The Long War: Attrition, Empty Bins and a Negotiated Settlement

Summary

This scenario starts where most published wargames stop. China's invasion of Taiwan does not succeed quickly, and the US–Japan–Taiwan coalition does not defeat it quickly either. The 2023 CSIS wargame (24 iterations) found that the invasion usually failed, but only at high cost to everyone: the US lost two aircraft carriers, 10–20 large surface combatants and about 3,200 personnel in roughly three weeks [1][2]. Here, China secures a lodgement in southern Taiwan but cannot take the island. The fighting then settles into a long contest at sea and in the air around Taiwan, combined with a partial blockade.

China moves first. Its amphibious assault comes after a short crisis over a Taiwanese political transition. The US intervenes within days, and Japan allows combat operations from its bases. The first weeks go roughly as CSIS projected: heavy Chinese naval losses, heavy US and Japanese losses, and Taiwan's economy shattered [2].

The war is decided by what happens next. US stocks of long-range anti-ship missiles run out within about a week, and stocks of long-range cruise missiles within about a month [5][6]. Replacing them takes years [4][6]. Lost US warships cannot be replaced in wartime at all. Chinese shipyards, with a capacity that US naval intelligence put at roughly 200 times America's [7], repair damaged ships and launch new hulls. China accepts heavy losses and rebuilds its amphibious and missile forces. The US fights on with a shrinking fleet and rationed munitions.

At about D+540 (18 months), with no prospect of breaking the blockade and growing domestic opposition, Washington accepts a ceasefire. Under the settlement Chinese forces stay on Taiwan, Taiwan's government accepts "consultations" on unification and demilitarisation, and US forces withdraw from combat operations west of Japan. This is a negotiated victory for China, not a conquest. It still marks a decisive change in the regional order.

Most research judges this path unlikely. It depends on Chinese political resolve, Chinese industrial mobilisation and the US industrial base all behaving as the pessimists fear.

At a glance

Loss figures are this scenario's extrapolation from the CSIS base case (three to four weeks of fighting) [1][2] to an 18-month war. They are not CSIS figures.

Background and assumptions

What the research says. The most detailed open wargame, CSIS's The First Battle of the Next War (Cancian, Cancian and Heginbotham, January 2023), concluded that China was unlikely to occupy Taiwan in 2026 if four conditions held: Taiwan fights, Taiwan is supplied before the war, the US can use Japanese bases, and the US can strike the Chinese fleet quickly and in large numbers from outside its defensive zone [1][2]. Even then, the allies lost dozens of ships, hundreds of aircraft and tens of thousands of servicemembers, and CSIS concluded that "victory is not enough" [1]. CSIS deliberately modelled only the first weeks, so its findings say little about a longer war.

The munitions problem. The CSIS study Empty Bins in a Wartime Environment (Seth G. Jones, January 2023) found the US would likely run out of some long-range precision munitions in under a week of a Taiwan conflict, and that a protracted war becomes an industrial war [3]. The congressional Commission on the National Defense Strategy (July 2024) found the US could "largely exhaust its munitions inventories in as few as three to four weeks", with anti-ship missiles lasting only days and replacement taking years [4]. A CSIS-run simulation for the House Select Committee on the CCP (November 2024) found that LRASM anti-ship missile stocks would be gone in 3–7 days and long-range cruise missiles within about a month. Committee members also noted that key missiles take roughly two years to produce, and that replacing a carrier would take decades [5][6].

The shipbuilding gap. A 2023 Office of Naval Intelligence slide, confirmed as authentic by the Navy, put Chinese shipyard capacity at about 23 million tons against under 100,000 for the US, roughly 232 times as much. The Navy cautioned that the slide was not a detailed analysis, and most of that capacity is commercial [7]. The Congressional Research Service (April 2025) reports DoD's count of more than 370 PLAN battle force platforms, projected to reach 395 in 2025 and 435 by 2030, against 296 US Navy battle force ships in September 2024 [8]. The DoD 2024 China Military Power Report also estimates more than 600 Chinese nuclear warheads in mid-2024 (most thought to be in storage), with more than 1,000 projected by 2030 [9]. This makes it harder for either side to escalate its way out of a stalemate.

Long-war thinking. RAND's Thinking Through Protracted War with China: Nine Scenarios (2025) argues that a US–China war could last much longer than force planners assume, including direct fighting confined to third-party territory [10]. CSIS's 2025 blockade study Lights Out? (26 iterations) found that blockade options create escalation pressures and could hit Taiwan's energy system hard [11].

