← Back to scans

Will China invade Taiwan by June 30, 2027?

0x057da165b096d1f3a59d87bcad68649aecd18dc25c89a6d083530d1aaac18146 · World · 2026-08-26
5%
Agent
8%
Market Price
-2.5%
Edge
73%
Confidence
Volume: 428,637
Spread: 1.0c
Days to resolution: 307
Markets in event: 1
Final Rationale
The de facto anchor (matching-ticker Polymarket contract) is 7.5% and trending down, while the far more liquid end-2026 contract at 3.65% implies only modest incremental risk accruing in H1-2027. The March 2026 US IC assessment of no active 2027 invasion plan, the absence of mobilization/reservist/economic war-footing indicators, and July 2026 PLA activity at a three-year low all point strongly to No; 76 years of PRC-Taiwan non-invasion under nuclear-shadowed deterrence suggests the generic ~6.1% great-power base rate may overstate this dyad. I take the critique's point that the resolution bar includes a limited seizure of an outlying island (Kinmen/Matsu/Pratas) and that escalation indicators (Type 076, median-line crossing, 'ready by 2027') justify not straying far below the market, so I settle at 5% rather than the 4% of Forecast 1 — a modest longshot-premium discount only, not a double-count of the declining-price trend. Stale data past August 2026 adds uncertainty but is symmetric and already partly reflected in the market's wide historical range.
Price vs. prediction
How the market moved after the agent's call. Solid blue line is the actual market price; dashed line is the agent's static forecast at $t{=}0$.
$n = 8$ follow-ups
Re-scan Context
This market has been scanned before. Previous predictions:
DatePredictedMarket PriceConfidence
2026-08-18 7% 8% 68%
1. Planner Opus · Fermi decomposition + tool plan
Tools chosen: polymarket_direct polymarket_related kalshi_related claude_news gdelt_news wikipedia code_execution
Sub-questions (Fermi decomposition)
  1. What is the current Polymarket YES price and recent price history for 'Will China invade Taiwan by June 30, 2027?', and what is its volume/liquidity?
  2. Do Kalshi or other Polymarket markets price a China-Taiwan military action (invasion, blockade, quarantine) over comparable horizons, and do they agree with this market?
  3. What is the historical base rate of a great-power invasion of a contested territory in any given ~2.5-year window, and what annualized hazard rate does the market price imply?
  4. What do current US/Taiwan/Japanese official intelligence assessments say about PLA readiness for a 2027 invasion (the 'Davidson window'), and have they shifted in the last 6 months?
  5. Are there current escalation indicators — large-scale PLA exercises (Strait Thunder/Joint Sword-style), blockade drills, amphibious lift buildup, mobilization, reservist call-ups, or actions against Kinmen/Matsu/Pratas?
  6. What is the state of US-China and Japan-China relations and Taiwan's political situation (2028 election cycle, Lai administration policy) that could raise or lower invasion risk before mid-2027?
  7. Has China taken any sub-invasion coercive steps (customs inspections, cable cutting, ADIZ incursions, seizure of an uninhabited feature) that would NOT resolve this market YES but signal rising risk?
Planner reasoning
This is a low-base-rate geopolitical tail event with a ~2.5-year horizon (from 2025 to June 2027); the market price is likely in the low single digits. Key research is anchoring on the Polymarket price, cross-checking Kalshi/other venues for similar Taiwan invasion/blockade markets, and updating on current PLA activity, US/Taiwan intelligence assessments (e.g., the 'Davidson window' 2027 readiness date), and escalation indicators.
2. Research Parallel tool calls
ToolStatusTimeItemsSummary
polymarket_direct OK 0.3s 1 ## This Market's Polymarket Data **Will China invade Taiwan by June 30, 2027?** - Current price (probability): 7.50% - 7-day price change: -1.00% - 30-day price change: -3.00% - Total volume: $428,637 (USD notional) - Price range: 7.50% - 15.50% - Data points: 130 days
polymarket_related OK 4.0s 1 Scanned 100 active Polymarket markets, kept 1 matches. keyword 'Taiwan': 1 markets | keyword 'China invade': 0 markets | keyword 'China Taiwan blockade': 0 markets | keyword 'US China war': 0 markets
kalshi_related OK 3.9s 3 3 related markets / summaries. keyword 'Taiwan': ok | keyword 'China Taiwan': ok | keyword 'China military': ok
claude_news OK 28.7s 16 ## Key Findings: China-Taiwan Invasion Probability by June 30, 2027 **Official intelligence assessments now downplay a 2027 invasion:** - The IC assesses that Chinese leaders do not currently plan to execute an invasion of Taiwan in 2027, nor do they have a fixed timeline for achieving unification
gdelt_news OK 175.2s 24 GDELT: 24 articles across 3 queries (lookback=45d). 'China Taiwan invasion 2027': error GDELT rate-limited after retries (429) | 'PLA exercise Taiwan blockade': 12 hits | 'Taiwan Strait tensions military': 12 hits
wikipedia OK 0.1s 2 Fetched 2 Wikipedia entries (1 missing pages).
