In August 2022, a visit to Taiwan by U.S. House Speaker Nancy Pelosi had set off the worst cross-strait crisis in a quarter century. China’s aircraft barreled across the center line of the Taiwan Strait; its ships prowled the waters around the island; its ballistic missiles splashed down in vital shipping lanes. Months after Russia’s full-scale invasion of Ukraine had reminded everyone that major war is not an anachronism, the Taiwan crisis made visceral the prospect that a Chinese attack on that island could trigger conflict between the world’s two top powers.

Washington certainly took note. A year earlier, the outgoing chief of U.S. Indo-Pacific Command, Adm. Philip Davidson, had predicted that a war in the Taiwan Strait could come by 2027. After the August crisis, this “Davidson window” became something like conventional wisdom, with Minihan, Secretary of State Antony Blinken, and other U.S. officials predicting that trouble might start even sooner. If the United States and China do clash over Taiwan, it will be the war everyone saw coming—which would make the failure to deter it all the more painful.

To be sure, U.S. President Joe Biden has made deterring that conflict a priority. Despite the long-standing policy of “strategic ambiguity,” Biden has publicly affirmed, four times, that the United States would come to Taiwan’s aid if it were attacked. Yet deterrence is about more than declaratory policy: It requires assembling a larger structure of constraints that preserve the peace by instilling fear of the outcome and consequences of war. More than a year after the August crisis and nearly three years into the Davidson window, the United States and its friends are struggling to build that structure in the limited time they may have left.

Taiwan is a “lock around the neck of a great dragon,” as Chinese military analyst Zhu Tingchang has written. It anchors the first island chain, the string of U.S. allies and partners that block China from the open Pacific. If China were to take Taiwan, it would rupture this defense perimeter, opening the way to greater influence—and coercion—throughout the region and beyond.

In 1972, Chinese leader Mao Zedong told U.S. President Richard Nixon that Beijing could wait 100 years to reclaim Taiwan. China’s current leader, Xi Jinping, is not so patient. He has said the island’s awkward status cannot be passed from generation to generation; he has reportedly ordered the People’s Liberation Army to be ready for action by 2027. Militaries constantly prepare for missions they never execute, of course. But the risk of war is rising as China’s capabilities—and urgency—grow.

A great-power war over Taiwan would be cataclysmic. It would feature combat more vicious than anything the United States has experienced in generations.

Beijing is reaping the rewards of a multidecade buildup focused on the ships, planes, and other platforms needed to project power into the Western Pacific; the “counter-intervention” capabilities, such as anti-ship missiles and sophisticated air defenses, needed to keep U.S. forces at bay; and now the nuclear capabilities needed to enhance China’s options for deterrence and coercion alike. The scale and scope of these programs are remarkable. Adm. John Aquilino, Davidson’s successor at Indo-Pacific Command, said in April that China has embarked on “the largest, fastest, most comprehensive military buildup since World War II.” As a result, the balance is changing fast. By the late 2020s, several recent assessments indicate, Washington might find it extremely hard to save Taiwan from a determined assault.

Xi would surely prefer to take Taiwan without a fight. He currently aims to coerce unification through military, economic, and psychological pressure short of war. Yet this strategy isn’t working. Having witnessed Xi’s brutal crackdown in Hong Kong, the Taiwanese populace has little interest in unification. Since 2016, the more hawkish, pro-independence Democratic Progressive Party (DPP) has thumped the more Beijing-friendly Kuomintang in presidential elections. If the DPP wins the next presidential race in January 2024—its candidate, Lai Ching-te, currently leads the polls—Xi might conclude that coercion has failed and consider more violent options.

Biden knows the threat is rising—he recently called China a “ticking time bomb”—which is why he has repeatedly said Washington won’t stand aside if Beijing strikes. But make no mistake: A great-power war over Taiwan would be cataclysmic. It would feature combat more vicious than anything the United States has experienced in generations. It would fragment the global economy and pose real risks of nuclear escalation. So the crucial question is whether Washington can deter a conflict it hopes never to fight.

China’s fundamental advantages are proximity and the mass of forces it can muster in a war off its coast. The U.S. advantage is that control is harder than denial, especially when control requires crossing large contested bodies of water. An invasion of Taiwan, with its oceanic moat and rugged terrain, would be one of history’s most daunting military operations, comparable to the Allied invasion of Normandy in 1944. Options short of invasion, such as blockade or bombardment, offer no guarantee of forcing Taiwan to submit. Given the risk that a failed war could pose to Xi’s regime and perhaps his life, the Chinese leader will probably want a high chance of success if he attacks. So the United States and other countries should be able to inject enough doubt into this calculus that even a more risk-acceptant Xi decides rolling the iron dice is a bad idea.

