I am not alone in finding this hypothesis highly improbable, even far-fetched. Here is the comment of the Australian specialist blogger (Pete)
Submarine Matters:
Australia buying 3-5 Virginias? Intentionally Unworkable. Pete Comment
Reports from various media outlets support Submarine Matters prediction, here and here that Australia will/may eventually buy UK designed SSNs from the UK and/or to build them in Australia.
But what Submarine Matters didn't predict is the variation below that Australia might also buy from 3 to 5 Virginia's during the 2030s as submarine capability gap fillers.
The article below, with similar/same information from numerous news outlets the original "exclusive", then here and the best article here etc. looks credible in terms of what President Biden and Prime Ministers of the UK Sunak and Australia Albanese will say at the AUKUS summit in San Diego next week.
But there are too many problems with a "quick solution" to buy Virginias.
US Political Change of Heart?
A good intention to sell Australia Virginias in 10 or more years is a long time in AUKUS politics, where political continuity is essential, but unlikely. Biden will not be President in the 2030s.
Even before the 2030s there is uncertainty over Biden being re-elected in November 2024. Might a Republican President (at worst the alliance downgrader Trump?) have different priorities, which might include bowing to already expressed USN pressure not to sell Virginias to Australia?
Biden has simply made a promise he cannot possibly honour during his sitting presidential term.
The USN being a parent navy that is already reluctant to spare Virginia's for Australian use spells big problems down the line. The USN has said it already has too few submarines to handle Russia and the increasing commitments against rising Chinese and North Korean submarine threats. Nuclear novice Australian crewman being trained in the USN system might also be seen by the USN as a resource problem.
The USN already has a long backlog of nuclear subs that need deep maintenance. Australian Virginia's with major problems like training accident eg. damage from seafloor collisions that are more likely in very fast subs the RAN is unfamiliar with, are likely to languish at the back of the queue, behind USN Virginias that need major work.
Analysts have already reported it would take 15 years to train a whole Australian nuclear submarine crew. Also a Virginia needs a vastly larger crew of 135 compared to a likely 100 in SSN(R)s and only 58 in our present Collins subs. Even if we bought Virginia's in 2033 we could not majority crew them with Australians until 2038. This all points to years of reliance on US officers and crewman - something the USN wouldn't be happy about unless the USN remained de facto owners of "Australia's" Virginias.
Another concern is the irrationality of Australia operating US and UK designed SSNs simultaneously. It may well be that after Australia buys 5 Virginias it would make more sense to buy 3 additional Virginias to makeup the complete and final fleet of 8.
The alternative of building 3 UK SSN(R)s in Adelaide, to make up the 8, just doesn't make financial, training or operational sense.
Early Virginias are an Intentionally Unworkable Non-Solution
It is as if the federal Australian Government, Biden, the USN and US nuclear sub industry have intentionally come up with a solution that the whole submarine building industry in Australia (not only in Adelaide) won't accept. It is taken as given the US won't allow Virginias to be built in Australia. But "Build our submarines in Australia", from the Collins onwards, has always been the essentail requirement for our Australia-wide shipbuilding interests.
Australian shipbuilding industry and South Australian State Government resistance to Australia buying already built Virginias from US shipyards will be such that Albanese can then honestly claim "Oh well! Then we'll need to wait for SSN(R)s to be fully designed and built".
Therefore Australia's only real option is to wait until the UK, following its Dreadnought-class SSBN build (which will dominate UK resources until the late 2030s) is ready to build SSN(R)s in the UK and in Australia. For that we are talking much mid-2030s design work, then one or two SSN(R)s laid down in the the UK in about 2039 and maybe the first Australian one laid down in Adelaide in 2042. Commissioning of that first Australian SSN(R) may be 2047. These are time spans so distant that the Albanese government cannot be held to account during the current Parliament and even the next. The Albanese government will be long gone before about 2035 when Australia will need to start tooling up for SSN(R) production. /end