code_execution OK 20.8s 0 ## Key Findings: Constant-Hazard Poisson Model for China–Taiwan Invasion by June 2027 **Setup:** Window = 19 months ≈ 1.583 years (Dec 2025 → Jun 30 2027). Using p = 1 − e^(−λT), solved for annualized hazard λ and monthly hazard h. **Implied hazard rates by market price p:** - **p = 0.02** → annua
3. Evidence Brief Sonnet · 7646 chars
# Current state Polymarket's identically-ticketed contract (0x057da165...) prices "China invades Taiwan by June 30, 2027" at **7.5% YES**, down from a 15.5% high; this is the primary market signal since no separate Kalshi order book was returned for this ticker. A shorter Polymarket contract ("by end of 2026") trades far lower (3.65%), consistent with rising cumulative risk over a longer window rather than imminent action. US intelligence (March 2026 report) now assesses Beijing has no fixed 2027 invasion plan, even as PLA readiness and gray-zone pressure continue to build. # Timeline of key events - 2024-05: PLA begins more frequent blockade-style exercises following Lai Ching-te's inauguration (confirmed, AEI/ISW). - 2025 (year, multiple rounds): Sixth major PLA war-game series since 2022; late-Dec 2025 "Justice Mission-2025" exercise — encirclement/blockade rehearsal with elevated aviation activity (confirmed, Global Taiwan Institute/AEI). - 2026-03: US intelligence report states China not currently planning a 2027 invasion, no fixed unification timeline; "Davidson Window" assessed as effectively closed (confirmed, InsideDefense). - 2026-06/07: Reporting reiterates Xi directed PLA to be "ready by 2027," distinct from a decision to invade (reported, The Diplomat/Aquilino testimony). - 2026-07-21/22: PLA helicopter crosses Taiwan Strait median line for apparent first time (reported, Taipei Times/GlobalSecurity). - 2026-07-24: Chinese military activity around Taiwan reported at a three-year low (reported, Times of India). - 2026-07-28: Taiwan's Han Kuang drills scenario outlined; defensive posture (confirmed, Taipei Times). - 2026-08-08 to 08-12: Taiwan's Han Kuang war games (coastal strike drills, internet throttling test) run in parallel with Chinese naval activity (confirmed, multiple). - 2026-08-11/13: China-Indonesia joint naval exercise east of Taiwan; US urges China to halt military pressure after the drill (confirmed, Global Times/FocusTaiwan). - 2026-08-18: PLA "wartime" drill involving cash/gold-bar drone drops reported (reported, Times of India). - 2026-08 (ongoing): Japan-China tensions elevated after PM Takaichi's Nov 2025 "survival-threatening situation" remarks; trade/travel retaliation by Beijing; Trump reportedly urges de-escalation; Japanese lawmakers visit China late August 2026 to repair ties (confirmed/reported, CSIS/Tribune India/China Global South Project). # Event Will China commence a military offensive to seize any inhabited ROC territory by June 30, 2027 ("Yes"/"No"). # Outcomes to forecast - Yes - No # Kalshi market anchor No distinct Kalshi order-book data was returned for this ticker; the identically-ticketed Polymarket contract shows **YES = 7.5%**, down 1pt (7d) and 3pts (30d) from a 15.5% high over 130 days, on $428,637 total volume. Treat this 7.5% as the de facto market anchor given the shared ticker. # Sub-question answers 1. **Polymarket price/history/liquidity** — YES = 7.5%, range 7.5–15.5% over 130 days, $428,637 volume; declining trend both 7d and 30d [polymarket_direct]. 2. **Other Kalshi/Polymarket China-Taiwan contracts** — Polymarket's shorter "by end of 2026" version prices YES at 3.65% on $40.2M volume (much more liquid), consistent with lower risk over a shorter window [polymarket_related]. No Kalshi blockade/quarantine-specific market found; related Kalshi contracts (Level-4 travel advisory 0.1%, Trump recognizing Taiwan 15%) don't directly price invasion [kalshi_related]. 