Australia is set to become the first state without nuclear weapons to acquire nuclear-powered submarines. This presents yet another unwelcome challenge to the nuclear nonproliferation regime, this time by three states that have traditionally been among its greatest supporters.

The foundations of the nonproliferation regime, which aims to prevent the spread of nuclear weapons to states that do not have them, are the 1968 Non-Proliferation Treaty (NPT) and the safeguards system operated by the International Atomic Energy Agency (IAEA). Safeguards entail declarations of nuclear holdings, nuclear material accountancy, containment and surveillance measures, on-site inspections, and increasingly remote sensing and sophisticated data analysis. As the implications of the Australian submarine proposal are studied by governments, nonproliferation experts, and the IAEA itself, the complexities and apparent dangers grow.

Australia, like other non-nuclear weapon states that are party to the NPT, has a Comprehensive Safeguards Agreement (CSA) with the IAEA that permits the agency to verify that Australia is not diverting any nuclear material from peaceful purposes to weapons purposes. Australia has an excellent compliance track record, and there is no suggestion that Australia seeks nuclear weapons. Its proposed submarines would be nuclear-powered but conventionally armed. The difficulty is that nuclear propulsion for submarines is not considered a peaceful use of nuclear energy, but rather a “non-explosive military use.” This is partly because submarine reactors use enriched uranium, either highly-enriched uranium (HEU) that can also be used for nuclear weapons, or low-enriched uranium (LEU) that with further enrichment can become bomb material. It is also because nuclear propulsion technology, including the reactor and fuel design, unlike civilian nuclear power plants, is highly classified. The US Navy regards its nuclear submarine propulsion technology, shared so far only with the United Kingdom, as one of its “crown jewels.”

Australia’s CSA, like all others, seeks to skirt this obstacle by providing, in paragraph 14, that nuclear material may be removed from safeguards for the duration of its use in a submarine reactor and returned to safeguards afterwards. The drafters of this article, widely described as a “loophole” apparently assumed that the state would be enriching its own nuclear material under safeguards and that all the material would be derived from and returned to peaceful uses. Paragraph 14 requires that the state notify the IAEA of its intention, the amount and composition of the material involved, and the estimated duration of its withdrawal from safeguards.

From the little detail we have of the AUKUS proposal, it does not fit the model suggested in paragraph 14. On the contrary, it appears to envisage a “military to military” transfer completely outside safeguards. Australia’s vessels are apparently likely to be based on British or American designs and constructed in South Australia. Imported sealed reactors with “lifetime cores” of HEU would be built into the hull. Australia has no current capacity for designing and building reactors, enriching uranium, or disposing of high-level nuclear waste or spent fuel. Since both the UK and the US are nuclear weapon states, the HEU would be military in origin and not withdrawn from peaceful uses under safeguards. At the end of the 30-year lifespan of the submarines, the reactors would be returned to the UK or US for decommissioning and disposal of the spent fuel outside of safeguards.

A military-to-military transfer to a non-nuclear weapon state conducted completely outside the IAEA verification system would, however, make a mockery of the entire nonproliferation regime, both in logic and in law. Fortunately, paragraph 14 requires a state contemplating a “non-explosive use of nuclear material” to negotiate an “arrangement” with the IAEA to satisfy it that the material is not being diverted to nuclear weapons. Such an agreement appears to require approval by the IAEA Board of Governors, of which Australia is more or less a permanent member, thanks to a request for clarification several years ago from, ironically, Australia. The director general of the IAEA, Rafael Grossi, has already been notified by Canberra of its intentions and has convened a group of his senior safeguards experts to examine the issue. The three partners in the AUKUS arrangement have committed themselves to “the highest standards for safeguards, transparency, verification and accountancy measures to ensure the non-proliferation, safety and security of nuclear material and technology.” Australian officials, along with their American and British partners, will undoubtedly work diligently to ensure that the unprecedented “arrangement” with the IAEA attempts to fulfill these requirements.

Yet because no state has ever triggered the implementation of paragraph 14, this is unknown territory. There is no model for Australia to follow. Canada did begin discussions with the IAEA in the late 1980s before abandoning its submarine plans, but safeguards have become infinitely more complex since then. Even though it is currently building its own nuclear-powered submarine, using its own enriched fuel, Brazil has not yet notified the IAEA of its intentions. South Korea, Japan, and Pakistan have all expressed interest in nuclear-powered submarines. Worryingly, and herein lies the problem, Iran has informed the IAEA that it intends at some unspecified time to acquire them, undoubtedly yet another ploy to justify its suspect enrichment activities.

As the IAEA director general has already warned, verification of the use of nuclear fuel for nuclear-powered submarines will be “very tricky.” Any agreed arrangement will need to provide the IAEA and its member states with sufficient assurance of non-diversion of nuclear material, while also avoiding revealing proliferation-sensitive information to the agency’s inspectors and analysts. Standard onsite inspection techniques applied to land-based nuclear power reactors, including nuclear material accountancy, the application of seals,  and the installation of cameras will be impossible. The US and UK may even be unwilling to declare even gross parameters of the amounts and type of nuclear materials involved. Satellite imagery will help verify that a submarine reactor and its fuel is not being removed while the vessel is in port as HEU-fueled submarines apparently need to be cut open for this to happen. But what occurs at sea is naturally beyond the ken of the IAEA.

Proposals have been made for the AUKUS partners to investigate the use of LEU for Australia’s submarines. This is technically feasible, as China and France use such fuel and Brazil is planning to do so. While this would be preferable from a non-proliferation standpoint in one respect, as LEU is not immediately usable in nuclear weapons, such vessels need to be periodically refueled. A verification scheme for this scenario would be even more challenging, especially if the refueling took place in Australia. For the IAEA, having to devote time, personnel, and resources to devising a suitable scheme for Australia’s benefit when the agency is confronted by the non-compliance cases of Iran and North Korea, along with a host of other challenges, is daunting. The three AUKUS partners should at least pay for the privilege. For Australia there is also an element of moral hazard. In creating a precedent and safeguards model it could be paving the way for the proliferation of nuclear-powered submarines to a wide variety of non-nuclear weapon states. Some of these will have ulterior motives, some will be less than scrupulous in complying with safeguards, and some may be located not far from our shores. Be careful what you wish for.

Trevor Findlay is a principal fellow at the School of Social and Political Sciences, University of Melbourne. His research focuses on global nuclear governance. His latest work, Transforming Nuclear Safeguards Culture: the IAEA, Iraq, and the Future of Nonproliferation, will be published by MIT Press for Harvard in 2022. His research may be accessed online through the University of Melbourne and through the Harvard Kennedy School’s Belfer Center.

This article is published under a Creative Commons Licence and may be republished with attribution.