Scan: US x Cuba military clash in 2026?

1. Planner Fable 5.1 · Fermi decomposition + tool plan

Tools chosen: polymarket_direct polymarket_related kalshi_related claude_news article_search gdelt_news wikipedia code_execution

Sub-questions (Fermi decomposition)

What is the current Polymarket YES price and how has it trended over the past 90 days?
Has the Trump administration made explicit military threats toward Cuba in 2025-2026, or is there a US naval/Coast Guard buildup in the Caribbean near Cuba (e.g., tied to Venezuela operations)?
Have there been any recent incidents between the US Coast Guard/Navy and Cuban Border Guard or Cuban military vessels/aircraft (interdictions, standoffs, warning shots)?
What is the historical base rate of direct US-Cuba armed encounters since 1962, and how many years have seen any such incident?
Do related markets (US strikes on Venezuela, Cuba regime collapse, US military action in the Caribbean) indicate elevated escalation risk that could spill over to Cuba?
Are there structural factors (Guantanamo Bay disputes, Cuban migration surges, US designation of Cuba as a state sponsor of terrorism, Russian/Chinese military presence in Cuba) increasing the chance of a kinetic incident in 2026?

Planner reasoning

This is a low-base-rate geopolitical event; the key drivers are current US-Cuba tensions (Trump administration posture toward Cuba, Venezuela spillover, Caribbean naval buildup, Coast Guard interdictions near Cuban waters) and whether any incident could plausibly meet the strict 'use of force between military forces' threshold. The Polymarket price is the primary anchor, with Kalshi and related Polymarket markets (e.g., US-Venezuela strikes, Cuba regime change) as cross-signals. Historical base rate of direct US-Cuba military exchanges since 1962 is essentially zero, so the research should focus on whether 2026 is anomalous.

2. Research Parallel tool calls

tool	status	time	items	summary
polymarket_direct	OK	0.5s	1	## This Market's Polymarket Data US x Cuba military clash in 2026? - Current price (probability): 7.50% - 7-day price change: -6.00% - 30-day price change: -12.00% - Total volume: $247,564 (USD notional) - Price range: 7.50% - 51.00% - Data points: 90 days
polymarket_related	OK	3.6s	0	Scanned 100 active Polymarket markets, kept 0 matches. keyword 'Cuba': 0 markets \| keyword 'Venezuela strike': 0 markets \| keyword 'Caribbean military': 0 markets \| keyword 'Guantanamo': 0 markets
kalshi_related	OK	3.4s	1	1 related markets / summaries. keyword 'Cuba': no matches \| keyword 'Venezuela': no matches \| keyword 'military strike Caribbean': ok
claude_news	OK	32.1s	20	## Research Findings: US-Cuba Military Tensions (2025–2026) - US intervention in Venezuela sparked the crisis: Following the 2026 United States intervention in Venezuela, in which U.S. forces ousted president Nicolás Maduro, Venezuelan oil exports to Cuba were cut off, leaving the country with
article_search	OK	0.9s	23	Fetched 23 unique articles across 5 queries (lookback=60d). 'US Cuba military tensions': 10 articles \| 'Cuban Border Guard Coast Guard incident': 10 articles \| 'Trump Cuba military threat': 10 articles \| 'US Caribbean naval buildup Venezuela Cuba': 10 articles \| 'Guantanamo Bay Cuba dispute 2026': 1
gdelt_news	OK	170.5s	20	GDELT: 20 articles across 3 queries (lookback=30d). 'Cuba United States military': 10 hits \| 'Cuban navy US Coast Guard': 10 hits \| 'Cuba Guantanamo tensions': error GDELT rate-limited after retries (429)
wikipedia	OK	0.2s	4	Fetched 4 Wikipedia entries (2 missing pages).
code_execution	OK	18.2s	0	## Base Rate Analysis: US-Cuba Military Clash Historical count (1963–2025, 63 yrs): 0 documented direct armed engagements (exchange of fire, missile strike, ramming) between US military/Coast Guard and Cuban military/Border Guard forces. (1996 Brothers to the Rescue shootdown excluded — civilia