3. **Historical base rate** — Model using 1 great-power contested-territory invasion per 25 years implies ~6.1% probability over this 19-month window (≈4.0% annual hazard) [code_execution]. Market's 7.5% implies annual hazard ~1.66× this base rate. 4. **Official intelligence on 2027 readiness** — March 2026 US IC report: China not currently planning 2027 invasion, no fixed timeline; "Davidson Window" seen as closed as a decision point, though PLA readiness target stands (Aquilino testimony) [claude_news/InsideDefense]. 5. **Escalation indicators** — Yes: frequent PLA exercises (Justice Mission-2025, Aug 2026 China-Indonesia drill), first-ever median-line helicopter crossing, new Type 076 amphibious carrier-like vessel; but overall PLA activity reported at a three-year low in July 2026 [claude_news/gdelt_news]. 6. **US-China/Japan-China/Taiwan politics** — Japan-China tensions spiked post-Takaichi remarks (Nov 2025) with trade retaliation; Trump urged de-escalation; Japanese delegation visited China Aug 2026 to mend ties. Taiwan's Lai administration continues defensive Han Kuang drills; no acute election-driven flashpoint noted before 2027 (next Taiwan presidential election is 2028) [claude_news]. 7. **Sub-invasion coercive steps** — Gray-zone activity continues: ADIZ/median-line incursions, encirclement drills, naval shadowing near Taiwan waters (July 2026 detection of PLA vessel/official ships); no seizure of ROC-administered features reported [gdelt_news/claude_news]. # Key facts (high-confidence, factual) 1. [polymarket_direct] Matching-ticker contract YES=7.5%, trending down. 2. [claude_news/InsideDefense] March 2026 US IC assessment: no active 2027 invasion plan. 3. [claude_news] PLA conducted repeated large-scale exercises through 2025-2026, but July 2026 activity level reported at 3-year low. 4. [polymarket_related] Shorter-horizon (end-2026) contract prices YES far lower (3.65%) with much higher liquidity ($40.2M). 5. [code_execution] Base-rate Poisson model: ~6.1% implied by generic historical hazard for 19-month window. # Cross-market signals - Kalshi related: No direct Taiwan-invasion Kalshi contract found; adjacent contracts (travel advisory, recognition) don't triangulate invasion risk directly. - Polymarket: 7.5% (this ticker) vs 3.65% (shorter, higher-liquidity end-2026 contract) — term structure suggests market sees marginal added risk in the 2026-H2/2027-H1 stretch beyond 2026, plausible given "2027 readiness" narrative. - Sportsbook implied: None available. # Analyst opinions and speculation - Analysts (AEI, Global Taiwan Institute) frame 2027 as a PLA capability benchmark, not a decision date; consensus is invasion unlikely by mid-2027 but gray-zone/miscalculation risk elevated (claude_news). - Some commentary (Responsible Statecraft) suggests Xi personally downplayed 2027 as a fixed target in private diplomatic exchanges. # Directional lean per outcome - **Yes**: Supported by continued PLA modernization (Type 076), median-line crossings, repeated blockade-style drills, Japan-China friction. Opposing: explicit March 2026 IC assessment of no invasion plan, declining exercise tempo (3-yr low, July 2026), market price falling over 30 days. - **No**: Strongly favored — market consensus (7.5%), intelligence assessments, base-rate modeling (~6%) all cluster in single-digit-to-low-teens probability, no mobilization/reservist call-up evidence. # Gaps / unknowns - No native Kalshi order book confirmed for this exact ticker (only cross-referenced via matching Polymarket ID). - No explicit blockade/quarantine-specific market found to triangulate a "gray-zone but not invasion" scenario probability. - Post-August 2026 developments (most recent data point) not covered; unknown near-term trajectory into 2027. # Calibration anchors - Polymarket YES price (proxy anchor): 7.5%, down from 15.5% high. - Shorter-horizon comparable (end-2026): 3.65%. - Base-rate model estimate: ~6.1% over comparable window. - Historical precedent: No PRC-Taiwan invasion has occurred since 1949; last major crisis escalation (1996 missile crisis, 2022/2025 exercises) never crossed into invasion.