What must be true for this path:

1. China secures and holds a lodgement. This happened in some CSIS iterations, though in most the beachhead was eventually defeated [1].
2. Neither side uses nuclear weapons. Both sides keep strikes on the other's homeland limited: the US does not systematically attack mainland shipyards, and China does not strike the continental US.
3. China's leadership accepts very heavy casualties without the regime breaking, and puts its commercial shipbuilding and industrial base on a war footing. Chinese access to energy and food holds up, through Russian overland supply and US reluctance to impose a full distant blockade.
4. The US cannot increase munitions production in time. The Iran-related missile drawdowns reported in 2026 have also depleted inventories further before the war starts (an assumption; reported 2026 drawdowns are not independently verified here).
5. US public and allied support wears down after a year of losses with no clear path to victory.

Likelihood. Most published analysis judges a quick Chinese victory unlikely and a long war possible but highly uncertain for both sides. China's own constraints (imported energy, export dependence, inexperienced forces, the risk to the regime) are as serious as America's. This scenario is a pessimistic case, not the central estimate.

Timeline

• D-180 – After a contested election in Taiwan and a constitutional dispute, Beijing declares Taipei's new government "separatist". The PLA begins large exercises around the island. The US sends a carrier group, and Taiwan raises its readiness level.
• D-120 – China quietly expands stockpiles of fuel, grain and key metals. Shipyard workforces shift toward naval repair. US intelligence notices but cannot confirm intent. Congress debates supplemental munitions funding.
• D-60 – The US announces multiyear munitions contracts. Industry says new LRASM and JASSM-ER output will take about two years to arrive [5][6]. Beijing reads this as a closing window.
• D-14 – The China Coast Guard begins "customs inspections" of Taiwan-bound shipping. Taipei refuses them. Washington warns of "grave consequences" [11].
• D-day – The PLA launches missile strikes on Taiwanese air bases, ports and command centres, and on US facilities at Okinawa and Guam. Amphibious and airborne assault on southern Taiwan follows. The US President orders intervention, and Japan authorises combat operations from its territory [1][2].
• D+1 – US submarines and bombers with JASSM-ER and LRASM strike the invasion fleet. China loses dozens of amphibious ships. A US carrier is badly damaged east of Taiwan [2].
• D+3 – Taiwan's ground forces contain the lodgement near Tainan but cannot eliminate it. China lands follow-on forces using roll-on/roll-off ferries [1].
• D+5 – INDOPACOM reports its LRASM inventory almost used up, as the 2024 committee simulation predicted [5][6]. The US shifts to submarine torpedoes and shorter-range weapons.
• D+9 – Two US carriers are lost to ballistic and cruise missile attacks. US surface ships pull back east of the Ryukyus. China declares a "closure zone" around Taiwan [2].
• D+14 – Japan has lost about 100 aircraft and several destroyers. Tokyo continues but restricts US strikes from Japanese soil on the Chinese mainland [2].
• D+21 – The CSIS base-case outcome point: the invasion has stalled, with very heavy losses on both sides [1][2]. China has not been driven out. Beijing announces a "people's war of resistance" and full industrial mobilisation.
• D+30 – US JASSM-ER and Tomahawk stocks in theatre are largely spent, as the 2024 committee simulation projected for long-range cruise missiles [5][6]. The Pentagon begins rationing long-range munitions and diverting Patriot and SM-6 interceptors from Europe and the Middle East.
• D+45 – Taiwan's power grid fails under bombardment and fuel shortage, as the Lights Out? scenarios warned [11]. Civilian deaths rise. US resupply runs reach Taiwan only intermittently by submarine and stealthy airlift.
• D+60 – Chinese shipyards begin returning damaged destroyers to service. Converted merchant hulls arrive as missile carriers and minelayers [7][8]. The US Navy can only repair ships, not replace them.
• D+75 – The US expands the distant blockade of Chinese seaborne trade at the Malacca Strait. China increases oil and grain imports overland from Russia and Central Asia. India refuses to close its sea lanes to Chinese shipping.
• D+90 – The US Congress passes a war production act. The defense industry warns that significant increases in munitions output will take years [3][4].
• D+120 – Chinese anti-ship missiles sink a third US carrier near the Philippine Sea. Washington stops operating carriers inside the second island chain.
• D+150 – A ground-war stalemate sets in on Taiwan. PLA forces hold the southwestern plain, and Taiwan holds the north and the mountains. Both sides suffer heavy infantry losses.
• D+180 – Chinese exports collapse and unemployment rises. Beijing tightens censorship and blames "US aggression". Regime stability holds through nationalist mobilisation. This is a key assumption of the scenario.
• D+210 – The US begins taking new submarines and munitions from the expanded production lines. Losses still outpace replacements [8].
• D+240 – Russia increases pressure in Europe. NATO allies refuse further transfers of air-defence interceptors to the Pacific.
• D+270 – China rebuilds its amphibious lift with commercial roll-on/roll-off vessels and lands a second wave of reinforcements in southern Taiwan. US submarines sink many of them but cannot stop them.
• D+300 – Opinion polls show a majority of Americans favour a negotiated end. Congress debates war funding renewal. Japan's government faces a no-confidence vote.
• D+330 – Taiwan's government relocates key ministries to the northeast. Fuel and food rationing are severe. The leadership privately asks Washington for a timetable for relief.
• D+360 – The UN Secretary-General, backed by Brazil, Indonesia and South Africa, calls for a ceasefire. Beijing signals openness on terms that "respect sovereignty". The US rejects this publicly but agrees to back-channel talks through Singapore.
• D+400 – A large Chinese missile salvo damages Yokosuka and Sasebo. Tokyo suspends US combat operations from Japan pending talks. This is the decisive allied fracture.
• D+430 – Without Japanese bases, US air operations fall back to Guam and Australia. The CSIS conditions for defending Taiwan no longer hold [1][2].
• D+460 – China offers a ceasefire in place: Taiwan's "autonomous administration" continues, with unification talks and demilitarisation. Taipei rejects it at first.
• D+500 – The US President announces that America "will not fight a war without end". Washington presses Taipei to negotiate.
• D+520 – Ceasefire takes effect. PLA forces stay on Taiwan, and US forces withdraw east of the Ryukyus.
• D+540 – Settlement signed in Singapore. Taiwan becomes demilitarised and neutral and enters unification "consultations" under Chinese military presence. The US keeps its alliances with Japan, Australia and the Philippines but withdraws from the Taiwan Strait.