3. Evidence Brief Sonnet 5 · 6974 chars

# Current state As of early September 2026, no confirmed direct armed engagement between US and Cuban military forces has occurred despite historically extreme rhetorical and military escalation (naval buildup, EO designating Cuba a threat, Pentagon contingency planning). The Feb 2026 Cuban Border Guard shootout involved a US-*registered* civilian boat, not US military forces, and the US denies government involvement — this does NOT meet the resolution definition. Diplomatic channels (talks, prisoner releases) remain open in parallel with threats. # Timeline of key events - 2025-09: Trump deploys ~10% of US Navy fleet to Caribbean; begins lethal strikes on alleged drug boats (confirmed, ongoing campaign, not Cuba-specific) — militarytimes/defensepriorities. - 2026-01: US strikes oust Maduro in Venezuela; Cuba reports 32 security personnel killed in those Venezuela strikes (confirmed per Wikipedia "2026 in Cuba"); Cuban oil supply cut off. - 2026-01-29: White House EO "Countering Threats Posed by the Government of Cuba" formally labels Cuba a national security threat (confirmed, whitehouse.gov). - 2026-02-04 onward: US intel flights/drones near Cuba ~25 times (reported). - 2026-02-25: Armed clash between Cuban Border Guard and US-registered civilian boat near Villa Clara; 5 killed. US denies government involvement (confirmed incident, but NOT a US-military encounter) — Wikipedia. - 2026-03-13: Cuba confirms talks with US amid energy crisis (reported). - 2026-03-23: USS Nimitz deploys to Caribbean for "Southern Seas 2026" exercise (confirmed, militarytimes). - 2026-04-03: Cuba releases 2,000+ prisoners as goodwill gesture (reported). - Late May 2026: US-Cuban military officials meet at Guantanamo amid tensions (reported). - 2026-05: Trump/Rubio issue explicit military threats ("Cuba is next," "sometimes you have to use it") (reported, multiple outlets). - 2026-06: Hegseth visits Guantanamo, warns Cuba against arms acquisition; Axios reports Cuba acquired 300+ drones (reported). - 2026-07-16/19: CBS/BBC report Pentagon examining Cuba military options incl. Army air assault by 101st Airborne; officials stress no decision made (reported). - 2026-08-17/18: Hegseth says Cuba intervention "on the table" (reported). - 2026-08-20: DOJ indicts Raúl Castro over 1996 Brothers to the Rescue shootdown (confirmed). - 2026-08-21 to 09-04: Additional US sanctions rounds on Cuban entities (confirmed, ongoing). - 2026-08-26/29: US strikes on drug boats/refueling ship in Caribbean (confirmed, not Cuba-targeted). - 2026-09-01/06: Cuba says US waging "war without bombs"; Raúl Castro reported critically ill then appears publicly (reported/rumored). # Event Will there be a direct military encounter (use of force) between US and Cuban military forces between market creation and Dec 31, 2026? # Outcomes to forecast Yes / No # Kalshi market anchor Kalshi-direct price data was NOT returned by tools for this ticker (kalshi_related found no Cuba-specific market). Best available cross-platform anchor: **Polymarket YES = 7.5%**, down sharply from a 90-day high of 51% (-6pp 7-day, -12pp 30-day), $247.5K volume — suggesting market has been de-escalating pricing after an earlier spike (likely around the Jan/Feb Venezuela-strike and boat-incident period). # Sub-question answers 1. **Polymarket trend** — Currently 7.5%, down from a peak of 51% over the 90-day window; falling steadily (-12pp/30d), indicating cooling probability despite ongoing tensions [polymarket_direct]. 2. **Explicit threats/buildup** — Yes: EO (Jan 2026) labels Cuba a threat; Trump/Rubio explicit threats (May 2026: "Cuba is next"); Hegseth says intervention "on the table" (Aug 2026); USS Nimitz deployed (Mar 2026); Pentagon examined options incl. 101st Airborne air assault (Jul 2026) [claude_news, article_search]. 3. **US-Cuban military incidents** — No confirmed direct US military-Cuban military clash. The Feb 2026 boat shootout involved a US-registered *civilian* vessel, not US armed forces; US denies government involvement [Wikipedia, claude_news]. 4. **Historical base rate** — 0 documented direct armed US-Cuba military engagements 1963-2025 (Brothers to the Rescue excluded as civilian aircraft). Laplace-smoothed baseline ≈1.54%/year [code_execution]. 5. **Related market spillover** — No Polymarket/Kalshi Cuba-adjacent markets found (Venezuela/Caribbean strike markets return zero matches); can't directly cross-check via markets, but real-world Venezuela strikes (Jan 2026) directly preceded Cuba tension spike [polymarket_related, kalshi_related]. 6. **Structural factors** — Guantanamo friction (US-Cuban officer meetings, Hegseth visit/warnings), Cuban drone acquisition (300+, per Axios), state sponsor of terrorism designation, Russian SIGINT facility in Cuba cited in EO — all present and active [claude_news]. # Key facts (high-confidence, factual) 1. [whitehouse.gov] EO formally designating Cuba a threat, Jan 29 2026. 2. [Wikipedia] Feb 25 2026 Cuban Border Guard vs. US-flagged civilian boat clash, 5 killed; not US-military. 3. [militarytimes] USS Nimitz in Caribbean since March 2026. 4. [CBS News] Pentagon examined Cuba military options July 2026; no decision made. 5. [code_execution] Zero historical direct US-Cuba military armed clashes 1963-2025. # Cross-market signals - Kalshi related: no Cuba-specific comparable found. - Polymarket: 7.5% YES, declining trend, high past volatility (peak 51%). - Sportsbook implied: none available. # Analyst opinions and speculation - CSIS: US has overwhelming capability advantage (Cuba "essentially no navy"); punitive strikes feasible within days if ordered. - Politico/military.com: Pentagon "positioned" for strikes, awaiting presidential green light. - Defense Priorities/American Conservative: caution against repeating "Iran mistake" in Cuba — dovish pushback exists. # Directional lean per outcome - **Yes**: Unprecedented rhetorical escalation, EO, carrier deployment, Pentagon contingency planning, Guantanamo friction, Raúl Castro indictment — all elevate risk far above historical baseline. - **No**: Zero historical precedent ever; parallel diplomatic talks/prisoner releases continue; Feb incident already didn't qualify; Polymarket pricing has fallen sharply to 7.5%, suggesting informed traders see de-escalation; no decision by Trump reported as of Sep 2026. # Gaps / unknowns - No direct Kalshi YES price retrieved — anchor uncertainty. - Whether events after Sept 2026 (research cutoff) show further escalation or de-escalation is unknown. - Definition ambiguity: Coast Guard/Border Guard incidents could count if genuinely both are "military," but Feb incident excluded due to civilian boat. # Calibration anchors - Polymarket YES 7.5% (proxy anchor, Kalshi unavailable). - Historical base rate ~1.5%/year; even 10x elevation implies ~15% ceiling. - Comparable "hot" precedent: 1962 Missile Crisis (no shots fired), 1996 Brothers to the Rescue (civilian, not military).

