Opinion: The nuclear order is fraying. Abandoning it would be worse

International Atomic Energy Agenc director general Rafael Mariano Grossi addressing the Preparatory Committee for the 2026 Review Conference of the Parties to the Treaty on the Non-Proliferation of Nuclear Weapons

In May, delegates from 191 countries gathered at the United Nations for the eleventh review of the Treaty on the Non-Proliferation of Nuclear Weapons. After four weeks of negotiation, they went home without a consensus outcome document – the third such failure in a row. The review conference president, Viet Nam's Do Hung Viet, put four successive drafts before delegations. None survived contact with the room.

The failure lands at a genuinely dangerous moment – and nowhere more so than in the Asia-Pacific, which has quietly become the world's most consequential nuclear region.

China's warhead count has climbed from around 500 in 2023 to roughly 620 today, with a trajectory toward 1,000 by 2030. North Korea has tested more than 100 missiles since 2022 and, by US intelligence estimates, now holds fissile material sufficient for 60 to 90 warheads. The United States is modernising its Pacific-deployed bomber and submarine forces, and for the first time in decades Washington is publicly discussing nuclear-sharing arrangements with Seoul.

According to the Stockholm International Peace Research Institute's 2026 Yearbook, all nine nuclear-armed states expanded or modernised their arsenals last year – but it is in Asia that the growth curves are steepest and the diplomatic guardrails weakest.

 It would be tempting, watching this unfold, to conclude that the non-proliferation regime built around the NPT has run its course. That temptation should be resisted.

The treaty remains the most widely joined arms control agreement in history, and its track record, however imperfect, is real: it helped forestall the "world of 25 or 30 nuclear-weapon states" that President Kennedy warned of in the 1960s, established the inspection architecture that still gives the world its clearest picture of who is building what, and created the diplomatic scaffolding that non-nuclear states have used for half a century to hold nuclear powers accountable.

Only four of the treaty's 11 review conferences have ever produced consensus documents – and yet the system has survived.

What has changed is the environment around it? Four trends are converging in ways the NPT's architects never anticipated – and all four are now most acute in the Asia-Pacific.

The first is the collapse of restraint among the great powers.

United States President George H. W. Bush and Russian President Boris Yeltsin sign the Start II Treaty in Moscow, Russia, 1993 (Image: Wiki Commons; George W Bush Library)

The New START Treaty between the United States and Russia lapsed in February, leaving the two largest nuclear arsenals – together still more than 80 percent of the global total – unconstrained by any bilateral agreement for the first time since 1972. But the more novel gap is between Washington and Beijing.

There is no equivalent of Cold War hotlines or pre-launch notification agreements linking the American and Chinese nuclear establishments.

Senior military-to-military contact has been sporadic at best. In a region where US and Chinese forces now operate in close proximity around Taiwan, the South China Sea and the Korean Peninsula, that silence is itself dangerous: misperception has fewer guardrails to catch it.

The second is the erosion of the taboo on nuclear testing.

No country besides North Korea has tested since 1996, but that informal norm is fraying.

The United States has accused China of covert low-yield testing at Lop Nur in 2020, and Washington has floated resuming its own tests to keep pace with Moscow and Beijing.

Whether or not any of these claims prove accurate, the return of testing to serious policy discussion is itself a regression – and one felt most acutely in places like the Marshall Islands, Kiribati and French Polynesia, which still live with the legacy of more than 300 Cold War-era tests conducted on Pacific atolls.

The United States carried out nuclear tests on Bikini Atoll in the Marshall Islands between 1946 and 1958, causing widespread radioactive contamination and displacing the local population

For Pacific Island nations now negotiating expanded security partnerships with the United States, China and Australia, that history makes any weakening of the testing moratorium an immediate domestic concern.

The third is the entanglement of nuclear forces with new technology.

Russia's nuclear doctrine, revised in November 2024, explicitly lowered the threshold for nuclear use in response to conventional attacks on strategic infrastructure.

Seven months later, Ukrainian drones struck Russian nuclear-capable bomber bases at Belaya and Engels in an operation that, by Moscow's own doctrine, could have justified a nuclear response.

Russia did not retaliate with nuclear weapons, but the episode illustrates a structural risk that will only grow as autonomous systems, AI-enabled early-warning networks and cheap drones proliferate: compressed decision timelines that leave less room for human judgment – the kind of judgment that, in 1983, led a single Soviet officer, Stanislav Petrov, to conclude that a satellite warning of an American missile launch was a false alarm rather than relay it up the chain of command.

