On Wednesday, Australia, the United Kingdom, and the United States announced the formation of a new trilateral security partnership, AUKUS. Its first project will be “to deliver a nuclear-powered submarine fleet for Australia.”

This initiative was hailed in the three countries themselves and in the Asia-Pacific—at least outside of China, whose growing power AUKUS seeks to offset. Nuclear-powered submarines offer various military advantages over the diesel-powered submarines that Australia had been planning to buy from France. More generally, AUKUS is seen in the Asia-Pacific as a steely manifestation of the United States’ commitment to the region.

Discussion of the nonproliferation implications of AUKUS has been largely relegated to a handful of nuclear experts who, like debate nerds at a high school party, are soberly arguing over the finer points of nonproliferation policy on Twitter while everyone else gets intoxicated on grand strategy.1 Among those experts, there is far from a consensus on how serious the nonproliferation implications of AUKUS are or, indeed, whether they are negative at all.

James M. Acton
Acton holds the Jessica T. Mathews Chair and is co-director of the Nuclear Policy Program at the Carnegie Endowment for International Peace.
More >

In my assessment, the nonproliferation implications of the AUKUS submarine deal are both negative and serious. For Australia to operate nuclear-powered submarines, it will have to become the first non-nuclear-weapon state to exercise a loophole that allows it to remove nuclear material from the inspection system of the International Atomic Energy Agency (IAEA). I have no real concerns that Australia will misuse this material itself, but I am concerned that this removal will set a damaging precedent. In the future, would-be proliferators could use naval reactor programs as cover for the development of nuclear weapons—with the reasonable expectation that, because of the Australia precedent, they would not face intolerable costs for doing so.

It is important to be honest about this negative consequence—even if Canberra, London, and Washington conclude that the military and strategic benefits are greater still—because it can be mitigated. First, however, it is necessary to explain why the AUKUS submarine deal will inflict serious damage on the nonproliferation regime. 

The Politics of Nonproliferation Enforcement

The Nuclear Non-Proliferation Treaty (NPT) aims to prevent nuclear proliferation without interfering with states’ “inalienable right” to benefit from nuclear energy’s peaceful applications. To this end, it requires non-nuclear-weapon states to declare all their nuclear material and tasks the IAEA with verifying that none of this material is used to build nuclear weapons. The goal of this system of declarations and inspections, known as safeguards, is deterrence—that is, to lead a would-be proliferator to conclude that, because of the risk of detection, the costs of trying to acquire nuclear weapons would outweigh the benefits.

Deterring proliferation, however, requires not only that a would-be proliferator believes its nefarious activities will be detected but also that it will suffer significant consequences for them. IAEA safeguards are not self-enforcing. There are no automatic consequences for noncompliance (except referral to the UN Security Council, which is under no obligation to act). As a result, the efficacy of IAEA safeguards in preventing proliferation hinges on the willingness of the international community as a whole and of individual states to enforce the rules.

The credibility of this enforcement process is of profound importance to the security of the United States—particularly in preventing proliferation to adversaries or potential adversaries. Washington has many tools to prevent allies from proliferating. Most importantly, it can offer security guarantees to allies, posture U.S. forces to defend them, and build up their own capabilities (as it is doing with Australia). It is also well-placed to bribe, threaten, and cajole allies not to proliferate.

However, Washington’s options to prevent adversaries from proliferating are more limited. It has toppled dictators in Libya and Iraq who had given up (voluntarily or otherwise) nuclear-weapon programs, and so provided adversaries with a potentially strong motivation to proliferate while undermining the credibility of potential U.S. promises not to invade them should they refrain from doing so. Meanwhile, export controls—regulations on the international commerce in sensitive technologies—are declining in effectiveness. As Scott Kemp notes, “the technologies needed to make nuclear weapons have remained static, whereas the indigenous capabilities of states have steadily grown over the last half-century.”

When the IAEA detects outright noncompliance with safeguards or even just more ambiguous forms of worrying behavior, an intensely political process is set in motion. Key states—Security Council members, in particular—can disagree about whether enforcement actions are needed and about what kind of response is appropriate. For many states, an important consideration is whether enforcement is seen as entrenching a double standard.

The United States doesn’t get hurt by double standards. Other states lack the power to create rules that would prevent Washington from developing any technology that it doesn’t yet have. As a result, U.S. officials and analysts don’t worry much about double standards. Their foreign counterparts, however, have good reason to be more concerned about them. 

