When China’s military launched its “Justice Mission 2025” exercises last December, simulating a blockade of Taiwan’s major ports and air routes, hundreds of civilian flights between Taiwan and its outlying islands were disrupted. For thousands of stranded passengers, the drills offered a brief, unsettling glimpse of what a real crisis might feel like.

This article is adapted from Defending Taiwan: A Strategy to Prevent War with China, Eyck Freymann, Oxford University Press, 432pp., $29.99, April 2026

They also exposed a question few in Washington want to confront: What happens to the hundreds of thousands of foreign civilians living in Taiwan—including roughly 11,000 Americans—when the next crisis isn’t a drill?

The honest answer is that no one knows—because no one has a plan.

The question matters because the most likely path to a Taiwan crisis is not a sudden invasion, but a slow-rolling, gray-zone escalation over the course of weeks or months. There would be time for Taipei, Washington, and others to act, but every action taken by the U.S. president, including the decision about whether to evacuate U.S. civilians, would itself become a signal in the escalation spiral.

Simply put: A crisis over Taiwan might not begin with the first shot. It could begin with the first flight out.

In a grim sense, the presence of foreign civilians in Taiwan is one of the strongest deterrents against a Chinese surprise attack. As of April 2023, Taiwan was home to more than 800,000 foreign residents representing 164 countries, plus an unknown number of Chinese nationals. Some 700,000 of these foreigners come from Southeast Asia, but there are also substantial communities of Americans, British, Canadians, French, Germans, Indians, Japanese, Singaporeans, and South Koreans, among others.

If Chinese strikes killed or injured large numbers of these civilians or made it impossible for them to be safely evacuated, Beijing would risk drawing their home countries into the conflict. For that reason, it would likely prefer to pressure foreign governments into pulling their people out of Taiwan before it made any military move.

But extracting those populations would be an enormous undertaking—and one the United States and its allies are strikingly unprepared for. A noncombatant evacuation operation (NEO) from Taiwan would dwarf anything the United States and its allies have ever attempted.

The Defense Department’s joint publication on NEOs lays out an elaborate doctrine for how the process is generally supposed to work, involving coordination between the State Department and the Pentagon, agreements with host countries, and staged phases of evacuation. None of this has been adapted for a Taiwan scenario. In 2023, the House Armed Services Committee asked the Pentagon to draw up plans for a potential Taiwan NEO—strongly suggesting that no such plan existed at that time. There is no public evidence that the U.S. government has since acted on this recommendation.

None of the countries with significant populations in Taiwan currently has anything approaching the resources and doctrine to quickly withdraw its nationals. The United States and its allies do not have a joint doctrine for NEOs, nor a framework for coordinating one. In a crisis, each country would essentially be improvising under the most stressful conditions imaginable.

History offers little reassurance.

The 1975 U.S. airlift out of Saigon, one of the largest emergency evacuations in U.S. history, carried just 45,000 people, including 5,600 Americans. That effort, however, still left thousands of refugees and many citizens behind—including the U.S. ambassador and his staff. An additional 7,800 people had to be evacuated by helicopter in the final two desperate days. The South Vietnamese government collapsed in the process, and the images of helicopters lifting off the embassy roof became an enduring symbol of the United States’ failure.

In the 2021 evacuation of Kabul, some 122,000 people were airlifted over 16 days. U.S. forces already had a large combat presence in and around the Afghan capital—a military footprint that would not exist on or near Taiwan when a crisis began. Even so, the evacuation was chaotic, and many Afghans were left behind to meet their fate at the hands of the Taliban. This became a political catastrophe for then-U.S. President Joe Biden, in part because it played out in full view of cameras.

That experience also forced officials to confront the limits of such operations outside similarly permissive conditions. “I think there may be a misperception that what we saw in Afghanistan is something that the U.S. government can undertake anywhere and everywhere in the world,” State Department spokesperson Ned Price said in the aftermath. In other contexts, referring to Ukraine at the time, he said no American “should expect that we may be in a position to undertake something similar to what we saw in Afghanistan.”

A Taiwan evacuation would unfold over far greater distances and under far less permissive conditions than any of these operations. And unlike Ukraine, where millions fled overland to neighboring countries following Russia’s full-scale invasion in 2022, Taiwan’s geography makes air evacuation the only viable option at scale. The island lies roughly 100 miles from the Chinese mainland and about 400 miles from the nearest major U.S. base on Okinawa. Moreover, the United States has had no organic strategic troop sealift capability since 1973, and the U.S.-flagged commercial passenger-shipping base that historically backstopped it has essentially disappeared.

