TLDR: No expert gives a clean percentage for “nuclear war in 2026 or 2027,” because the event is too rare to calculate that way. The closest numbers we have: superforecasters peg the chance of a nuclear catastrophe by 2045 at around 1%, domain experts put it at 5%, and forecasting platforms have repeatedly assigned near-0% odds to specific short-term scenarios like a Russian strike on Ukraine. The risk environment in 2026 is elevated compared to the 1990s and 2000s, but still below Cold War peaks.
Key Takeaways
- Most experts refuse to give a single probability for nuclear war in a specific year
- A major 2024 survey put expert median odds of a nuclear catastrophe by 2045 at 5%, and superforecaster odds at 1%
- Metaculus forecasters consistently rated specific 2025–2026 nuclear scenarios at well under 1%
- The Doomsday Clock sits at 89 seconds to midnight entering 2026, the closest setting ever
- New START expires February 2026, removing the last cap on US and Russian deployed warheads
- The most plausible path is escalation from a conventional war, not a deliberate first strike
Why You Cannot Get One Clean Number
Ask ten nuclear analysts for the odds of nuclear war in 2026 and you will get ten different answers, or more often, a refusal to answer at all. The reason is statistical. Nuclear strikes are a sample size of two in history, both in 1945. You cannot run a frequentist probability calculation on an event that has happened twice in 80 years against a backdrop of constantly shifting geopolitics.
Brookings analysts have argued it is a mistake for experts to assign a hard percentage to nuclear war at all. They prefer relative claims like “higher than last year” or “elevated” over numbers that imply more precision than the data supports.
That said, several serious efforts have produced numbers worth knowing.
The Most Cited Expert Estimate Right Now
The Open Nuclear Network study published through the Forecasting Research Institute surveyed 110 nuclear domain experts and 41 superforecasters. They were asked to estimate the probability of a nuclear catastrophe, defined as nuclear weapons killing at least 10 million people, by 2045.
The numbers:
- Domain experts: median 5%
- Superforecasters: median 1%
That is for the next two decades, not a single year. If you crudely annualize the expert estimate, you get something in the range of 0.25% per year. The superforecaster estimate annualizes to roughly 0.05% per year. Neither group thinks nuclear war is likely in any given 12-month window, but neither thinks it is negligible either.
What Forecasting Platforms Say About Specific 2026 Scenarios
Metaculus runs public forecasting markets where calibrated forecasters bet on specific outcomes. The track record on recent nuclear questions:
- Russia testing a nuclear device before January 2026: community estimate 0.1%, resolved no
- North Korea conducting a nuclear test before May 2026: community estimate around 0.6%, resolved no
- Russia detonating a nuclear weapon in Ukraine before 2026: community estimate 0.1%, resolved no
These are narrow questions, not full nuclear war scenarios. But they show forecasters consistently see specific near-term nuclear events as very low probability, even in active conflicts.
What Has Changed Going Into 2026
The risk profile is not static. Several things shifted in 2025 and early 2026 that nuclear analysts are watching closely.
New START Expires in February 2026
The last remaining treaty capping US and Russian deployed strategic warheads at 1,550 each ran out in February. There is no successor agreement. Arms control experts argue this opens the door to a triangular arms race between the US, Russia, and China.
China’s Stockpile Is Growing
The Pentagon estimates China will field roughly 1,000 nuclear warheads by 2030, up from a few hundred a decade ago. Some US strategists are now openly discussing the possibility of fighting two nuclear wars simultaneously.
Tactical Nukes Are Back in the Conversation
Russia deployed new short-range non-strategic nuclear weapons during the Ukraine war and threatened their use multiple times. The US has responded with its own short-range nuclear cruise missiles. Lowering the threshold for “limited” nuclear use is itself an escalation risk.
Testing May Resume
President Trump ordered US preparations for new nuclear tests if Russia or China test first. China has been expanding its Lop Nur test site. No major power has tested since 1996, but the moratorium is fraying.