4. Forecasters Opus + GPT · in parallel
claude-opus-5
0.76
Yes 4%
No 96%
The de facto market anchor (matching-ticker Polymarket contract) sits at 7.5% YES and is trending down, while the far more liquid shorter-horizon 'by end-2026' contract prices only 3.65%, implying modest incremental risk in H1-2027. The March 2026 US intelligence assessment explicitly states Beijing has no active 2027 invasion plan and no fixed unification timeline, and PLA activity around Taiwan hit a three-year low in July 2026 with no mobilization, reservist call-up, or economic war-footing indicators. A generic great-power base-rate model implies ~6% over this 19-month window, and no PRC-Taiwan invasion has occurred in 76 years despite repeated crises (1996, 2022, 2024-26 drills). I shade modestly below the 7.5% anchor because low-probability catastrophic-event contracts on prediction markets carry a well-documented longshot/hedging premium and because the trend in both price and observable escalation indicators is downward.
gpt-5.6-sol
0.70
Yes 7%
No 93%
The 7.5% Polymarket YES price is the best available consensus anchor, while the more liquid end-2026 contract at 3.65% supports a modestly rising rather than imminent risk. The March 2026 US intelligence assessment of no fixed invasion plan, combined with no reported mobilization or seizure of ROC territory, points strongly toward No. Repeated blockade exercises, PLA modernization, and median-line crossings preserve a meaningful tail risk, but July's reported three-year low in activity and the roughly 6.1% historical base-rate estimate justify placing YES slightly below the market anchor.
5. Devil's Advocate Sonnet · red-team critique
1. Both forecasters shade below the 7.5% anchor citing a "longshot premium," but Forecast 1's larger downward shade (to 4.0% vs market's 7.5%) isn't well justified given the brief explicitly notes escalation indicators (Type 076 carrier, first-ever median-line helicopter crossing, Xi's "ready by 2027" directive) that argue for staying closer to or above the anchor, not well below it — this looks like anchoring-adjustment overconfidence rather than evidence-driven. 2. Neither forecast addresses the event definition nuance: the question resolves YES for "commence a military offensive to seize any inhabited ROC territory," which could include a limited seizure of an outlying island (e.g., Kinmen, Matsu, or Pratas) rather than a full Taiwan invasion — this is a materially lower bar than "invade Taiwan" and may be underpriced by both the market anchor and both forecasters, who treat it as synonymous with full invasion. 3. Both forecasters treat the base-rate model (~6.1%) and market price (7.5%) as mutually reinforcing, but neither questions whether a "1 great-power contested-territory invasion per 25 years" reference class is appropriate for a case with 76 years of non-invasion, nuclear deterrence dynamics, and no historical precedent — this generic base rate could be a significant overestimate for this specific dyad. 4. Neither rationale grapples with the "Gaps/unknowns" caveat that data is stale as of August 2026, leaving 10+ months uncovered before the June 2027 close — both express fairly high confidence (0.76, 0.70) despite this data gap and the brief's own admission that "near-term trajectory into 2027" is unknown. 5. Forecast 1's citation of "longshot/hedging premium" as a reason to shade below market price is somewhat inconsistent with the brief, which shows the price has already fallen from 15.5% to 7.5% — this trend could reflect genuine information updating (declining risk) rather than a persistent structural premium, making the extra downward shade partly a double-count of the same declining-trend evidence already used elsewhere in the rationale. 6. Neither forecaster examines the term-structure signal in more granular detail: the end-2026 contract (3.65%) implies only ~3.85 percentage points of incremental risk accrue in the 2027 H1 stretch alone, which is a large jump relative to the base rate for a single half-year — this could indicate the market is pricing a specific catalyst (e.g., a "2027 readiness" deadline effect) that neither analysis explores as a scenario driver.
6. Aggregator Opus · final calibrated forecast
5% (73% confidence)
Yes: 5% No: 95%
The de facto anchor (matching-ticker Polymarket contract) is 7.5% and trending down, while the far more liquid end-2026 contract at 3.65% implies only modest incremental risk accruing in H1-2027. The March 2026 US IC assessment of no active 2027 invasion plan, the absence of mobilization/reservist/economic war-footing indicators, and July 2026 PLA activity at a three-year low all point strongly to No; 76 years of PRC-Taiwan non-invasion under nuclear-shadowed deterrence suggests the generic ~6.1% great-power base rate may overstate this dyad. I take the critique's point that the resolution bar includes a limited seizure of an outlying island (Kinmen/Matsu/Pratas) and that escalation indicators (Type 076, median-line crossing, 'ready by 2027') justify not straying far below the market, so I settle at 5% rather than the 4% of Forecast 1 — a modest longshot-premium discount only, not a double-count of the declining-price trend. Stale data past August 2026 adds uncertainty but is symmetric and already partly reflected in the market's wide historical range.
Pipeline Timing
Total pipeline time: 273.8s
Per-tool research timings shown in the Research section above.