Public Notes

Current affairs explainers · Note 060

After New START: the world's nuclear guardrails come down

Security · Reference edition 1 November 2026 · 2 min read

Study context: GS III — Security challenges and their management in border areas; role of external state and non-state actors; various security forces and agencies and their mandate.

Background explainer from the November 2026 study collection. Events discussed may predate the collection; this is not a monthly news update.

On 5 February 2026, the New Strategic Arms Reduction Treaty — New START — expired. Signed in April 2010 and in force from February 2011, it was the last treaty limiting US and Russian strategic arsenals, capping each side at 700 deployed delivery vehicles, 1,550 deployed warheads and 800 launchers and bombers, with inspections and data exchanges. Extended for five years in 2021, it was suspended by Russia in February 2023 and expired as a hollow shell. For the first time since the early 1970s, there are no binding restrictions on the strategic nuclear arsenals of the two largest nuclear-weapon states.

The expiry did not arrive unannounced. In September 2025, President Vladimir Putin offered to keep observing the treaty's central limits for a year if the United States did the same. Washington instead announced plans to negotiate a new, improved, modernised treaty with an expanded scope covering Russia's tactical nuclear warheads as well as China's arsenal; China refused to participate, and no follow-on agreement had been reached by mid-2026. The Federation of American Scientists put Russia's 2026 inventory at about 5,420 warheads, the largest of any country, and the United States at just over 5,000; together they hold roughly 86 per cent of the world's nuclear weapons, while SIPRI estimated China's arsenal at about 620 warheads and growing.

The practical consequence is uncertainty. Without data exchanges and on-site inspections, outside estimates of Russian forces will become less certain, and each side must plan against worst-case assumptions about the other — the classic dynamic that arms control was invented to break. Russia also withdrew its ratification of the nuclear test ban treaty in 2023, though it has not carried out an explosive test since the Soviet era. The erosion is broader than one treaty: over fifteen years other agreements have unravelled, and new domains — cyber, space, AI and hypersonics — were never covered by the old framework.

For India, which holds its own nuclear arsenal outside the Non-Proliferation Treaty framework and maintains a doctrine of credible minimum deterrence and no first use, the post–New START world changes the strategic environment rather than India's own posture directly. A world without verified caps is a world where the two largest arsenals can grow unchecked, where China's expansion draws American responses, and where crises — in Ukraine, in the Taiwan Strait, in the Middle East — unfold without the transparency measures that once gave leaders time and information. The question for 2026 is not whether arms control is dead, but whether a new kind can be built: one that covers new weapons, includes China, and values inspections and information-sharing as much as numerical limits.

Key takeaways

  • New START, the last US–Russia strategic arms treaty, expired on 5 February 2026, leaving no binding limits on the two largest arsenals for the first time since the early 1970s.
  • The treaty capped each side at 700 deployed delivery vehicles and 1,550 deployed warheads; Russia suspended participation in 2023 and offered a one-year limits extension that Washington declined.
  • Without inspections and data exchanges, estimates of Russian forces grow less certain, and China — estimated at about 620 warheads — remains outside any cap.

Why it matters for civil studies

Arms control is information as much as limitation: verification and data exchange prevent worst-case assumptions from driving arms races.

Think it through

If a new treaty must include China and tactical weapons to be worthwhile, but neither is achievable now, is a limited extension better than no agreement at all?

From the Public Notes companion study collection. Editorial context and corrections.

WhatsApp