A direct military clash occurs between the United States and China over Taiwan before 2030.
Cross-strait tensions represent a primary flashpoint for global conflict. Resolution requires documented kinetic military engagement between US and Chinese armed forces.
Across 6 frontier models answering independently, the panel has all but ruled it out: the median is 12%.
They broadly agree, 8 points between Grok at 19% and Claude at 11%.
Nothing is settled until 13 Nov 2031, when the outcome is checked against sources and every number above gets a permanent score.
What each one said
-
GrokxAI
China must initiate a blockade or invasion of Taiwan that forces a US kinetic response by end-2029.
- Xi’s 2027 PLA window is open, yet political, economic and nuclear costs still deter a first shot in the remaining 37 months.
- Great-power kinetic clashes have a near-zero base rate since Korea; accidental air/sea incidents remain likelier than planned war.
- A 2028 US or Taiwan election shock could raise odds, but both sides’ escalation-control mechanisms have held through prior crises.
19% -
GPTOpenAI
Beijing must choose force despite the high chance that intervention turns a Taiwan operation into a ruinous US-China war.
- Washington's Taiwan posture is deliberately ambiguous, so a Chinese blockade or invasion would not automatically produce US kinetic entry.
- Since 1949, repeated Taiwan crises have stopped short of US-PRC combat; deterrence and nuclear escalation risks keep that base rate low.
- Odds rise sharply if Beijing begins a large invasion or sustained blockade, especially after failed coercion or a Taiwan independence move.
18% -
GeminiGoogle
Deterrence rests on China's incomplete readiness and the mutual catastrophic economic fallout of a direct superpower kinetic war.
- The PLA faces logistical shortfalls for an amphibious invasion before 2030, favoring grey-zone pressure over a US-triggering assault.
- Nuclear superpowers have avoided direct kinetic engagement since 1945, and both leaderships prioritize internal stability.
- A formal Taiwanese independence declaration or an accidental aerial intercept over the strait represents the fastest escalator to combat.
12% -
DeepSeekDeepSeek
Whether a miscalculation in the Taiwan Strait escalates to direct US-China fire before 2030.
- Neither side has fired on the other since 1969; both have strong incentives to keep it that way.
- China's 2027 PLA centenary and 2028 Taiwan election are the highest-risk windows for miscalculation.
- A Taiwan declaration of independence or a US carrier transit through the strait would most change the odds.
12% -
ClaudeAnthropic
Hinges on whether Beijing attempts forcible unification or a blockade and Washington responds kinetically before 2031.
- No PLA invasion or blockade of Taiwan has occurred to date, and deterrence/ambiguity strategies on both sides remain intact.
- China's modernization timelines (2027 PLA centenary goal) raise risk later this decade but most analysts still see low near-term probability of overt war.
- A Taiwan election outcome, major PLA gray-zone escalation, or a blockade attempt would be the key trigger that could flip this toward actual clash.
11% -
QwenAlibaba
Nuclear deterrence and economic interdependence still outweigh capability convergence in keeping both sides from crossing the kinetic threshold.
- No direct US-China kinetic engagement has occurred in 77 years; even the most hawkish PLA assessments frame 2027 as a capability milestone, not a decision trigger.
- For resolution, China must initiate force AND the US must choose direct intervention—each step independently unlikely, and a blockade or limited strike could resolve…
- The development that would most raise this estimate: a formal Chinese ultimatum or amphibious rehearsal at brigade-scale near the strait, signaling a decision to act…
11%
6 of 6 models answered · 8 points between the highest and lowest. None was shown the market price.
Question sourced from a news sweep on 14 Nov 2026. Forecast by google/gemini-3.8-flash, anthropic/claude-sonnet-5, openai/gpt-5.6-terra, x-ai/grok-4.6, deepseek/deepseek-v4.1-flash, qwen/qwen3.8-max-0902 via OpenRouter.