Following the discovery of Iran’s clandestine nuclear program in 2002, for example, former U.S. president George W. Bush’s administration called on Tehran to “abandon uranium enrichment and reprocessing activities.” Looking at the world through (at the time) British eyes, I thought this demand made sense. A nuclear-armed Iran would pose a real threat to the security of the United States and its allies, including the United Kingdom. This outcome could be prevented if Iran abandoned enrichment and reprocessing. It wasn’t a concern to me that other non-nuclear-weapon states—like Japan, Germany, and the Netherlands—had enrichment programs. They didn’t seem likely to flout international nonproliferation rules and, even if they did, their acquisition of nuclear weapons wouldn’t feel threatening to me.

This logic wasn’t persuasive for most of the world, however. Iran asserted its “right to enrich” and exploited a perceived double standard by comparing itself to Japan. This campaign was successful. Few other countries, even in Europe, were willing to support calls for Iran to abandon its enrichment and reprocessing programs. Germany, in particular, strongly resisted pressuring Iran to abandon its enrichment program—probably because it enriched uranium itself and feared it could come under similar pressure in the future. Eventually, in 2006, the United States resigned itself to supporting a Security Council resolution that demanded that Iran “suspend,” rather than abandon, enrichment and reprocessing. (In spinning this outcome, John Bolton, then U.S. ambassador to the UN, insisted that the suspension be “sustained,” though the resolution’s text said no such thing.)

In short, key states—including close U.S. allies—are less likely to respond robustly to proliferation threats when doing so would entrench a double standard. Understanding this, potential proliferators may exploit such double standards to further their pursuit of nuclear weapons.

Many Americans may think, “So what? If the Security Council won’t enforce nonproliferation, the United States will—unilaterally, if necessary—through sanctions or military action.”

This reaction, however, is shortsighted. U.S. sanctions are more likely to be effective if they attract widespread support. In fact, if they are perceived as illegitimate too frequently, an alternative to the dollar as the global reserve currency is likely to emerge over the medium term, severely degrading the effectiveness of U.S. unilateral sanctions. Military action also becomes less costly and thus more likely if it is seen as legitimate—not least because other countries become more likely to participate. For reasons of self-interest, therefore, the United States should want to avoid creating new double standards that could affect nonproliferation enforcement.

The Naval Propulsion Loophole

This (finally!) is where the AUKUS submarine deal enters the equation. Specifically, it will create a new double standard—or, at least, severely exacerbate an existing one—that a would-be proliferator could exploit to build nuclear weapons.

The NPT does not prohibit non-nuclear-weapon states from building or operating nuclear-powered ships. However, for practical reasons, the IAEA cannot safeguard naval reactors (especially on submarines whose locations are kept secret and which are essentially inaccessible because they are, well, sub marine). As a result, the Comprehensive Safeguards Agreement (the IAEA’s basic safeguards agreement) permits non-nuclear-weapon states to withdraw nuclear material from safeguards for use in a “non-proscribed military activity,” that is, naval reactors.

This is a glaring and worrying loophole in IAEA safeguards. To date, however, naval reactors have been operated only by nuclear-weapon states and by non-signatories to the NPT. As a result, since the Comprehensive Safeguards Agreement was first drafted nearly fifty years ago, no non-nuclear-weapon state has ever exercised this loophole (though some, including Brazil and Canada, seriously considered it).

Nonetheless, there has always been the danger that this loophole could be exploited by a would-be proliferator. Indeed, in 2018, Iran informed the IAEA it was planning “to construct naval nuclear propulsion in [the] future”—creating an obvious pretext to remove nuclear material from safeguards in the future. That said, had Iran actually taken this step, I believe it would have faced a fierce backlash, including from China and Russia. Indeed, the expectation of such a backlash, which could include new sanctions or even military action, has probably helped to deter Iran from acting on its implied threat—so far, at least.

With the announcement of the AUKUS deal, this potential backlash would likely be somewhat weaker. If Australia ultimately withdraws nuclear material from safeguards, the backlash would probably be even weaker still. In the future, therefore, a state that wants to acquire nuclear weapons—Iran or otherwise—could reasonably calculate that, by using the Australia precedent, it could reduce the costs of removing nuclear material from safeguards to a tolerable level.

The comeback to this—that “Australia is not Iran or Brazil”—is both entirely fair and also, when it comes to international diplomacy, largely irrelevant. Much of the rest of the world will see a double standard in encouraging Australia to exploit the naval propulsion loophole while trying to punish a U.S. adversary for doing the same. It will, therefore, resist strong international enforcement actions and oppose the United States’ own unilateral efforts. The result is that the AUKUS deal is likely to weaken the deterrence value of safeguards and make proliferation more likely.