4. Forecasters Fable 5.1 + GPT-5.6 Sol · in parallel

claude-fable-5-1

48% conf

Yes	9%
No	91%

No direct US-Cuban military engagement has occurred as of early September 2026; the Feb 2026 Villa Clara shootout involved a US-registered civilian boat and does not qualify. The best available anchor is Polymarket YES at 7.5%, down sharply from a 51% peak with a steady declining trend, indicating traders see de-escalation despite the Jan EO, Nimitz deployment, Pentagon contingency planning, and Hegseth's 'on the table' remarks. The historical base rate is ~1.5%/year with zero direct clashes since 1963, and parallel diplomatic channels (talks, prisoner releases, Guantanamo officer meetings) remain active. I nudge slightly above the market to reflect genuine tail risk from an unusually hawkish administration with forces positioned and ~4 months remaining, but the 'No' outcome remains strongly favored.

gpt-5.6-sol

68% conf

Yes	8%
No	92%

The 7.5% Polymarket YES price is the best available consensus anchor, and its sustained decline points toward No despite elevated tensions. Explicit threats, regional US deployments, Pentagon contingency planning, and Cuban drone acquisitions keep the risk well above the roughly 1.5% historical annual baseline. However, no qualifying clash had occurred by early September, the February shootout involved a civilian vessel, and diplomatic and military communication channels remained open. With only several months left in 2026 and no reported decision to attack, I assign only a small premium over the market anchor.

