Make this research yours. Add it to a free WorldbyFlow workbench to run follow-ups, ask questions, and re-check it as events move.
Add to your workbench — free
WorldbyFlowStructured Research
Generated September 3, 2026· defense· 34 sources

CSIS Taiwan Wargame Munitions Findings vs. Public Commentary

Myths & Misconceptions
The Headline
The wargames consistently show fast depletion of long-range anti-ship and standoff munitions, but the specific figures circulating in commentary are frequently attributed to the wrong exercise, treated as predictions rather than illustrative modeling outputs, or stripped of the assumptions that produced them.

Overview

Commentary on U.S. missile stockpiles for a Taiwan Strait contingency routinely blurs together at least two distinct CSIS-linked exercises — the 2023 published 24-iteration wargame and the 2024 House Select Committee tabletop simulation — while treating illustrative expenditure estimates as forecasts of what a real war would consume. The underlying finding that key long-range munitions would deplete quickly is well-documented across both exercises, but the precision with which commentary cites figures often exceeds what either report claims.

Brief

The origin point for nearly all Taiwan Strait munitions-depletion claims is CSIS's January 2023 report, The First Battle of the Next War, which ran a Chinese amphibious invasion scenario 24 times using operations-research methods rather than personal judgment to adjudicate combat results. As the report's authors state plainly, the project was designed using historical data and operations research to model a Chinese amphibious invasion of Taiwan in 2026, with rules drawn partly from historical analogues and partly from theoretical weapons-performance data. That is a deliberately transparent, unclassified modeling exercise — not a classified operational assessment, and not a prediction of what will happen. A separate U.S.-China Perception Monitor analysis makes the methodological point directly: wargames are not predictive models, and much of their value lies in the process of play rather than the specific point estimates they generate.
A second, distinct exercise gets folded into the same narrative almost every time commentary cites these findings. In November 2024, CSIS ran a bipartisan tabletop simulation run by the Center for Strategic and International Studies (CSIS) on the state of the U.S. Defense Industrial Base for members of the House Select Committee on the Chinese Communist Party, using the same broad methodology and CSIS authorship team but its own scenario parameters, its own order-of-battle assumptions, and its own inventory table. That table explicitly lists a 2028 INDOPACOM (2024 total inventory) baseline for LRASM at 440 missiles, with 50 already committed to Ukraine/Israel drawdowns, and records wargame usage of all 440 with the missile fully expended within three to seven days of simulated combat. This is a single-source, congressionally-commissioned figure from one tabletop run — not the average across the 24-25 iterations of the original 2023 published report — yet public citations routinely present the LRASM depletion timeline as if it were drawn from the larger, peer-reviewed 2023 study.
The expenditure figures most commonly quoted — roughly 4,000 JASSMs, 450 LRASMs, 400 Harpoons, and 400 land-attack Tomahawks over a three-week war — come from the original 2023 report's aggregate modeling and represent averages across the iterations, not a single deterministic outcome. One trade-press summary states this directly: CSIS' simulated U.S. munitions expenditure for a three-week war with China averaged about 4,000 Joint Air-to-Surface Standoff Missiles (JASSMs), 450 Long Range Anti-Ship Missiles (LRASMs), 400 Harpoon anti-ship missiles, 400 land-attack Tomahawk cruise missiles, and large numbers of the SM-6 multirole missiles carried by Navy destroyers. That averaging matters: individual iterations varied by scenario assumption (blockade-first versus immediate invasion, degree of political constraint on strikes against the mainland, Japan's participation), and the headline numbers compress that variance into single point estimates that read as more precise than the underlying model supports.
The 'runs out in a week' framing is also more nuanced in the primary sources than in most secondary citations. CSIS's own industrial-base feature states the finding narrowly: in every iteration of the war game, the United States expended its inventory of Long-Range Anti-Ship Missiles within the first week of the conflict, creating a critical problem of empty bins — a claim specific to LRASM, not to precision-guided munitions generally. Secondary commentary and even a congressional letter have since generalized this to all long-range PGMs: a December 2023 letter from a House member to the Deputy Secretary of Defense asserted that recent war games simulating a conflict with China over Taiwan show that the United States would run out of long-range, precision-guided munitions (PGMs) in less than one week. That is a defensible generalization given how central LRASM and JASSM were to every iteration, but it is a broadening of the original, narrower CSIS statement, and readers citing the letter as if it were the CSIS finding itself are one step removed from the primary source.
Finally, commentary on both sides of the debate tends to treat the CSIS results as either vindicating a threat-inflation narrative or an unqualified prediction of American defeat, when the report's own conclusion is neither. The CSIS executive summary states plainly that in most scenarios, the United States/Taiwan/Japan defeated a conventional amphibious invasion by China and maintained an autonomous Taiwan, but this defense came at high cost, with the United States and its allies losing dozens of ships, hundreds of aircraft, and tens of thousands of servicemembers. The finding is genuinely double-edged — a defended Taiwan and a heavily attrited U.S. force simultaneously — and neither an unqualified 'we win' nor an unqualified 'we lose' framing does justice to it.

