The Anatomy of a Gridlock: Inside the Final Day of the NPT 2026 Review Conference
How the Conference failed to agree an outcome and and what the resulting gridlock leaves behind
The 2026 NPT Review Conference ended on Friday, 22 May, after four weeks of gruelling diplomatic negotiation, without agreeing on a substantive outcome. For the first time in the half-century history of the treaty’s review process, three consecutive conferences have failed to reach an agreement.
The conference had started with low expectations, partly reflecting the polarised state of contemporary international politics. President Do Hung Viet’s calm, deliberative demeanour, alongside his carefully executed plan of presenting a very early first draft, brought states parties remarkably close to an agreement that many thought was not possible at the outset. Until the very end, an agreement seemed within reach and sharp differences whittled down to a few issues. But as the clock ran out, consensus was ultimately out of reach.
Here is my take on how the final hours unfolded in New York, and what this failure means for the future of the treaty.
Manoeuvres of the Last Conference Day
The final day of any NPT Review Conference is known for its long waits and occasional suspense. Will an agreement be reached? If so, what would it say? If not, who raised objections and on what issues? By this stage, and where there is a draft text that has garnered some level of acceptance as a baseline, the task would be to do the final tweaks and gain the formal acceptance of sceptical voices.
The last day is also when the ultimate decisions on controversial segments are made probably after consulting national capitals rather than by delegations in New York. It is when governments must decide exactly what concessions they are willing to tolerate. All this means that much is happening in the background, feeding into the presidency of the conference, but very little in the plenary until a decision is reached and then announced.
The last day is also when strategies to deflect blame comes to the fore particularly if an agreement becomes increasingly untenable. No delegation wants to be seen as the sole spoiler breaking a consensus. As the end of the conference approaches, the posturing of the early weeks gives way to difficult but also practical decisions about how far a country or a group of countries is willing to go to defend a position.
On the final day, the President also gets to exercise one of their most formidable levers: the ability to formally table a draft text for consideration, forcing states to take a yes/no position on the draft. In a situation where a draft has widespread support, tabling it puts the remaining objectors under the public spotlight, forcing them to formally block consensus in front of the entire international community.
The Midnight Clean-up
At midnight, as the final day of the conference began, President Viet presented a new draft for consideration. It carried the air of near-finality. It was no longer framed as a mere revision of the preceding texts. It was issued under a new code: CRP.4. With only hours before the end of the conference, everyone understood there was very time for any additional major edits that would require another cycle of consultations.
CRP.4 revealed a significant shift from earlier drafts. In its attempt to move closer to consensus, the draft had either purged or watered down paragraphs on topics that had generated heat throughout the conference. The President made the decision that the differences on these issues were so big that they could not be bridged by finessing the language or falling back to general principles. Instead, he decided to drop many issues entirely. Here are some of the notable deletions:
The paragraph on nuclear sharing and extended deterrence arrangements (Paragraph 20 in Rev. 3).
The paragraph on the denuclearisation of the Korean Peninsula (Paragraph 24 in Rev. 3).
The explicit reference to the IAEA Director General’s Five Concrete Principles for securing the Zaporizhzhia nuclear reactor (Paragraph 36 in Rev. 3).
Notably, references to international humanitarian law and the UN Charter were demoted from the operative sections of the document to the preamble.
The section on the Strengthened Review Process, arguably the most significant and action-oriented element of the earlier drafts, was substantially diluted. In previous drafts, the nuclear-weapon states (NWS) were mandated to present their national implementation reports in interactive plenary sessions in each review cycle.
Under CRP.4, this obligation became confined to just “the next review cycle”, the requirement for “interactive discussion” was dropped, and the term “private plenary” was inserted. While the definition of a “private plenary” remained vague, it probably meant excluding international organisations, regional bodies, and civil society.
To many non-nuclear-weapon states, it was difficult not to view these sweep-clean deletions as the result of heavy-handed lobbying by the nuclear-armed states. The frustration in the room was palpable, and it ultimately dampened any enthusiasm for the text.
Crucially, one paragraph in CRP.4 remained heavily contested. It was also the only one presented with brackets in the document, signalling that it was unfinished business. That paragraph dealt with Iran’s nuclear programme:
Underscore the importance of compliance by States parties with the non-proliferation obligations of the Treaty and of addressing all cases of non-compliance with safeguards obligations in a timely manner and in full conformity with the Statute of IAEA and the parties’ respective legal obligations, express serious concern with unresolved non-compliance with obligations under safeguards agreements and its repercussions for the non-proliferation regime, call for a prompt return to compliance through the full, timely and effective implementation of safeguards obligations, and emphasize, in this context, that Iran can never seek, develop or acquire any nuclear weapons.