5. Devil's Advocate Sonnet 5 · red-team critique

Both forecasters anchor almost entirely on the Polymarket 7.5% figure, but the brief itself flags this as a weak proxy ("Kalshi-direct price data was NOT returned"; no Cuba-specific Kalshi market found), yet neither forecast discusses why Polymarket's thin $247.5K volume and high recent volatility (peak 51%) might make it an unreliable predictor over a full remaining ~4-month window with active escalation events (Hegseth "on the table" Aug 17/18, DOJ indictment Aug 20, sanctions through Sep 4) occurring *after* much of that price decline.

Neither forecast explicitly grapples with the compressed time-frame effect: the Polymarket price at 7.5% likely reflects annualized-style de-escalation sentiment, but the event resolves by Dec 31, 2026 with only ~4 months left and clear evidence of *continued* escalation into September (indictment, new sanctions, "war without bombs" rhetoric) — this recency of hawkish signals is underweighted relative to the market's lagging decline.

Both treat the "Pentagon examined options, no decision made" (Jul 2026) and "on the table" (Aug 2026) statements as already priced in without asking whether these are unusually concrete relative to historical rhetoric-only episodes (e.g., 1962 Missile Crisis) — the brief notes Pentagon is "positioned... awaiting presidential green light," which is a stronger operational readiness signal than typical base-rate scenarios, warranting more probability mass than a ~9% figure implies.

Neither forecaster addresses the definitional ambiguity flagged in "Gaps/unknowns" — that Coast Guard/Border Guard incidents "could count if genuinely both are military" — meaning a repeat of a Feb-style incident but with confirmed government-linked (not civilian) vessels could flip a "No" into "Yes," and given Cuba's 300+ drone acquisition and border friction, this pathway deserves explicit discussion rather than a blanket "does not qualify" dismissal.

The two forecasts are nearly identical (9.0% vs 8.0%) despite F2 claiming much higher confidence (0.68 vs 0.48) — this discrepancy in stated confidence given near-identical probability estimates and largely the same reasoning is unexplained and internally inconsistent.

Both anchor near the historical base-rate ceiling implied by the brief ("even 10x elevation implies ~15% ceiling") but neither explains why they land near the low end of that range rather than nearer the ceiling, given the brief describes conditions as "unprecedented" and far exceeding any prior year in the 1963-2025 reference class (carrier deployment, EO, indictment, Guantanamo friction, drone buildup) — a case could be made for weighting closer to the 15% ceiling rather than ~8-9%.

6. Aggregator Fable 5.1 · final calibrated forecast

10% (58% confidence)

Yes: 10% No: 90%

Both forecasters converge near 8-9%, slightly above the Polymarket 7.5% anchor, and the critique's strongest points are that the anchor is a thin, volatile proxy and that concrete operational signals (Pentagon options, 'on the table', indictment) arrived after much of the price decline. However, the Polymarket contract resolves on this same event and window, so the critique's 'annualized sentiment' argument does not hold — traders were pricing the remaining ~4 months and still moved down through the August escalation, indicating they read the rhetoric as coercive rather than pre-attack. No qualifying clash had occurred by early September, diplomatic and military channels remained open, and the historical base rate is near zero. I nudge modestly above both forecasts to reflect an administration that has actually used force in the region, forces positioned awaiting only a presidential decision, and a plausible definitional pathway via a Coast Guard/Border Guard or drone incident, but 'No' remains strongly favored.

Pipeline Timing

Total pipeline time: 271.9s

Per-tool research timings shown in the Research section above.