Prolepsia forecastYes / no

Will North Korea conduct a nuclear test before 31 December 2027?

As of 17 November 2026, Prolepsia puts the chance of yes at 14%.

14%chance of yes

Yes 14%No 86%
Prolepsia forecast made

How Prolepsia reasons

The report Prolepsia wrote with this forecast.

How forecasts are made · Track record

TL;DR

Probability of Yes: 14.1%. Probability of No: 85.9%. North Korea is likely to go without a confirmed nuclear explosive test through December 30, 2027, UTC: it has had the option to test for years without using it, and the latest readiness reports do not establish a decision to proceed.

Key evidence

The CTBTO record contains six North Korean tests from 2006 to 2017, with the last on September 3, 2017. That earlier cluster shows that North Korea has tested repeatedly when it chose to do so. But the more relevant comparison is the period since 2022, when U.S. officials reported indications that preparations were complete. No confirmed test followed.

In June 2026, the IAEA reported ongoing enrichment at Yongbyon and Kangson but no significant change at Punggye-ri, which remained ready. The August safeguards report likewise found no significant activity at the test site during its reporting period. Continued weapons development keeps another test possible; it does not put one on the calendar.

November diplomacy also gave little reason to expect a sharp change. Pyongyang left dialogue open in principle, but the U.S. overture had not produced a mutually confirmed meeting or a test-specific pledge by November 17. Continued restraint is the most likely outcome even without a deal.

Uncertainties

The main risk is a change in incentives: a severe regional crisis, a decision by Kim Jong Un that a test is worth the diplomatic cost, or an explosive test by another nuclear power. UN statements made the last scenario plausible but did not establish a test schedule or a North Korean commitment to respond in kind. A private authorization could also precede any public signal. Credible evidence of such a decision would matter more than another readiness warning; a reciprocal commitment not to test would move the outlook the other way. Only a corroborated, deliberate nuclear explosion attributable to North Korea before 00:00 UTC on December 31, 2027 qualifies. Missile launches, preparations, and unsupported claims do not.

Read the full report

TL;DR

Probability of Yes: 14.1%. Probability of No: 85.9%. The baseline for planning is no North Korean nuclear explosive test before 00:00 UTC on December 31, 2027: the test site is ready, but years of readiness have not produced a detonation.

Context

As of November 17, 2026, North Korea was still expanding its nuclear capabilities, according to the October 13 Congressional Research Service report. The IAEA reported continued enrichment at Yongbyon and Kangson and assessed that Punggye-ri remained prepared to support another test. The question is whether Pyongyang will use that option in the forecast window, not whether it has the option.

The window runs from November 17, 2026, through December 30, 2027, inclusive, in UTC. A deliberate nuclear explosion attributable to North Korea counts at any yield and in any location, including a test that falls short of its intended yield. Preparations, missile launches without a nuclear detonation, simulations, subcritical experiments, accidents, and combat use that is not a test do not count.

Evidence

North Korea has conducted six confirmed nuclear explosive tests: October 9, 2006; May 25, 2009; February 12, 2013; January 6 and September 9, 2016; and September 3, 2017 (CTBTO chronology). The five repeat tests came in just under eleven years, with three in 2016 and 2017 alone. Then came more than nine years without a confirmed test. Extrapolating from the active testing years gives too much weight to a campaign that has not continued. Averaging the active years and the pause also misses the break between them.

The more telling stretch begins in 2022. U.S. officials warned about preparations in June and reported indications in December that preparations were complete. Yet no confirmed detonation followed. That period is a better guide than treating every year since 2017 as equally comparable: it shows what North Korea has actually done while it was publicly assessed to have a usable test site.

The IAEA’s June 8, 2026 statement reported no indication of significant change at Punggye-ri even as it described ongoing enrichment elsewhere (IAEA briefing). Its August 28 safeguards report likewise found no significant activity at the site during the reporting period despite continued readiness. Those findings reinforce the distinction between a weapons program that is advancing and a decision to conduct an explosive test. The public Ninth Party Congress report called for stronger nuclear forces and continued tests and drills, but did not specify a nuclear explosion or give a date.