Throughout the four weeks, the United States had adamantly insisted that a reference to Iran’s safeguards non-compliance be explicitly highlighted in the text. Iran, with equal vehemence, objected. As the final hours ticked away, this paragraph became the key source of friction.
The Plenary & the Final Gavel
In the morning of the final day, the plenary took care of some key procedural business. The dates and venues for the next cycle were set. the Thirteenth Review Conference will convene in New York in 2031, with Preparatory Committee sessions scheduled for Vienna in 2028, Geneva in 2029, and New York in 2030.
The proposal tabled the day before by the Netherlands and other states to alter the cycle, proposing two PrepComs in 2028 and 2029 and holding the next RevCon in 2030, failed to find widespread support. It was a surprising move to table the proposal, as even basic consultations would have revealed a lack of consensus support which would be required for such a proposal to be adopted.
Towards the end of the morning session, the President announced that no consensus existed and therefore more consultations would be needed. This was to be expected.
As the hours dragged on, my colleague and fellow NPT watcher, Gaukhar Mukhatzhanova, and I mused on what kind of ending awaited this conference. We recalled the euphoria of 2010, the last time an NPT Review Conference successfully adopted a substantive outcome document. Back then the General Assembly hall erupted into spontaneous applause. We contrasted that memory with the grim final hours of 2015 and 2022, where clashes on the Middle East and Ukraine masked a wide disappointment with what was widely seen as weak drafts.
When the plenary finally resumed around 5:00 PM, it became obvious that consensus was out of reach. Much of the immediate attention in the room was directed at the clashing views between the US and Iran over the bracketed paragraph.
Confronted with an unbridgeable gulf, President Viet faced a key choice. He could formally table CRP.4 or a modified version of it for a decision by the conference. In doing so, he would be forcing the dissenting delegations (understood to be either the US or Iran depending on the drafting) to take the floor, register their opposition, and take the public heat for blocking consensus. Alternatively, he could withhold the draft and announce that his consultations had yielded no agreement and in the process sparing anyone the public embarrassment.
Ultimately, President Viet chose not to put anyone on the spot and did not put a draft forward. This must have been a disappointment for Ambassador Viet, who throughout the conference ran proceedings with a collected and calm poise while maintaining strong control on the drafting pen. It was not difficult to feel the disappointment behind his diplomatic language as he opened the evening session to announce the result of his deliberations.
The final reveal came when he declared:
Despite our best efforts, however, it is my understanding that the conference is not in a position to achieve agreement on its substantive work. Given that this is now apparent, I do not intend to put the document forward for adoption as the conference is not in a position to achieve agreement.
With those words, he sealed the fate of the 2026 Review Conference. There was no substantive outcome document, but also no single, clear target to blame; at least straight away as the conference closed.
Ultimately, and aside from US-Iran differences on the wording of that paragraph, many unresolved differences existed that could have also torpedoed the delicate compromises that President Viet built in his draft. Yet the draft, in its latest watered-down form, had also triggered widespread indifference about its adoption among many of the traditional supporters of the NPT. Everyone was disappointed, but no one was shedding any metaphorical tears.
What Does This Mean for the NPT?
It will take time to fully digest what this third consecutive failure means for the treaty and the broader non-proliferation and disarmament regime.
In the lead-up to the conference, some have tried to redefine what “success” of a review conference might look like. The argument was frequently made that we do not actually need a consensus outcome document, and that its absence does not affect the daily technical operation of the treaty.
I believe this self-soothing argument is misguided, and it papers over some emerging hard truths about the treaty and the regime built around it.
The three consecutive failures to agree on an outcome document, or even see eye-to-eye on the basic principles around what the treaty is about, what its current challenges are, and how it can be sustained, all point to deepening fissures within the treaty.
Treaties do not derive their strength merely from their letter or their history. They derive it from the political authority they hold and the continued acceptance and investment by states parties. By failing to bridge these divides once again, the 2026 Review Conference leaves behind a weakened treaty and a creeping uncertainty about whether global nuclear politics can still be managed collectively.