World politics and allies

Japan. Japan is the most important ally in every wargame, because without bases there the US loses [1][2]. In this scenario Japan fights for over a year at high cost. It then restricts and finally suspends US combat use of its bases after direct strikes on its home islands. Afterwards it begins a large rearmament and an open debate on nuclear weapons.

Australia. Australia provides bases (Darwin, Tindal) and submarine support but has small munitions stocks of its own. After Japan's suspension, Australia becomes the main US staging area. Canberra supports the settlement because it is the least bad option.

Philippines. The Philippines grants access to its northern bases early, but Chinese missile strikes on them push Manila toward neutrality by the second half of the war.

South Korea. Seoul is constrained by North Korea's posture and refuses combat roles. It provides shipyard repair capacity, which partly offsets the gap with China [7], and some munitions.

NATO/EU. European states provide diplomatic support and some industrial capacity. Russian pressure in Europe and their own empty stockpiles (flagged by the NDS Commission [4]) limit transfers. The EU continues trade restrictions on China but cannot replace US war material.

India. India stays neutral, refuses to join a maritime blockade, and buys discounted Chinese and Russian goods. It gains strategic room as both great powers weaken.

China's partners. Russia supplies energy, food and some components overland, which makes China's long war sustainable. It also ties down NATO. North Korea threatens South Korea and Japan without attacking. Iran and its partners raise pressure in the Gulf, drawing off US interceptors. Most Global South states call for a ceasefire and avoid sanctions on China.

UN. China's veto blocks Security Council action. The General Assembly passes ceasefire resolutions. A neutral group (Brazil, Indonesia, South Africa, Singapore as host) provides the diplomatic route to the settlement.

How China wins

China does not conquer Taiwan outright. It wins by outlasting the United States. The end state is a negotiated settlement:

• a ceasefire in place, with PLA forces on Taiwanese soil;
• Taiwan's demilitarisation and neutrality, with unification "consultations" on Beijing's timetable;
• an end to US arms sales and to US military operations west of the Ryukyus.

Why the US accepts it. There are four reasons:

1. Munitions exhaustion. The weapons the US needs to sink ships from a distance ran out in weeks, and replacements arrive at a pace measured in years [3][4][5].
2. The fleet cannot be replaced. Each lost warship is gone for the war, while Chinese yards keep producing [6][7][8].
3. Allies fall away. Once Japan suspends base access, the CSIS conditions for success no longer hold [1][2].
4. Domestic politics and escalation risk. Breaking the stalemate would require large strikes on mainland China against a nuclear power with a growing arsenal [9]. After a year with no prospect of victory, the American public and Congress will not accept that risk.