This will require two mutually reinforcing types of deterrence. “Deterrence by denial” convinces an enemy not to attack by persuading him that the effort will fail. The ability to deter invasion, in this sense, is synonymous with the ability to defeat it. “Deterrence by punishment” convinces an enemy not to attack by persuading him that the effort—even if successful—will incur an exorbitant price. The strongest deterrents blend denial and punishment. They confront an aggressor with sky-high costs and a low likelihood of success. The U.S. task in the Western Pacific, then, is to show that Taiwan can survive a Chinese attack—and that any such war will leave China far poorer, weaker, and less politically stable than before.

In practice, this approach would rest on five pillars: first, a Taiwan that can deny China a quick or easy victory because it is bristling with arms and ready to resist to the end; second, a U.S. military that can sink a Chinese invasion fleet, decimate a blockade squadron, and otherwise turn back hostile forces trying to take Taiwan; third, a coalition of allies that can bolster this denial defense while raising the strategic price China pays by forcing it to fight a sprawling, regionwide war; fourth, a global punishment campaign that batters China’s economy—and perhaps its political system—regardless of whether Beijing wins or loses in the Taiwan Strait; and fifth, a credible ability to fight a nuclear war in the Western Pacific—if only to convince China that it cannot use its own growing arsenal to deter the United States from defending Taiwan.

If this sounds like a tall order, it is. Deterring determined revisionists is never easy. If these steps sound awful to contemplate, they are. Deterrence involves preparing for the unthinkable to lessen the likelihood it occurs. The United States and its friends are making real, even historic progress in all these areas. Alas, they are still struggling to get ahead of the threat.

In fairness, Taiwan faces an epic task in hardening itself against its hulking neighbor. To do so, it has adopted a smart, asymmetric defense concept that emphasizes using “large numbers of small things,” as former U.S. defense official David Helvey termed it—sea mines, anti-ship missiles, mobile air defenses—to slow and attrite Chinese forces; it is building an army that can surge troops to invasion beaches; and it is raising a reserve force that can fight guerrilla-style in Taiwan’s complex terrain. The United States is selling—and, now, simply giving—Taiwan missiles, drones, and other weapons to hasten this transformation. It is quietly increasing its training and advisory presence on the island. Given time, Taiwan can make itself a prickly porcupine. The question is how much time that will take.

Taiwan’s promising defense reforms have been dogged by political and bureaucratic opposition, just as U.S. arms sales have lagged for years due to backlogs in the military supply pipeline. Yet the underlying problem is more fundamental. It is hard to claim that a country that spends just 2.4 percent of its GDP on defense, that is only slowly preparing the sort of all-of-society resistance that has sustained Ukraine, and whose military spends precious dollars on expensive, easy-to-kill capabilities that could be useless in the event of war is entirely serious about its own defense. According to the Rand Corp., Taiwan’s ability to hold out until help arrives is becoming more tenuous—which will make it a more tempting target for Beijing.

The United States reportedly lacks enough anti-ship missiles and other munitions to blunt the first Chinese attack, let alone keep fighting after a few days or weeks.

For the U.S. military, the story is also one of smart reforms and glaring weaknesses. The Pentagon is doing many of the right things to turn geography against Beijing by transforming the Western Pacific into a killing zone for attacking forces: buying more missiles and munitions, hardening its bases and learning to disperse its forces, investing in loitering shooters and sensors, exploring creative ways of delivering firepower from longer ranges, and even making the Marine Corps into a ship-killing force that operates from tiny islands. As new capabilities, such as a next-generation stealth bomber, and new basing opportunities come online in the late 2020s and 2030s, the United States may stand a good chance of stymying a Chinese attack. Yet these changes are still years or more from fruition, and striking deficiencies remain.

Modern combat remains a matter of mass. Recent investments aside, the United States reportedly lacks enough anti-ship missiles and other munitions to blunt the first Chinese attack, let alone keep fighting after a few days or weeks of high-intensity combat. Amphibious ships, attack submarines, and other critical platforms are all too scarce. Rapidly surging production of any of these capabilities is difficult, thanks to decades of disinvestment in the defense industrial base—and because even now, defense spending is roughly as low, relative to GDP, as at any time since World War II. As aging ships, planes, and submarines are retired in the late 2020s, in fact, U.S. firepower in the Western Pacific will decline, just as China’s current military reforms reach fruition. The Pentagon is working hard to address the China challenge, but it is still a long way from closing the window of vulnerability that is opening up.