How can anyone opt for such an uncertain and expensive solution?
That is precisely what strikes me: they did not really choose a ‘solution’ in the sense that a systems engineer would understand it. They chose a strategic direction — SSNs plus highly advanced integration in the United States and the United Kingdom — even though the industrial solution to achieve it did not yet exist.
The Australian press release of 16 September 2021 is almost staggering in hindsight: Canberra announced the abandonment of the Attack-class submarines and the acquisition of at least eight SSNs, then explained that an 18-month consultation period would follow to determine the design, costs, timelines, industrial requirements, training, infrastructure, etc.
In other words, at the very moment they burned their bridges with Naval Group, they still did not know precisely:
- what sort of submarine they were going to build;
- how they were going to fill the capability gap;
- how much it would cost;
- or when they would actually be able to produce an Australian SSN.
This is exactly the opposite of the approach one would expect for a programme of this scale.
But there was a rationale behind it.
1. The choice of nuclear power, however, is perfectly justifiable
For Australia, the SSN is extremely attractive: vast distances, long patrols in the Indo-Pacific, high transit speed, endurance, and the ability to remain in remote areas for extended periods.
This is precisely what Canberra stated explicitly in 2021: the strategic environment was deteriorating much faster than anticipated, and the government believed that conventional submarines were no longer suited to its future needs.
On that point, I do not find the decision absurd. The real question is:
Why was it absolutely necessary to scrap the French programme in order to acquire an SSN?
And that is where the problem begins.
2. They confused the choice of capability with the choice of alliance
The official Australian rhetoric at the time is very revealing. AUKUS is consistently presented in terms of industrial integration, interoperability, technology sharing, AI, cyber, quantum technology and military rapprochement with Washington and London.
The submarine is the first programme of this alliance, not merely a naval procurement.
I therefore believe that the actual line of reasoning went something like this:
China is becoming the main strategic problem → the United States is our essential safeguard → we must integrate much more deeply into the American military system → AUKUS is the means → the SSN is the central tool.
In this line of reasoning, cost and industrial risk become secondary considerations.
This is very different from:
‘We need a nuclear-powered submarine; let’s examine the French, American and British options and choose the best one. This second competition never actually took place.
3. And the Attack-class submarines provided them with a convenient justification
Nor should the French programme be idealised. Even before its cancellation, the ANAO had already identified an overall risk classified as high;
- a nine-month delay in the design phase;
- difficulties in the relationship between Defence and Naval Group;
- an estimated acquisition cost of around AUD 80–89 billion;
- and a risk of a capability gap if the timetable slipped further.
So Morrison could say:
‘The French programme is already expensive and problematic; if we’re going to pay a high price anyway, we might as well go straight to nuclear power.’
Taken in isolation, that’s a valid argument. However, they have replaced a risky programme with one that is infinitely more complex, involving simultaneously:
a complete change in propulsion technology, a nuclear programme that does not currently exist in Australia, a transition spanning several decades, a reliance on US shipyards, a reliance on the UK for the future design, the extension of the Collins-class submarines, and the creation of an Australian SSN industry.
This is where the trade-off becomes far less convincing.
4. The fundamental bias was probably: ‘the Americans will surely find a solution’
And this assumption seemed far less risky in 2021 than it would in 2026. The Australians could reasonably have thought:
the United States has ~70 SSNs/SSBNs, a massive defence industry and a budget of hundreds of billions of dollars; three Virginia-class submarines for Australia surely cannot be an insurmountable problem.
Except that today the Congressional Research Service notes that US shipyards are producing only around 1.1–1.2 Virginia-class submarines per year, whilst US requirements were already based on a target of two per year.
And that is where the entire intermediate framework becomes fragile.
Because the Virginia-class submarines promised to the Australians do not come from a magic stockpile: each vessel must be built or taken from a US fleet that is itself short of them.
5. Once committed to, AUKUS becomes almost impossible to abandon
This is now the ‘lock-in’ phenomenon. Australia has:
- broken ties with France;
- paid compensation;
- invested in nuclear infrastructure;
- begun training its sailors in the United States;
- invested in the US submarine industry;
- prepared Stirling to accommodate US SSNs;
- built its entire doctrine around AUKUS.
To backtrack now by saying:
‘on reflection, it wasn’t a good idea after all’ would carry a colossal political cost.
And every additional dollar invested makes withdrawal even more difficult, both psychologically and politically. It is the classic sunk-cost trap.
The 2026 Australian audit is, in fact, beginning to reveal the collateral damage: following the abandonment of the Attack-class programme, the Collins-class extension programme had to be radically overhauled, costs have risen sharply, and the ANAO now states that Defence is unable to properly demonstrate the value for money of this programme.
What strikes me as extraordinary, then, is this: I would really distinguish between two decisions.
Decision A: ‘Australia must switch to SSNs’.
→ This is perfectly defensible.
Decision B: ‘To acquire SSNs, we must immediately cancel the French programme and commit to AUKUS even before we have defined the industrial pathway’.
→ That one seems much harder to defend in hindsight.
Because there was an obvious question to ask Paris:
‘You are building the Suffren-class submarines. How much would an Australian nuclear-powered version cost, and how long would it take to build?’ Even if the French response had ultimately been unsatisfactory, it would have provided a point of comparison.
They didn’t do that.
And that’s what leads me to believe that the submarine wasn’t really the dominant factor in the decision. The main consideration was geopolitical: to bind Australia much more closely to the United States.
The paradox, five years on, is that this choice – which was supposed to give Canberra greater strategic security – is in fact creating an extraordinary dependence on Washington’s future decisions.
This is probably the fundamental flaw of AUKUS: Australia sought operational sovereignty through the SSNs by choosing an industrial architecture that reduces its strategic sovereignty.
And the more we pull on this thread, the more I find that this — even more so than the 368 billion — is the real problem.
