Two dates mark a visible acceleration in the deterioration of the post‑Cold War arms‑control framework.
On 5 February 2026, New START expired. For the first time in more than half a century, no binding treaty limits the strategic arsenals of the United States and Russia — together accounting for around 90 percent of the world’s roughly 12,241 warheads (SIPRI 2025). Weeks earlier, the Bulletin of the Atomic Scientists had moved its Doomsday Clock to 85 seconds to midnight, its most alarming setting.
On 22 May 2026, the eleventh NPT[1] Review Conference concluded in New York without a consensus document — for the third time in succession, following 2015 and 2022. Conference President Do Hung Viet acknowledged that ” State Parties need to take the three consecutive failures very seriously if they want to preserve this regime” and chose not to put the final draft to a vote. The next review cycle will not begin in earnest until preparatory meetings are expected to begin around 2028, with the next Review Conference scheduled for 2031.
These two dates are not isolated shocks, but milestones in a longer trajectory in which bilateral limits, multilateral norms and verification institutions have all lost key supporting elements. INF and Open Skies are gone, the Vienna Document’s confidence-building framework has been substantially eroded, and the NPT review process has now failed three consecutive times to produce substantive outcomes.
That does not mean the treaty system has ceased to operate, or that IAEA safeguards have stopped functioning. But it does suggest that the political mechanism intended to update the system every five years has lost its capacity to deliver shared progress in an era of intensifying nuclear competition.
What the NPT Was — and What It Was Not
The NPT, in force since 1970, was built around a core transaction. Nuclear‑weapon states committed not to transfer nuclear weapons and to negotiate in good faith towards disarmament; non‑nuclear‑weapon states committed not to acquire them and to accept IAEA inspections; all parties were recognized the right to peaceful nuclear energy.
From the outset, this transaction was embedded in a three‑pillar structure: non‑proliferation, nuclear disarmament, and the peaceful uses of nuclear energy. Over time, the non‑proliferation and peaceful‑use pillars have continued to function relatively well through safeguards and technology cooperation, while the disarmament pillar has become the politically most fragile.
The NPT was never a universal prohibition treaty in the sense later embodied by the Treaty on the Prohibition of Nuclear Weapons. It does, however, legally prohibit non‑nuclear‑weapon states from acquiring nuclear weapons and nuclear‑weapon states from transferring them, while allowing the five recognized nuclear‑weapon states to retain their arsenals under a commitment to pursue disarmament. It was a bargain — non‑proliferation in exchange for a disarmament commitment — rather than an immediate, universal norm mandating a nuclear‑free world.
That balance has come under growing pressure. All five recognized nuclear‑weapon states are now modernizing their arsenals rather than reducing them, while new systems are being developed and doctrines adjusted. Legally, this may still fall within Article VI, which sets no timetable or form for disarmament. Politically, however, many non‑nuclear‑weapon states experience it as a breach of the treaty’s spirit — which in turn gradually erodes the argument that has long legitimized their own restraint.
New York 2026: Diagnosis from Inside the Conference
Opening the Review Conference on 27 April, UN Secretary‑General António Guterres warned that the treaty was eroding, commitments were unfulfilled, and proliferation pressures were accelerating: ”Today, a collective amnesia has taken hold. Nuclear sabers rattle once more. Mistrust rules the day. Hard-won norms are eroding. Arms control is dying.” France’s Foreign Minister Jean‑Noël Barrot described proliferation risk as at a record high “And the threat posed by the programs of Iran and North Korea must not be tolerated by any State Party to this treaty.” IAEA Director‑General Rafael Grossi cautioned against what he described as a growing perception that nuclear weapons benefit national security — ”The risk of nuclear disaster has risen to levels not seen since the height of the Cold War. … In today’s nuclear domain, we face a precarious standoff, with more actors, more risks, and less clarity.”
The European Leadership Network’s April 2026 compilation of European and P5 perspectives, published as the conference opened, reflected a similar diagnostic with important national variations. European contributors — from Italy, Ireland, Germany, Poland and France — converged on the need to preserve the multilateral non‑proliferation regime while acknowledging the geopolitical obstacles. P5 contributors offered more guarded assessments shaped by their own strategic priorities.
The conference’s eventual breakdown was publicly attributed primarily to the Iran dispute. The final draft contained language stating Iran ”can never seek, develop or acquire nuclear weapons” — wording the United States and several other states insisted must remain, while Iran objected to being singled out in a text that, in Iranian representatives’ framing, did not address what they described as recurring military strikes against their nuclear infrastructure. The US had already signalled displeasure at the opening session by objecting to Iran’s vice president attending. Iran thus became the immediate procedural breaking point.
Reports of successive drafts, however, suggest that the failure had more structural roots. Language on North Korea was softened; the IAEA Director‑General’s five principles on Zaporizhzhia nuclear safety did not appear in the final text; and references to nuclear‑sharing, extended deterrence, no‑first‑use, a New START successor, a fissile material cut‑off moratorium, and stronger negative security assurances had all been progressively weakened or removed. Iran was therefore the point where the process finally stalled, not the sole cause of the regime’s deeper malaise.