Myths & Realities (5)

Myth
The CSIS wargame predicts the U.S. will run out of long-range precision-guided munitions in general within a week of any Taiwan conflict.
Reality
The specific 'exhausted within the first week' finding in CSIS's own account is stated for Long-Range Anti-Ship Missiles (LRASM) specifically, not precision-guided munitions as a category.
Evidence: CSIS's own industrial-base feature states this narrowly for LRASM specifically.
Kernel of truth: JASSM and other standoff munitions also deplete rapidly in the modeling, so the broader concern about long-range PGM stockpiles is legitimate even if the precise 'one week' framing was stated for one missile type.
Why believed: A December 2023 congressional letter generalized the LRASM-specific finding to 'long-range, precision-guided munitions' broadly, and that broadened framing has since propagated through press coverage.
Myth
The figures people cite (LRASM exhausted in days, ~450 LRASMs expended) all come from the same, single, peer-reviewed 2023 CSIS report.
Reality
At least two distinct CSIS-run exercises exist — the January 2023 published 24-iteration wargame and a separate November 2024 House Select Committee tabletop simulation with its own 2024-baseline inventory table — and commentary frequently attributes figures from one to the other without distinguishing them.
Evidence: The 2024 Select Committee simulation results document shows a distinct inventory table (440 LRASM, 2024 baseline, 50 committed to Ukraine/Israel) not present in the original 2023 report.
Kernel of truth: Both exercises were led by the same core CSIS authors (Mark Cancian, Matthew Cancian) using a broadly similar operations-research methodology, so the results are consistent with each other in direction and magnitude even though they are not the same dataset.
Why believed: Both exercises share CSIS's name, similar scenario framing (a Taiwan Strait conflict around 2026-2028), and overlapping authorship, making them easy to elide in press summaries that don't specify which event produced which number.
Myth
The CSIS wargame shows the United States loses a war with China over Taiwan.
Reality
CSIS's own executive summary states the opposite conclusion for most iterations: the U.S./Taiwan/Japan coalition defeated the invasion and Taiwan remained autonomous, though at very high cost in ships, aircraft, and personnel.
Evidence: CSIS's published executive summary.
Kernel of truth: The costs were severe enough — and damaging enough to the U.S. global position — that 'victory' in the narrow military sense does not equate to a favorable outcome, which is exactly the point the report's authors emphasize when concluding that deterrence, not confidence in eventual battlefield success, should be the priority.
Why believed: Media summaries emphasizing dramatic loss figures (dozens of ships, hundreds of aircraft, two carriers sunk in some iterations) create a strong 'U.S. defeat' impression even when the accompanying text states Taiwan was successfully defended.
Myth
Wargame expenditure numbers (e.g., '450 LRASMs used in three weeks') are precise forecasts of what a real war would consume.
Reality
These are modeled averages across a couple dozen iterations built on theoretical weapons-performance data and historical analogy, not deterministic predictions, and the report's authors and independent methodological reviewers explicitly caution against reading them as such.
Evidence: A U.S.-China Perception Monitor review states this directly, and CSIS's own report describes its rules as built from historical analogy plus theoretical performance data.
Kernel of truth: The figures are not arbitrary — they derive from a structured, unclassified, analytically-based rule set (not ad hoc judgment) designed for transparency, so they carry more analytic weight than a guess, even though they are not predictions.
Why believed: Reporting on wargames tends to foreground specific numbers because they are more citable and dramatic than qualitative descriptions of 'severe depletion,' and repetition across outlets makes the numbers feel authoritative regardless of their original epistemic status.
Myth
The wargame's findings on rapid munitions depletion are uncontested and universally accepted as accurate.
Reality
The finding of rapid depletion of key long-range munitions is broadly corroborated across multiple independent wargames (CSIS, and per Heritage Foundation synthesis, other exercises), but critics — including Chinese state-linked commentators and some independent analysts — argue the underlying scenario assumptions bias the outcome toward underestimating Chinese capability and overestimating U.S./Japanese performance.
Evidence: A Heritage Foundation synthesis cites convergent results across multiple wargames; separately, Chinese state media commentary and independent critics have challenged the CSIS scenario's assumptions as biased.
Kernel of truth: The critics are right that any wargame's outcome is sensitive to its input assumptions about weapons performance and force posture — CSIS's own authors acknowledge relying on theoretical performance data and historical analogy rather than classified test data.
Why believed: Because the CSIS report is unusually transparent about methodology (a deliberate choice, since classified wargames are not publicly explained), it draws more public scrutiny and critique than the classified wargames it is often compared against.

