
14. Rear Admiral Peter Briggs AO CSC RAN (Ret’d), AUKUS infrastructure, parts 1 to 5, Pearls and Irritations, July and August 2026: the three tiers of maintenance and the missing graving dock; the prohibition on mechanical ship lifts for nuclear overhauls and the abandonment of the Faslane lift concept after the United Kingdom Defence Nuclear Safety Regulator forbade intrusive nuclear work on it; the rejection of floating docks for planned depot overhauls because of hull flexing and the requirement for unbroken, redundantly powered shore cooling of a core generating one to two megawatts of decay heat; a nuclear-capable floating dock requiring seven to ten years and exceeding $1 to $2 billion; the land-tenure logjam at Henderson and the requirement to build replacement commercial facilities along the Western Trade Coast before tenants can be displaced; the absence of an embedded graving dock at Osborne; and the assessment that if Henderson slips past 2032 Washington retains both the legal duty and the political justification to defer or cancel the transfers.
15. Congressional Budget Office analysis, as cited in the above: Virginia-class depot availabilities averaging 760 days in dry dock against United States Navy schedules budgeting fifteen to eighteen months.
16. 22 U.S.C. § 10431, AUKUS Submarine Transfer Authorization Act, and 10 U.S.C. § 8677: the certification requirement that Australia has demonstrated the domestic capacity to fully perform all the associated activities necessary for the safe hosting and operation of nuclear-powered submarines, and the bar on any transfer that would degrade United States undersea readiness.
17. Australian Radiation Protection and Nuclear Safety Agency and the Australian Naval Nuclear Power Safety Regulator: the licensing of a Controlled Industrial Facility at HMAS Stirling in July 2024, and the Regulator’s first substantive licence, issued 12 June 2026 to Australian Naval Infrastructure for site preparation at Osborne.
18. Parliamentary Standing Committee on Public Works, Submarine Rotational Force — West, Priority Works, HMAS Stirling: approval on 30 November 2023 of Early Works valued at $74.9 million and of additional living-in accommodation, associated car parking and the Navy Training Systems Centre — West Stage 2 valued at $212.3 million; the committee report chapter on the Priority Works; and the category of exempted works. Defence records that the project received parliamentary expediency on 22 August 2024.
19. Submissions to the same inquiry, including the concern that the second-stage application to construct a Controlled Industrial Facility at HMAS Stirling preceded disclosure of the results of the Submarine Rotational Force — West Infrastructure Project Community Survey.
20. Australian Strategic Policy Institute, AUKUS spending should be fully accountable at Osborne. It is at Henderson, July 2026, and the Australian Naval Institute: Defence projects above $75 million generally referred to the Public Works Committee, which examines scope, costs, risks and delivery arrangements before significant expenditure proceeds; Henderson delivered through traditional federal capital-spending pathways with business case development, investment assurance and parliamentary scrutiny; Australian Naval Infrastructure excluded under the Public Works Committee Act as a legacy provision from the Australian Submarine Corporation’s exemption on commercial confidentiality grounds; and the description of the Osborne funds, including $4.6 billion announced in July 2026, as going into a black box.
21. House of Commons Defence Committee, AUKUS, First Report of Session 2025-26, HC 841, 28 April 2026: investment of about £4 billion at HMAS Stirling to deliver wharf upgrades and new submarine maintenance, logistics and training facilities, and a new Defence Precinct within the Australian Marine Complex at Henderson at an estimated cost of about £12.5 billion.
22. Australian Submarine Agency and ministerial statements on programme cost: a total of $268 to $368 billion over thirty years, composed of a base estimate of about $244.7 billion and a contingency of about $122.9 billion, being a 50 per cent uplift; payments to the United States submarine industrial base totalling about $3 billion delivered in instalments from February 2025; about $4.6 billion to United Kingdom industry for reactor production capacity at Rolls-Royce; and estimates of $6 to $8 billion for each Australian-built SSN-AUKUS.
23. Strategic Analysis Australia and ASPI The Strategist, 2024 and 2026: the 2026 Integrated Investment Program containing about $425 billion over the decade with maritime programmes taking 41 per cent, a larger share than Army, Air Force, space and cyber combined, on budgets locked decades ahead; the Defence budget at $62.6 billion in 2026-27, 2.02 per cent of gross domestic product on the traditional measure and about 3 per cent by 2033 on the NATO methodology; the absence of any published breakdown of the ten-year or thirty-year figures; the finding that of every dollar of the $53 billion additional decade investment announced only about four cents is appropriated, with the remainder in forward estimates promises, out-year profiles, Contingency Reserve allocations and unspecified alternative financing; and the precedent of the 2009 Defence White Paper, in which roughly $48 billion of committed funding was later abandoned.
24. Pillar 2 funding of $151.6 million over the forward estimates, against the Pillar 1 programme, as recorded in the register at R-12.
25. Agreement among the Governments of Australia, the United Kingdom and the United States for Cooperation Related to Naval Nuclear Propulsion, tabled in the Australian Parliament 12 August 2024: Australia to indemnify both governments against any liability, loss, costs, damage or injury, including third-party claims, arising from nuclear risks connected with the design, manufacture, assembly, transfer or utilisation of any material or equipment transferred, including the propulsion plants and spare parts, subject only to narrow exclusions; a separate indemnity for patent infringement, with no warranty by either supplier that transferred material does not infringe a patent; Australia responsible for spent fuel and radioactive waste from operations, maintenance, decommissioning and disposal; in force to 31 December 2075, terminable on one year’s notice.
26. Organisation for Economic Co-operation and Development, Nuclear Energy Agency, Nuclear Legislation in OECD Countries: Australia: no specific provisions in Australian legislation governing nuclear third party liability, and Australia not a party to any of the conventions on the subject, having signed but not ratified the 1997 Convention on Supplementary Compensation for Nuclear Damage.
27. World Nuclear Association, Liability for nuclear damage: the Price-Anderson Act in force since 1957, providing about US$15.5 billion of cover at no cost to the public or government and without fault needing to be proven, across power reactors, research reactors, enrichment plants, waste repositories and all other nuclear facilities, extended to December 2065.
28. ABC News, 12 August 2024: the head of the AUKUS submarine programme declining to say whether an almost $5 billion payment to the United States would be refunded if no nuclear-powered boats were delivered to Australia.
29. Michael Shoebridge, Strategic Analysis Australia, March 2026: Collins problems compounded by submariner retention and recruitment difficulties such that even when a submarine was available the Navy may not have had a complete crew to operate it; the need to siphon personnel into the nuclear submarine training pipeline meaning those challenges could return to the Collins class; and the Chief of Navy having stated that already, six or more years before Australia receives its first nuclear submarine, more sailors are entering the nuclear submarine pipeline than the Collins pipeline.
30. Andrew McLaughlin, PS News: each Collins-class boat having a crew complement of about 56 sailors and officers, with the Navy having always struggled to fill those ranks, and the Virginia-class submarines the Royal Australian Navy is expected to operate requiring a crew of 132.
31. Naval News and Naval Technology, July 2026: Canada selecting the Type 212CD as preferred supplier for up to twelve submarines, with the manufacturer putting the core crew at roughly 28 to 30, and the Prime Minister stating that only one of Canada’s four Victoria-class boats is currently operational.
32. ASPI The Strategist, February and July 2026: an estimated Australian shortfall of 110,000 skilled trades workers; the absence of a surplus workforce available for mobilisation; and the United Kingdom Defence Committee recommendation of a dedicated AUKUS visa.