Mitigating the Damage

I have concluded that the nonproliferation costs of the AUKUS deal outweigh the military and strategic benefits. I’m also realistic enough to understand that, in Australia, the United Kingdom, and the United States, few others feel the same way. Nonetheless, Canberra, London, and Washington should try to mitigate the nonproliferation costs. Here are two ways to do so.

First, Australian nuclear-powered submarines should be fueled by low enriched uranium (LEU), which cannot be used directly in nuclear weapons, rather than high enriched uranium (HEU), which can be. Although the three states say that they still need to work out all the details, it is crystal clear that Australia’s submarines are currently on track to use HEU. Both the United Kingdom and the United States fuel their naval reactors with HEU. Moreover, Canberra has indicated that its reactors will not need refueling—which, again, strongly suggests the use of HEU.

Some analysts have argued that the use of HEU is preferable for nonproliferation precisely because it precludes the need for refueling. I’d probably agree with them if I were worried about Australia proliferating (which I’m not). Instead, I’m concerned about precedent. And removing HEU—which can be used directly in nuclear weapons without further enrichment—from safeguards sets a significantly worse precedent than removing LEU.

Because neither the United Kingdom nor the United States currently has an LEU-fueled naval reactor design, France should be invited to join AUKUS. French naval reactors are fueled with LEU and could be used to power Australian submarines.

There would be disadvantages to such an arrangement. French reactors need refueling periodically—every ten years or so—though that seems very manageable. More importantly, four-way industrial collaboration would be extremely complicated. Indeed, the additional complications of including France would probably be prohibitive if the alternative was to supply Australia with off-the-shelf U.S. or British submarines. In practice, however, the politics of the situation will demand that Australia’s submarines use a mix of U.S. and British technology. With the United Kingdom and the United States merging their own designs, it should be possible to also integrate a French reactor. (French involvement would also help rebuild bilateral relationships that, following the announcement of the AUKUS deal, have deteriorated to the point where Paris recalled its ambassadors to Australia and the United States.)

Second, the United Kingdom and the United States should set out objective criteria that they believe should be met for non-nuclear-weapon states to have unsafeguarded naval reactors. Doing so now would help combat accusations of double standards later, should a state with weaker nonproliferation credentials than Australia seek to exploit the naval reactor loophole. Specifically, London and Washington could argue that a non-nuclear-weapon state party to the NPT should be permitted to withdraw nuclear material from safeguards for use in naval reactors only if it has:

  • a Comprehensive Safeguards Agreement and an Additional Protocol (a strengthened safeguards agreement) in force;
  • received its broader conclusion from the IAEA (that is, the IAEA has concluded that there is no undeclared nuclear material in the state);
  • not been found in noncompliance with its IAEA safeguards agreement for fifty years;
  • not been the subject of a state specific report about safeguards compliance from the IAEA director general to the Board of Governors for at least twenty-five years;
  • signed and ratified the Comprehensive Test Ban Treaty and the Convention on the Physical Protection of Nuclear Material and its amendment;
  • committed to fuel its naval reactors with LEU;
  • committed not to enrich uranium or reprocess spent fuel domestically for as long as it operates unsafeguarded naval reactors;
  • concluded an agreement for the supply of fabricated naval reactor fuel by a nuclear-weapon state; and
  • committed to allow full IAEA safeguards on irradiated naval fuel as soon as it is removed from a reactor.2

Australia does not meet these commitments today—though it should be able to. Canberra has never pledged not to enrich uranium domestically. However, because it does not do so and has no plans to do so, there’s no reason it can’t make such a promise. (If you’re Australian, think very carefully before objecting to this condition as a double standard.) Moreover, the AUKUS partnership still has to work out the details of reactor design, fuel supply, and spent fuel management. Providing Australia with an LEU-fueled reactor would certainly be challenging, but it should be possible with French involvement.

It is important to be clear-eyed here, however. These two proposals would only go so far. They would not be enough to prevent the AUKUS submarine deal from establishing a bad precedent that could be exploited by future proliferators, but they would mitigate its most damaging aspects.


1 No offense is intended to debate nerds. I was one and never got invited to high school parties so hold debate nerds who were in high esteem.

2 It should be permissible to terminate such safeguards if and when the fuel is transferred to a nuclear-weapon state.