Diplomacy offers a reason to defer a test, but not a basis for assuming a freeze. China confirmed Kim Jong Un’s October 24 meeting with Wang Huning and a separate November meeting between Wang Yi and a North Korean party delegation. Those contacts show that channels were open; they do not show what was conveyed. Kim Yo Jong’s November 4 statement left dialogue open on conditions, and Trump raised a possible meeting. By the forecast date, there was neither mutual confirmation of a meeting nor a test-specific commitment. The case for no test does not depend on a diplomatic breakthrough.

An explosive test by another nuclear power is a plausible way for the picture to change. The October 8 U.S. statement at the UN clarified that the 2025 directive concerned preparation for explosive testing, but supplied no schedule (UN transcript). Russia reiterated a conditional response. North Korea rejected test-ban language at the UN on October 26 without announcing a test or saying it would follow a foreign detonation. These statements describe a route to escalation, not a sequence already under way.

What's non-obvious

A ready site sounds like a countdown when the assessment is read on its own. In this case, readiness is the standing condition, not the new signal. The strongest evidence for another year of restraint is not the lack of a public announcement. It is the record of not testing for more than four years after preparations were first reported. A new readiness assessment confirms that North Korea can still act; it says much less about when or why it would act.

The same distinction applies to diplomacy. Confirmed contacts with Chinese officials are firmer evidence of an open channel than a public invitation to meet. Neither is evidence of a bargain. That leaves a wide middle ground between a formal freeze and a return to testing: North Korea keeps the option, rejects demands to give it up, and still chooses not to use it during the window.

Uncertainties

Leadership intent is the biggest unknown. Public statements do not establish what Kim has decided in private. The record of the sixth test shows why the absence of a public order is weak evidence: Kim formally approved that test on the day it took place (UN report). Credible evidence of a fresh authorization, a severe regional crisis, or a foreign explosive test accompanied by a clear North Korean response would matter more than another missile launch or a missed summit. A reciprocal agreement that explicitly commits North Korea not to test would be stronger evidence in the other direction than talks alone.

Confirmation is a separate uncertainty. The CTBTO detected all six past tests, but that record does not guarantee that an unusually small event would be identified and attributed promptly. YES requires an authoritative technical assessment or corroboration from at least two independent, credible sources. The normal review date is January 31, 2028, with possible deferral to March 31 for a specific suspected event under active authoritative investigation. Those dates allow time to establish what happened; they do not extend the window in which a test counts.

Forecast history

How Prolepsia’s forecast has movedOne forecast so far: 14% on 17 Nov 2026.17 Nov 2026: 14%
DateForecast
14% chance of yes

Sources Prolepsia read

  1. IAEA Director General's Introductory Statement to the Board of Governors | IAEAwww.iaea.org
  2. Detecting Nuclear Weapon Test Explosions | CTBTOwww.ctbto.org
  3. U.S. Special Representative to the Democratic People’s Republic of Korea Sung Kim On Recent DPRK Missile Launches - United States Department of State2021-2025.state.gov
  4. Report on Ninth Congress of Workers’ Party of Korea – KKF Online Englishkkfonline.com
  5. Wang Huning Accompanies DPRK Supreme Leader Kim Jong Un on Visit to Dandong_Ministry of Foreign Affairs of the People's Republic of Chinawww.mfa.gov.cn
  6. 김여정 “선택지에 대화 있지만…미국 대북 적대시 정책 포기가 전제” - 경향신문www.khan.co.kr
  7. Trump raised a possible meetingrollcall.com
  8. First Committee, 4th plenary meeting - General Assembly, 81st session — UN Transcriptstranscripts.un.org
  9. United Nationsdigitallibrary.un.org
  10. (LEAD) S. Korea-U.S. combined exercise to conclude ahead of schedule after Trump order | Yonhap News Agencym-en.yna.co.kr