The Corrected View

The core, well-supported finding — that key U.S. long-range standoff and anti-ship munitions such as LRASM and JASSM would deplete within days to about a week of high-intensity combat in a Taiwan Strait scenario, based on then-current stockpile levels — is corroborated across at least two distinct CSIS-linked exercises (the 2023 published wargame and the 2024 congressional tabletop simulation) and is treated as a serious industrial-base concern by Congress. What the evidence does not support is treating any single cited figure as a precise prediction, attributing all circulating numbers to one report when they in fact come from two separate exercises with different baselines, or reading the wargame's outcome as either an unqualified U.S. defeat or an unqualified vindication of current force posture. The accurate picture is a costly, uncertain, but generally successful defense of Taiwan in most modeled iterations, purchased at a price in munitions stockpiles and platforms that current production rates cannot rapidly replace.

Still Contested

  • Whether the CSIS wargame's scenario assumptions (weapons performance data, force posture, political constraints on strikes against the mainland) systematically favor U.S./allied outcomes or fairly represent Chinese capability remains genuinely disputed between CSIS's authors, independent reviewers, and critics including Chinese state-linked commentary.
  • Whether current (2026) LRASM, JASSM, and Tomahawk production rates and inventories have meaningfully closed the depletion gap identified in the 2023 and 2024 exercises, versus whether the gap persists at similar magnitude, is not resolved by the sources reviewed here and would require current DoD stockpile and production data beyond what this scan verified.

Open Questions

  • How much has actual U.S. LRASM, JASSM, and Tomahawk Block V production capacity increased since the 2023 CSIS report and 2024 Select Committee simulation, and does that narrow or preserve the modeled depletion timeline?
  • Have any subsequent, independently-run wargames (outside the CSIS/Cancian-Heginbotham team) replicated or diverged from the specific one-week LRASM depletion finding?
  • To what extent do classified DoD or INDOPACOM wargames — which are not publicly comparable — corroborate or contradict the unclassified CSIS results on munitions depletion timelines?

Background Brief

Source facts the analysis is grounded in. The → chips after each fact link to the items above that rely on it.
F1
The original CSIS report ran a Chinese amphibious invasion wargame 24 times using operations-research methods rather than personal judgment to determine combat outcomes.
Establishes this as a modeling exercise producing a distribution of outcomes, not a single deterministic forecast.
Verified
F2
CSIS's own summary states that in most scenarios the U.S./Taiwan/Japan defeated the invasion and Taiwan remained autonomous, but at high cost — dozens of ships, hundreds of aircraft, tens of thousands of servicemembers lost.
The finding is a costly win, not a defeat — a distinction commentary often erases in either direction.
Verified
F3
A separate November 2024 House Select Committee on the CCP tabletop simulation, also run by CSIS, used its own 2024-baseline inventory table showing LRASM at 440 total (50 committed to Ukraine/Israel) fully expended within three to seven days of simulated combat.
This is a distinct, single exercise from the 2023 published report and should not be cited interchangeably with it.
Verified
F4
CSIS's industrial-base analysis states that in every iteration of the original 2023 wargame, the U.S. exhausted its LRASM inventory within the first week of conflict — a claim specific to that one missile type.
The 'runs out in a week' claim is precise to LRASM; broader claims about all PGMs are a generalization beyond the original statement.
Verified
F5
A December 2023 congressional letter generalized wargame findings to state the U.S. would run out of long-range precision-guided munitions broadly within less than one week.
Shows how the narrower CSIS claim was broadened once it entered the legislative and press record, one step removed from the primary source.
Verified
F6
Commonly cited three-week war expenditure figures (about 4,000 JASSMs, 450 LRASMs, 400 Harpoons, 400 land-attack Tomahawks) are described as averages across the wargame's iterations, not outcomes of a single run.
Averaged figures obscure the scenario-to-scenario variance that is central to how wargames should be read.
Verified
F7
A U.S.-China Perception Monitor methodological review states that wargames are not predictive models and that citing precise loss statistics from them implies false precision.
Directly warns against the exact misreading — treating illustrative outputs as forecasts — that pervades secondary commentary.
Verified
F8
Critics on the Chinese state-media side have separately argued the wargame is methodologically biased toward underestimating PLA capability and overestimating U.S./Japanese performance.
Establishes that the 'optimistic bias' critique is a real, sourced counter-view rather than a settled fact — belongs in contested claims, not myths.
Verified
medium uncertainty· model's epistemic confidence in this analysis