That risk is intensifying fastest in Asia. China is building early-warning satellite constellations and shifting part of its arsenal to higher-alert postures. North Korea's growing inventory of solid-fuelled missiles shortens launch preparation times. Taiwan has expanded its indigenous drone programme, and Japan is developing counter-strike capabilities that could, in a crisis, blur the line between conventional and strategic targets.

The region's geography, characterised by short flight times, overlapping missile ranges, dense concentration of forces, all compresses decision windows further still.

The fourth is the quiet unravelling of confidence in extended deterrence – the promise that nuclear-armed allies will defend non-nuclear partners under their security umbrella.

Nowhere is this more visible than in South Korea. The Asan Institute's 2025 poll found a record 76.2 percent of South Koreans support their country developing its own nuclear weapons, with 66.3 percent favouring redeployment of American tactical warheads to the peninsula.

These are not soft preferences: follow-up questions show that a majority would still back nuclear armament even if it meant international sanctions or increased defence cost-sharing.

The Washington Declaration of 2023, which created a Nuclear Consultative Group and promised Seoul greater visibility into American nuclear planning, has not closed the gap. Persistent North Korean missile tests, growing scepticism about long-term American commitment, and the Trump administration's transactional approach to alliances have kept the domestic debate alive.

 A South Korean breakout would be catastrophic for the NPT: it would mark the first defection by a technologically advanced, treaty-compliant democracy and it would almost certainly trigger reconsideration in Japan and Taiwan.

The threat is not imminent, but the norm against allied proliferation, once assumed to be settled, is now visibly eroding – and the Asia-Pacific is the region where that erosion could cascade fastest.

 None of this means the grand bargain at the NPT's core – that non-nuclear states forgo weapons, nuclear states pursue disarmament, everyone gets access to peaceful nuclear technology – still satisfies anyone fully. It doesn't.

Disarmament has stalled badly enough that a majority of the world's non-nuclear states have thrown their support behind the Treaty on the Prohibition of Nuclear Weapons, a ban treaty now backed in some form by roughly 99 countries.

Global nuclear weapon numbers 2023 (Image: Wiki Commons; Stockholm International Peace Research Institute)

In the Pacific, the Treaty of Rarotonga has kept the South Pacific a nuclear-weapon-free zone since 1985, and its members have been among the most vocal advocates for the prohibition treaty. No nuclear-armed state will join the ban, and it will not eliminate a single warhead on its own. But its growth is a signal the NPT's nuclear-weapon states cannot ignore indefinitely.

 Given all this, the temptation to declare the regime broken and start over is understandable. But it is wrong.

Building a new treaty architecture from scratch, in today's fractured geopolitical climate, is far harder than sustaining an imperfect one.

The more realistic path is adaptation on several fronts simultaneously:

• getting nuclear-armed states to publicly recommit to the testing moratorium at leaders' level

• keeping risk-reduction channels open through track-two and military-to-military dialogue even where formal arms control has stalled

• extending robust safeguards to new arrangements like the AUKUS submarine deal, so that legitimate technology-sharing doesn't become a template for future loopholes

• investing now in the verification and monitoring capacity the International Atomic Energy Agency will need as nuclear energy expands and weapons technology evolves.

 Non-nuclear states have shaped this order before, and they can again – not by matching nuclear arsenals, but by defending nuclear-weapon-free zones like Rarotonga and Bangkok, using collective diplomatic pressure at the UN and NPT review cycles, and pushing domestic audiences that remain under-informed about nuclear risk to pay attention.

The regime on offer is not the one anyone would design today. It is, however, the one we have, and letting it collapse would leave the world, and the Asia-Pacific in particular, with nothing in its place at precisely the moment restraint matters most.

About the author

Dr Julia Macdonald is director of research and engagement at the Asia New Zealand Foundation.

Prior to joining the Foundation she held senior roles in government and academia.

In government, Julia worked for the Department of Prime Minister and Cabinet, leading work on New Zealand’s first National Security Strategy, and before that at the New Zealand Ministry of Defence.

From 2017-2021, Julia was an Assistant Professor at University of Denver’s Korbel School of International Studies where she taught and published on a range of international security issues. 

The Foundation's Asia in Focus initiative publishes expert insights and analysis on issues across Asia, as well as New Zealand’s evolving relationship with the region.