Taken together, the failed Review Conferences in 2015, 2022 and now 2026 point to more than diplomatic misfortune. Each failure had its own proximate trigger — the Middle East WMD‑free zone in 2015, the war in Ukraine in 2022, and the Iran paragraph in 2026 — but the pattern suggests that consensus‑based reviews have become increasingly incompatible with a strategic environment marked by active conflicts with nuclear dimensions and a P5 reluctant to accept new binding constraints.
Three Reinforcing Pressures
The erosion around the NPT does not occur in isolation. Three parallel dynamics appear to be intensifying pressure on the regime’s legitimacy.
China and a multipolar nuclear landscape. China has substantially expanded its nuclear arsenal over the past decade. According to US Department of Defense assessments, its arsenal grew from around 200 warheads in 2019 to approximately 600 by the mid‑2020s, with a projected trajectory towards roughly 1,000 by 2030, alongside a more complete nuclear triad. China has so far declined proposals for trilateral arms‑reduction talks with Washington and Moscow. Several European analyses note that this is accelerating a tripolar structure for which the classical bilateral arms‑control architecture was not designed.
Russian nuclear signalling. Since 2022, Russia has repeatedly used nuclear threats in its communications regarding Ukraine and its relationship with NATO, while developing new delivery systems and declaring intentions to station nuclear weapons in Belarus. At New York, Russian delegates criticized what they described as attempts to ”regulate political disagreements” through the NPT while opposing proposals that might have preserved parts of the review process. How much of this constitutes deliberate psychological signalling versus operational preparation is debated, but the normalization of such language makes nuclear weapons a more routinely invoked instrument — with consequences that remain difficult to predict.
The Middle East and attacks on nuclear infrastructure. Military strikes against nuclear facilities have become not merely a theoretical risk but a documented feature of recent conflicts. US and Israeli strikes against Iran’s nuclear sites in 2025 and 2026 destroyed significant parts of Iran’s nuclear infrastructure — yet Iran retained its stockpile of highly enriched uranium, its technical knowledge base, and, by all public accounts, its strategic intent to rebuild. Iran suspended IAEA cooperation in July 2025 and by March 2026 was openly reviewing withdrawal from the NPT itself. The logic is straightforward: if military attack, not safeguards and diplomacy, is the operative enforcement mechanism, states may conclude that NPT membership offers obligations without protection. That inference — visible in Iranian statements and in broader regional debates from the Gulf to Northeast Asia — represents one of the most direct challenges to the treaty’s incentive structure since its entry into force.
Together, these three dynamics reinforce a picture of an order in which both the disarmament promise and confidence in long‑term restraint mechanisms are gradually weakening.
What Remains — and Where the Argument Points
Despite the review process’s difficulties, the NPT remains near‑universal, and the IAEA safeguards system continues to function as a core verification mechanism. The absence of New START and other bilateral agreements does not render arms‑control thinking obsolete, but it does suggest that the field may need new formats and sometimes different entry points.
The risk of horizontal proliferation — long the NPT’s central concern — has moved from a structural background condition to an acute foreground problem. The Budapest Memorandum’s failure to protect Ukraine, and military strikes against previously safeguarded nuclear facilities, together undermine the two central arguments that have sustained non‑nuclear‑weapon states’ restraint: that security assurances offer meaningful protection, and that IAEA membership confers a degree of inviolability.
Several analyses — including those gathered by the ELN and recent work at Chatham House and CSS ETH Zürich — converge on a similar set of priorities. Over the period between New York 2026 and the next Review Conference in 2031, a minimum agenda seems to emerge from these debates.
First, protecting what still works: ensuring the financial and political resilience of IAEA safeguards, and reinforcing norms against attacks or threats against peaceful nuclear facilities in armed conflict.
Second, keeping nuclear risk‑reduction and crisis‑communication channels alive in regional and P5 formats when global forums are blocked — including work on doctrines, new technologies and command‑and‑control vulnerabilities. The P5 process — though substantially weakened and largely dormant in recent years — remains one of the few venues where even modest confidence‑building measures can be discussed. The joint P5 statement that ”a nuclear war cannot be won and must never be fought”, reaffirmed in 2022, represents a minimal but real shared anchor that has not yet been formally abandoned.
Third, sustaining pressure for transparency and accountability from the nuclear‑weapon states, so that the disarmament pillar does not become entirely hollow even as traditional arms‑control treaties lapse. Many non‑nuclear‑weapon states will see not only future negotiations, but concrete reporting, restraint in modernization, and defence of the Comprehensive Test‑Ban Treaty as tests of whether the ”core transaction” of the NPT still holds.
On the longer arc, much will depend on how the central actors assess the value of that core transaction. If the disarmament commitment continues to be perceived as hollow, pressure for alternative paths will grow: regional security arrangements, norm initiatives such as the Treaty on the Prohibition of Nuclear Weapons, or gradual drift towards greater strategic independence in specific regions.
The window between New York 2026 and the next review cycle is neither an ending nor a clean new beginning. The question may be less whether the NPT ”survives” in formal terms — it almost certainly will — than what function it will serve: continuing core of the non‑proliferation order, or primarily a ceremonial reference to an earlier era of managed competition. Which direction is taken will be shaped less by a single negotiating round than by the accumulated choices — in capitals, in Vienna, and in New York — of the states that assembled, and failed to agree, this May.