This year marks the 50th anniversary of the entry into force of the Treaty on the Non-Proliferation of Nuclear Weapons ^[1] (NPT). The treaty has three separate but inter-related objectives: preventing the spread of nuclear weapons and nuclear-weapon technologies to more countries; promoting cooperation in the peaceful uses of nuclear energy; and pressing the existing nuclear-weapon states to disarm.

With 191 states parties, it enjoys almost universal membership. Today, only India, Pakistan, Israel, North Korea and South Sudan remain outside it. (North Korea did join, in December 1985, but never came into compliance and left in January 2003.) Of course, the first four of those countries have all built nuclear weapons, which severely constrains their options for future membership: the treaty contains no provision under which they could be admitted as nuclear-weapon states, entitled to the same status and privileges as the five officially recognised nuclear powers (the US, Russia, China, France and the UK).

Treaty members will vent about a range of issues during the upcoming ^[2] five-yearly review conference, scheduled to take place in New York from 27 April to 22 May. North Korean nuclear and missile developments, the trials of the Iranian nuclear deal, the collapse of the Intermediate-Range Nuclear Forces Treaty, and the continuing modernisation programs of the five recognised nuclear weapon states—instead of disarmament—provide plenty of new grist for the mill.

Still, every review conference seems to precipitate a sense of unease over the future of the treaty. This year’s no different. Yet the NPT will survive this conference, too, because most treaty members continue to believe that an uncontrolled spread of nuclear weapons would make for a more dangerous international security environment than the one we already have. Moreover, few states would want the NPT to collapse and leave the Treaty on the Prohibition of Nuclear Weapons (which still hasn’t been ratified by enough states to enter into force) in sole possession of the field.

Harder to calculate, and more worrying, is the shifting relationship between the NPT and geopolitics. Let’s see the treaty for what it is—an attempt to freeze the world in the nuclear status quo of 1968, when the treaty was first opened for signature. As such, the NPT represents an attempt to disconnect nuclear weapons from the geopolitical power shifts that have occurred since then.

True, the freeze wasn’t perfectly solid. The treaty gave some wriggle room. Once the treaty entered into force, there were only two ways that a country not previously protected by nuclear weapons could acquire such protection: it could leave the treaty under the ‘supreme national interest’ clause (or never join in the first place) and build its own nuclear arsenal in the teeth of international disapproval; or it could enter into an alliance relationship with one of the recognised nuclear-weapon states and gain the protection of extended nuclear deterrence.

India, Pakistan, Israel and North Korea took the first course. The ‘new NATO’ states took the second. NATO was an alliance of 15 countries in 1968; today it’s an alliance of 29. US extended nuclear deterrence, which covered perhaps 20 countries worldwide in 1968, now covers almost double that number.

Of course, the attraction of the second course turns heavily upon a set of judgements about the credibility and resilience of extended nuclear assurances. For such assurances to be a substitute for a national indigenous program, they have to be more than superficially plausible—and the benchmark tends to climb steeply as the sense of imminent threat grows.

And here we come to the central quandary. Looked at in the broadest sense, the NPT has led something of a sheltered geopolitical existence. Born into a bipolar strategic world, where the bulk of nuclear weapons were in the hands of two risk-averse superpowers, both of which knew well the terrible costs of great-power war, the treaty added an ordering layer to a geopolitical environment in which caution, and recognised spheres of influence, already prevailed.

The treaty was originally intended to last for 25 years, which meant it expired, and was indefinitely extended, in the year 1995—probably the single best year in the last 50 ideally suited to achieving that outcome. The world of 1995, remember, was a unipolar one, where US dominance and the ‘end of history’ thesis suggested global geopolitical convergence.

No such favourable geopolitical environment now exists. Neither bipolarity nor unipolarity has prepared us for the emerging security challenges that now confront us. An NPT which held back proliferation incentives during those earlier, less strategically demanding eras might find it harder to contain such pressures in an age of multipolarity and risk-tolerant actors. And, given the weight it’s now carrying, a US less willing to run nuclear risks on behalf of its allies could throw a serious spanner into the works.

Nuclear proliferation is rare. And it’s always treated on a case-by-case basis. But the four successful non-NPT proliferators have proven that the world has no strategy—save another Desert Storm—to halt proliferation by a determined proliferator.

Nor does it have a strategy for managing proliferation pressures if US extended nuclear deterrence were to falter. The steady expansion of US nuclear commitments in the post–Cold War era suggests a world where the demand for nuclear protection could outrun the supply. Indeed, there are growing concerns that the supply itself might be contracting as US engagement declines—which may explain some of the teasing hints in French President Emmanuel Macron’s recent speech ^[3] on nuclear doctrine, holding out the possibility of closer nuclear links between France and its neighbours.

Can we reinvent the NPT for the world of 2020? Frankly, I don’t see how we do that—not without reopening the whole question of nuclear identity. The NPT forced states to choose their nuclear identity in the geopolitical world of 1968, when only five countries—the Permanent Five of the UN Security Council—had conducted nuclear tests. Repeating the exercise now would open a veritable Pandora’s box. Even Australia, a country patently unenthusiastic about a more densely proliferated world, would probably find a decision now more difficult than the one it took five decades ago.