The question

This question asks whether the Democratic People's Republic of Korea (DPRK), commonly known as North Korea, will conduct at least one nuclear explosive test between November 17, 2026, and December 30, 2027, inclusive. The deadline is 00:00 UTC on December 31, 2027, meaning that tests conducted on December 31 itself will not count. A qualifying test must involve the actual detonation of a nuclear explosive device, whether underground, underwater, in the atmosphere, or in space. As of November 17, 2026, North Korea's most recent confirmed nuclear explosive test occurred on September 3, 2017, its sixth since 2006. According to an October 2026 Congressional Research Service report, North Korea continues to expand its nuclear weapons capabilities, while the International Atomic Energy Agency (IAEA) has reported ongoing uranium enrichment activities at Yongbyon and Kangson. The IAEA also reported in June 2026 that North Korea's Punggye-ri nuclear test site remained prepared to support another nuclear test. These developments suggest that North Korea retains the technical capability to resume nuclear explosive testing, although its political decision to do so remains uncertain. The question concerns an actual nuclear detonation, not preparations, threats, missile launches, or demonstrations involving simulated nuclear warheads. Background sources include the Congressional Research Service report of October 13, 2026 (https://www.everycrsreport.com/files/2026-10-13_IF10472_fa964aaa8c971acbe11a0e1fbd733515a18b3ca6.pdf) and the Comprehensive Nuclear-Test-Ban Treaty Organization (https://www.ctbto.org/our-work/detecting-nuclear-tests).

How it resolves

The question resolves YES if credible evidence establishes that North Korea conducted at least one nuclear explosive test between November 17, 2026, and December 30, 2027, inclusive, using UTC to determine the date of the event. A qualifying test requires a physical nuclear explosion attributable to North Korea, regardless of the device's explosive yield, whether the test succeeds in achieving its intended yield, or whether North Korea publicly acknowledges it. Evidence will be evaluated primarily using reports and technical assessments from the Comprehensive Nuclear-Test-Ban Treaty Organization (CTBTO), including its International Monitoring System, supplemented by assessments from the IAEA and official statements from the governments of South Korea, Japan, and the United States. A North Korean government announcement may support a YES resolution, but an unsupported claim alone is insufficient. A YES resolution requires either an authoritative technical assessment identifying a nuclear explosion attributable to North Korea or corroborating evidence from at least two independent, credible sources establishing that such an explosion occurred. Evidence may include seismic signatures, radionuclide detections, other technical observations, and authoritative intelligence assessments. The question resolves NO if no qualifying nuclear explosive test occurred within the specified period, as established by the available evidence. Because evidence of a test may emerge after the deadline, final resolution should normally occur no earlier than January 31, 2028, allowing time for technical analysis and official reporting. If credible evidence published by that date establishes that a qualifying event occurred before the deadline, the question resolves YES even if the event was not publicly identified until afterward. If no qualifying evidence has emerged by January 31, 2028, the question resolves NO unless a specific suspected event remains under active authoritative investigation, in which case resolution may be deferred until March 31, 2028. If an event remains genuinely disputed at that point and available evidence cannot establish whether a nuclear explosion occurred, the question will be annulled rather than resolved on speculation.

Fine print

A nuclear explosive test includes a deliberately initiated nuclear fission or fusion explosion, including an unsuccessful or low-yield nuclear detonation. Conventional explosives used to simulate nuclear detonations, hydrodynamic experiments, subcritical nuclear experiments, nuclear-capable missile launches without nuclear detonations, and nuclear reactor operations do not qualify. An accidental nuclear detonation does not count because the question concerns deliberate nuclear testing. A nuclear weapon detonated during an actual military attack rather than a test does not qualify unless credible evidence establishes that the detonation was conducted as a nuclear explosive test. Tests conducted outside North Korean territory count if the nuclear explosive device was tested by North Korea or under its direction. A test conducted jointly with another country counts if North Korea is established to have participated as an operator or sponsor of the nuclear explosive test, rather than merely observing it. A natural earthquake or conventional explosion incorrectly reported as a nuclear test does not qualify. The event must occur before 00:00 UTC on December 31, 2027. The location of the detonation does not change the deadline's time zone. Subsequent revisions to technical assessments may be considered until the question is formally resolved.