The settlement leaves US alliances in place but damaged. China becomes the leading military power in the Western Pacific. Many states read the outcome as a sign that America's security guarantees in Asia can be outlasted.

Losses

All figures beyond the CSIS three-week base case are scenario estimates, not published findings.

What would change the outcome

1. Munitions depth before the war. Larger stockpiles of LRASM, JASSM-ER, Tomahawk and SM-6, and production lines already running at wartime rates, would remove the "empty bins" problem the scenario relies on [3][4][5].
2. Affordable mass and allied production. Cheap drones, mines and missiles, plus co-production in Japan, South Korea and Australia, would let the US keep up the fight without depending on scarce high-end weapons. Using allied shipyards for repair would narrow the shipbuilding gap [7].
3. Japan holds. If Japan keeps allowing combat operations from its bases throughout, the CSIS conditions for defending Taiwan stay in place [1][2].
4. China's own limits. If China's energy imports, export economy or regime stability break under a long war and a distant blockade, attrition works against Beijing. Analysts judge this at least as plausible as the US breaking first [10].
5. Taiwan's preparation. Stockpiled fuel, food and munitions, distributed power and a strong territorial defence would let Taiwan hold out longer than China can sustain a lodgement [1][11].

References

1. Mark F. Cancian, Matthew Cancian, Eric Heginbotham, "The First Battle of the Next War: Wargaming a Chinese Invasion of Taiwan", CSIS, 9 January 2023. https://www.csis.org/analysis/first-battle-next-war-wargaming-chinese-invasion-taiwan
2. Radio Free Asia, "Chinese invasion of Taiwan would fail, but at a high cost for all, CSIS says", 10 January 2023 (summary of base-case losses from [1]). https://www.rfa.org/english/news/southchinasea/csis-taiwan-01102023061038.html
3. Seth G. Jones, "Empty Bins in a Wartime Environment: The Challenge to the U.S. Defense Industrial Base", CSIS, 23 January 2023. https://www.csis.org/analysis/empty-bins-wartime-environment-challenge-us-defense-industrial-base
4. Patrick Tucker, "The US might lose a war with China, congressional commission says", Defense One, 29 July 2024 (on the Commission on the National Defense Strategy report, July 2024). https://www.defenseone.com/threats/2024/07/us-might-lose-war-china-congressional-commission-says/398418/
5. House Select Committee on the CCP, "CSIS Simulation Results" (slide deck, November 2024). https://chinaselectcommittee.house.gov/sites/evo-subsites/selectcommitteeontheccp.house.gov/files/evo-media-document/CSIS%20Simulation%20Results.pdf
6. Adam Kredo, "In a War Against China, the US Runs Out of Missiles in a Matter of Weeks, House Committee Finds", Washington Free Beacon, 9 December 2024. https://freebeacon.com/national-security/in-a-war-against-china-the-us-runs-out-of-missiles-in-a-matter-of-weeks-house-committee-finds/
7. The War Zone (TWZ), "Alarming Navy Intel Slide Warns Of China's 200 Times Greater Shipbuilding Capacity", 11 July 2023 (ONI slide, confirmed authentic by the Navy). https://twz.com/alarming-navy-intel-slide-warns-of-chinas-200-times-greater-shipbuilding-capacity
8. Ronald O'Rourke, "China Naval Modernization: Implications for U.S. Navy Capabilities—Background and Issues for Congress", Congressional Research Service RL33153, 24 April 2025. https://www.everycrsreport.com/reports/RL33153.html
9. US Department of Defense, "Military and Security Developments Involving the People's Republic of China 2024", 18 December 2024; as analysed by Hans Kristensen et al., Federation of American Scientists, "The 2024 DOD China Military Power Report", December 2024. https://fas.org/publication/the-2024-dod-china-military-power-report/
10. RAND Corporation, "Thinking Through Protracted War with China: Nine Scenarios", February 2025 (publisher's product page). https://shop.rand.org/products/thinking-through-protracted-war-with-china-nine-scenarios
11. Mark F. Cancian, Matthew F. Cancian, Eric Heginbotham, "Lights Out? Wargaming a Chinese Blockade of Taiwan", CSIS, 31 July 2025. https://www.csis.org/analysis/lights-out-wargaming-chinese-blockade-taiwan

← The Ayungin Withdrawal: a Second Thomas Shoal clash that the United States chose not to widen The Systems War: Coercion Without Invasion →