In a crisis, hundreds of thousands—if not millions—of Taiwanese nationals would likely also attempt to flee before hostilities began, competing with foreigners for limited flights. Even under the most optimistic assumptions, it would still take weeks to evacuate all foreign nationals. Meanwhile, Chinese military pressure would steadily make Taiwan’s airspace less permissive and commercial carriers less willing to operate.

Taiwan might declare martial law, banning men of fighting age from leaving, as Ukraine did. Its leaders might also interpret a U.S.-led evacuation as a sign of abandonment, reducing their incentive to cooperate. Under such strain, political collapse is not unthinkable, just as the South Vietnamese and Afghan governments collapsed during evacuations. Refugees could attempt dangerous sea crossings to Japan’s southwestern islands, overwhelming local communities without capacity to absorb them.

Taking all of these uncertainties into account, the humanitarian and political consequences would reverberate across the entire Indo-Pacific.

As tensions rose, the U.S. president would face a decision with no good options.

Encouraging Americans to leave Taiwan voluntarily while a conflict still appeared weeks or months away, as Washington did before Russia’s full-scale invasion of Ukraine, might seem prudent. But Taiwan is not Ukraine, and the playbook for one won’t necessarily work for the other. Ensuring the safety of its citizens abroad is a fundamental obligation of the U.S. government, particularly when the host country is unable or unwilling to protect them. But this must be balanced against the destabilizing effect that an evacuation order could have on the crisis itself.

An evacuation order would signal imminent conflict, which Beijing could interpret as a sign of U.S. hesitancy, encouraging further escalation. This dilemma illustrates why NEOs are so difficult to execute. Keeping civilians in place can, paradoxically, serve as a form of deterrence: Their presence forces Beijing to factor in the risk of killing or injuring the citizens of dozens of countries, any of which might enter the war in response. But leaving them at risk carries obvious moral and political costs that cannot be set aside in the name of strategic logic.

The president would face a true quadrilemma, with each option carrying devastating consequences:

Evacuate early and risk inviting escalation, alienating Taiwan’s government and public in the process.

Evacuate late—when flights may already be canceled, airports overwhelmed, and China’s naval cordon tightening—and risk a chaotic, highly visible Kabul-scale disaster, with the spectacle of Americans scrambling to leave a besieged island playing out on live television.

Refuse to evacuate, leaving Americans exposed to harm. This option preserves the deterrent value of foreign civilians on the island, but at an extraordinary moral and political cost. If Americans died in Chinese strikes, the domestic political fallout would be severe. If they were detained or used as bargaining chips, Washington would face a hostage crisis layered on top of a geopolitical one.

Or attempt to evacuate quietly, hoping to reduce exposure without signaling retreat. But in the age of social media, nothing stays quiet for long. Flight bookings, embassy advisories, and troop movements would leak within hours. A botched attempt at a discreet drawdown could combine the worst of all worlds: the strategic damage of signaling withdrawal, without actually getting people to safety.

Each option that includes evacuation would also force an immediate reckoning among U.S. allies, particularly those with citizens on the island, who would face domestic pressure to do the same. This could trigger a cascading international withdrawal that would devastate Taiwan’s morale and economy long before any shots were fired. And because there is no joint allied evacuation framework, each country would be making its decisions in an information vacuum, with minimal coordination and maximum potential for confusion.

None of this is theoretical. Since 2022, China’s military has conducted a series of increasingly realistic exercises targeting Taiwan, simulating blockades, missile strikes, and encirclement operations. It has steadily pushed closer to the island each time, the line between exercise and operation growing thinner with every iteration.

The evacuation question is neither a secondary concern nor a logistics problem to be solved after a crisis begins. It is inseparable from the larger question of whether the United States has a coherent strategy for the gray zone between peace and war where China already operates. If Washington lacks a plan for something as basic as protecting its own citizens on the island, what does that imply about its readiness to manage financial shocks, coordinate with allies, or communicate resolve to Beijing as a crisis unfolds? The evacuation gap is a symptom of a deeper failure: the absence of an integrated strategy that treats political, military, and economic deterrence as a single problem.

The question could define a presidency. It could determine whether the United States enters a war or stands aside. It could shape whether allies rally or fracture. And it is a question that Washington, by its own admission, has barely begun to answer. The time to build a plan is before the next drill becomes the real thing.