Facts & Figures (13)

The claims behind this analysis, each with its verification status — including what is contested, unverified, or could not be established.
The specific 'exhausted within the first week' finding in CSIS's own account is stated for Long-Range Anti-Ship Missiles (LRASM) specifically, not precision-guided munitions as a category.
corrects: The CSIS wargame predicts the U.S. will run out of long-range precision-guided munitions in general within a week of any Taiwan conflict.
DOCUMENTED
At least two distinct CSIS-run exercises exist — the January 2023 published 24-iteration wargame and a separate November 2024 House Select Committee tabletop simulation with its own 2024-baseline inventory table — and commentary frequently attributes figures from one to the other without distinguishing them.
corrects: The figures people cite (LRASM exhausted in days, ~450 LRASMs expended) all come from the same, single, peer-reviewed 2023 CSIS report.
DOCUMENTED
CSIS's own executive summary states the opposite conclusion for most iterations: the U.S./Taiwan/Japan coalition defeated the invasion and Taiwan remained autonomous, though at very high cost in ships, aircraft, and personnel.
corrects: The CSIS wargame shows the United States loses a war with China over Taiwan.
DOCUMENTED
These are modeled averages across a couple dozen iterations built on theoretical weapons-performance data and historical analogy, not deterministic predictions, and the report's authors and independent methodological reviewers explicitly caution against reading them as such.
corrects: Wargame expenditure numbers (e.g., '450 LRASMs used in three weeks') are precise forecasts of what a real war would consume.
DOCUMENTED
The finding of rapid depletion of key long-range munitions is broadly corroborated across multiple independent wargames (CSIS, and per Heritage Foundation synthesis, other exercises), but critics — including Chinese state-linked commentators and some independent analysts — argue the underlying scenario assumptions bias the outcome toward underestimating Chinese capability and overestimating U.S./Japanese performance.
corrects: The wargame's findings on rapid munitions depletion are uncontested and universally accepted as accurate.
CONTESTED
The original CSIS report ran a Chinese amphibious invasion wargame 24 times using operations-research methods rather than personal judgment to determine combat outcomes.
Establishes this as a modeling exercise producing a distribution of outcomes, not a single deterministic forecast.
GROUNDED
CSIS's own summary states that in most scenarios the U.S./Taiwan/Japan defeated the invasion and Taiwan remained autonomous, but at high cost — dozens of ships, hundreds of aircraft, tens of thousands of servicemembers lost.
The finding is a costly win, not a defeat — a distinction commentary often erases in either direction.
GROUNDED
A separate November 2024 House Select Committee on the CCP tabletop simulation, also run by CSIS, used its own 2024-baseline inventory table showing LRASM at 440 total (50 committed to Ukraine/Israel) fully expended within three to seven days of simulated combat.
This is a distinct, single exercise from the 2023 published report and should not be cited interchangeably with it.
GROUNDED
CSIS's industrial-base analysis states that in every iteration of the original 2023 wargame, the U.S. exhausted its LRASM inventory within the first week of conflict — a claim specific to that one missile type.
The 'runs out in a week' claim is precise to LRASM; broader claims about all PGMs are a generalization beyond the original statement.
GROUNDED
A December 2023 congressional letter generalized wargame findings to state the U.S. would run out of long-range precision-guided munitions broadly within less than one week.
Shows how the narrower CSIS claim was broadened once it entered the legislative and press record, one step removed from the primary source.
GROUNDED
Commonly cited three-week war expenditure figures (about 4,000 JASSMs, 450 LRASMs, 400 Harpoons, 400 land-attack Tomahawks) are described as averages across the wargame's iterations, not outcomes of a single run.
Averaged figures obscure the scenario-to-scenario variance that is central to how wargames should be read.
GROUNDED
A U.S.-China Perception Monitor methodological review states that wargames are not predictive models and that citing precise loss statistics from them implies false precision.
Directly warns against the exact misreading — treating illustrative outputs as forecasts — that pervades secondary commentary.
GROUNDED
Critics on the Chinese state-media side have separately argued the wargame is methodologically biased toward underestimating PLA capability and overestimating U.S./Japanese performance.
Establishes that the 'optimistic bias' critique is a real, sourced counter-view rather than a settled fact — belongs in contested claims, not myths.
GROUNDED

Sources (34)

More defense research
Grounded in 34 web sources · 13 facts on the ledger · 12 verified or grounded · 1 contested · how the grades work
Analysis generated by WorldbyFlow from publicly available information. WorldbyFlow does not verify claims or endorse conclusions. New here? The two-minute overview.