The NPT Is Broken. What Comes After Multilateral Nuclear Order?
Eight people died in a Chinese coal mine gas explosion on 22 May 2026, hours before diplomats at the United Nations issued a terse communique confirming what most observers had expected: the triennial review conference of the Nuclear Non-Proliferation Treaty had failed to produce a substantive agreement. The timing was coincidental. The contrast was not.
The NPT, opened for signature in 1968, rests on a grand bargain: five nuclear-weapons states would work toward disarmament while the remaining 190-odd signatories forswore acquiring weapons of their own. In exchange, they received access to civilian nuclear technology under International Atomic Energy Agency safeguards. That bargain has been fraying for decades. The 2026 collapse suggests it may be finished as a functioning diplomatic framework.
The immediate cause is familiar. The treaty requires consensus among all participating states. Any single participant can block an outcome. In practice, that veto has long belonged to nuclear-armed states unwilling to accept binding timetables for disarmament, and non-nuclear states unwilling to accept indefinite exceptions for those who already have weapons. The conference floor has become a ritualised theatre of mutual accusation, each side holding positions calibrated for domestic consumption rather than negotiation.
What has changed is the surrounding environment. The 2022 Russian invasion of Ukraine, the first large-scale conventional war in Europe since 1945, reframed nuclear weapons in the minds of most NATO-aligned governments from instruments of strategic anachronism back into tools of coercive diplomacy. Moscow's explicit nuclear signalling — threats issued openly, tactical systems placed on alert — demonstrated that even a diminished Russian arsenal retains its deterrent function in the most brutal possible sense. Meanwhile, China's nuclear build-up has accelerated to a pace that US intelligence assessed in 2025 as far exceeding prior projections, and North Korea has conducted four successful ICBM-range tests since January 2026.
Western delegations arrived in Vienna this year under instructions to address these realities. Non-NPT members — North Korea, India, Pakistan, Israel — are expanding and diversifying their arsenals without legal constraint. The treaty's non-proliferation architecture offers no leverage over them. The disarmament provisions offer no timeline that nuclear states will honour. What remains is the fiction that the NPT contains the problem when it demonstrably does not.
The conference failure exposes a structural tension the treaty was never designed to resolve: the legal asymmetry between the five recognised nuclear powers and everyone else. Non-nuclear states have complied with the bargain. The nuclear states have not disarmed. For fifty years, this asymmetry was papered over with procedural language, confidence-building measures, and the shared assumption that gradual progress was better than no progress at all. The assumption is no longer shared.
A bloc of non-nuclear states, led diplomatically by Brazil, South Africa, and a coalition of smaller nations, pushed this year for a formal commitment to a nuclear weapons convention — a binding legal instrument requiring the elimination of all arsenals. This is the position the International Court of Justice endorsed in 1996 as a legal obligation. It has never been closer to realisation, and it has never been further from political possibility. The nuclear five will not accept mandatory timetables. The conference cannot proceed without them.
The realignment of global nuclear politics is not simply a product of great-power rivalry. It reflects a deeper question about what nuclear weapons are for. During the Cold War, the logic was MAD: mutual assured destruction made them too dangerous to use, so their only function was deterrence. That logic held because both sides believed it. Today, Russia has shown that nuclear coercion — threatening use without actually using — is a viable instrument against a non-nuclear adversary backed by a nuclear alliance. The lesson will not be lost on other states calculating whether a bomb would alter their own security calculus.
What this means, concretely, is that the non-proliferation regime is bifurcating. Inside the NPT, compliance remains near-universal among non-nuclear states — an achievement that should not be dismissed. Outside it, the number of nuclear-armed states is growing, their arsenals expanding, and the legal architecture governing their behaviour increasingly vestigial. The treaty's defenders argue that containing the club to five members for fifty years was always the point, and that the NPT has succeeded at exactly that task. Critics note that the five have used their legal status to legitimise the weapons themselves rather than working to eliminate them.
Both arguments are correct. That is the problem. A treaty that legitimises five arsenals while demanding their eventual abolition, but provides no mechanism to achieve that abolition, is a device for managing appearances rather than changing facts. The 2026 conference did not fail because anyone wanted it to fail. It failed because the premises underlying the bargain — that nuclear powers would negotiate genuine reductions, that the non-proliferation norm was self-reinforcing, that time was on the side of disarmament — have all been proven wrong.
The question now is not whether the NPT can be repaired. It cannot, in its current form, because its foundational compromise is no longer politically sustainable. The question is what architecture might replace it, or whether the world will simply manage an increasingly ungoverned nuclear landscape through bilateral deterrence relationships, extended nuclear guarantees, and quiet acceptance that some states will have bombs and some will not, with the distinction determined by geography and alliance rather than law.
For the states that have abided by the treaty faithfully — and paid the price of foregoing their own deterrent — that outcome represents a profound inequity. The conference's failure is not, in the end, about process. It is about whether the international system still believes that disarming nuclear states is a serious goal, or whether it has quietly accepted that nuclear weapons are permanent features of international order, to be regulated but not eliminated. The delegates went home without an agreement. The bombs did not go anywhere.
This publication chose to frame the NPT collapse as a structural failure of multilateral governance rather than as a news peg about diplomatic disappointment. The dominant wire framing treated the conference outcome as a procedural story. We think the underlying story is larger: a treaty regime whose premises have been overtaken by geopolitical reality, and whose collapse is accelerating the very proliferation it was designed to prevent.