The Doomsday Clock Sits at 89 Seconds
The Bulletin of the Atomic Scientists set the clock to 89 seconds to midnight for 2026, the closest setting in its 78-year history. That is a qualitative judgment by a panel of scientists, not a probability, but it tracks how the expert community feels about the trajectory.
What Is the Most Likely Path to Nuclear Use?
Almost no serious analyst thinks the highest-probability scenario is a bolt-from-the-blue strategic exchange between superpowers. Deterrence still works for that case. The pathways experts actually worry about:
Escalation From a Conventional War
A losing side in a major conventional conflict reaches for tactical nuclear weapons to reverse battlefield losses or signal resolve. Russia and Ukraine is the obvious case. A Taiwan scenario is the other.
Miscalculation or Accident
False warning systems, cyber intrusion into command and control, or a leader misreading an adversary’s intent. The historical record contains multiple near-misses, from the 1983 Soviet false alarm to the 1995 Norwegian rocket incident.
Regional Powers
India and Pakistan, Israel and Iran, or a North Korean miscalculation. These conflicts would not directly involve the US or Russia in a first-strike sense, but a regional nuclear exchange would still kill millions and reshape global politics.
The Open Nuclear Network study found experts believe these risks are spread roughly evenly across five flashpoints: Russia and NATO, China and the US, the Korean Peninsula, India and Pakistan, and Israel and Iran.
How Should You Think About This Personally?
A low annual probability across many years adds up. A 1% to 5% chance over 20 years is not a number you can dismiss. But it is also not a number that should drive daily decisions for most people.
For prepping and survival planning, the practical framing is this: the same fundamentals that prepare you for a hurricane, an earthquake, or a grid-down event cover most of what you would need in a regional nuclear scenario. Water, food, sheltering in place, communications, and a plan to get your family on the same page. Distance from major counterforce targets and military bases gives you a meaningful edge. Iodide tablets are cheap. Knowing how to shelter for the first 48 to 72 hours after a detonation is free.
If you live in a city near a major military installation, your risk profile is different than if you live in a small town in the rural Midwest. Geography matters more than people think.
The Honest Bottom Line
No one credible is going to tell you the probability of nuclear war in 2026 or 2027 is X percent. The serious people who have tried to put numbers on the question are talking about 20-year horizons, and the numbers are low single digits.
What we can say:
- The risk in 2026 is higher than at any point since the early 1980s
- The most likely pathway is escalation from a conventional conflict, not a deliberate strike
- Specific short-term scenarios consistently get rated under 1% by calibrated forecasters
- The expiration of arms control architecture and the breakdown of testing norms are real, measurable changes
- Basic preparedness covers most of what you would actually need
Plan for what you can control. Stop scrolling for what you cannot.
Frequently Asked Questions
What are the odds of nuclear war in 2026?
No credible expert assigns a single percentage to a specific calendar year. The closest available numbers are aggregate estimates for nuclear catastrophe over 20-year windows, which range from 1% to 5%.
Has the probability gone up recently?
Most analysts say yes, qualitatively. The Doomsday Clock is at its closest setting ever, arms control treaties are expiring, and active conflicts in Ukraine and the Middle East create escalation pathways that did not exist five years ago.
Who is most likely to use a nuclear weapon first?
Experts are divided. Russia and NATO scored highest in one major survey, but the risk is spread roughly evenly across five potential flashpoints including China and the US, India and Pakistan, North Korea, and Iran and Israel.
Is a full-scale nuclear war or a limited strike more likely?
A limited or tactical use scenario is considered more plausible than a full-scale strategic exchange. Deterrence is still strong against the latter. The lowered threshold for tactical use is the bigger concern in 2026.
What is New START and why does it matter?
New START was the last treaty limiting US and Russian deployed strategic warheads. It expired in February 2026 without a replacement, removing the cap that has shaped the nuclear balance for 15 years.
