The German high command learned a key lesson after losing World War I: Never fight a two-front war. That’s why Germany signed the 1939 Molotov-Ribbentrop pact with the Soviet Union, which stated that neither country would attack the other for a decade. But Adolf Hitler couldn’t count to 10, and Germany ended up in World War II—another two-front war that ended badly for Germany.

The same dictum goes for trade wars. It’s okay to fight a one-front war, but not a war with the whole world. As comedian Norm Macdonald joked on The Late Show in 2015, “In the early part of the previous century, Germany decided to go to war, and who did they go to war with? The world … Then about 30 years pass and Germany decides, again, to go to war. And, again, it chooses as its enemy the world!”

Now, U.S. President Donald Trump has decided to start a trade war, and who does he choose to attack? The world. On his so-called “Liberation Day” in April, Trump imposed tariffs on most countries, even islands with penguins and allies that run trade deficits with the United States.

The result is that the rest of the world appears to resent the United States now and is beginning to plan for a post-American trading system. Japan, South Korea, China, and members of the Association of Southeast Asian Nations are in talks to cooperate on trade, as are the European Union and China. Canada is also looking to the EU for closer trade and investment ties in light of Trump’s tariffs.

Regardless, Trump seems to think that the United States is still the top dog, the head honcho, the boss. But it’s time to face the facts: The United States will lose massively from his trade war.

The reason is simple. Too many U.S. companies, especially in advanced industries, need global market access to survive and thrive. The modal Trump company is a low to moderate tech and nontraded American company selling almost all its output domestically. Think kitchen knife makers, furniture producers, and golf club companies. (Does Trump know that he and his fellow country club members will soon pay more for a new set of irons?) If Trump protects these companies from overseas competition, his thinking goes, these companies will thrive. How could they not with a large protected market?

Not so fast. The United States cannot be powerful without robust advanced industries. And those industries will seriously struggle in Trump’s new world.

There are multiple vectors of challenge. The first is that U.S. firms that produce for export—think Boeing, Merck, General Electric, and others—will see their input costs increase significantly as they pay tariffs on imported components and materials.

Second, as other nations enact reciprocal tariffs, these companies’ products will be priced out of foreign markets. Companies in other countries will buy South Korean memory chips rather than those manufactured by Micron. Airbus jets, not Boeing. Hitachi machinery, not Caterpillar. You get the idea.

On top of that, Trump is opening the floodgates for blatant discrimination against U.S. firms. The EU is using Trump’s aggression to justify its “EuroStack”: a plan to eventually replace almost all U.S. tech products, from computer chips to severs, with those made in Europe. And while the bloc has finally announced plans to boost military spending, it will do so by purchasing European-made weapons.

This gets to a related challenge. While U.S. exporters under Trump’s trade war will be increasingly shut out of foreign markets, their competitors will find themselves in a globally integrated market, albeit sans the United States. That will be particularly true as other nations take steps to create more integrated markets through new trade agreements. Foreign companies will have the scale needed to innovate and thrive. By contrast, U.S. producers, dependent on the relatively smaller U.S. market, will slowly shrivel and perhaps ultimately die.

Trump’s import substitution strategy (ISI) has been tried elsewhere before. It failed. The international development community widely embraced ISI in the 1950s and ’60s as a growth strategy—and many developing nations held onto it until the 1980s and beyond. One reason it failed was that countries, even relatively large ones like Brazil, did not have big enough markets to efficiently make increasingly complex goods. A recent analysis of global industrial policies by the International Monetary Fund found that it was nations that adopted export strategies, such as South Korea and Taiwan, that thrived—not those that championed ISI.

Even though the U.S. economy is much bigger than, say, South Korea’s, today’s advanced industries require larger markets than the United States can provide to survive—if only to generate the revenues needed to pay for continuous research and development.

But there’s more. As the United States’ role as the global trade hegemon wanes, China will surely take its place. Beijing has spent the last 15 years penetrating every global institution. With Trump leading the U.S. exit from many of them—such as the World Health Organization, the Paris Agreement, and the United Nations Human Rights Council—victory is China’s.

We are already seeing, and will likely continue to see, vastly more countries traveling to Beijing and kowtowing to Chinese President Xi Jinping to cut trade deals. Brazil and Colombia, once the largest recipient of U.S. economic assistance in South America, have already gone down that road. Chinese advanced industries may not have U.S. market access after the trade war, but they will have the rest of the world’s. American companies, meanwhile, will be left with the scraps of the U.S. market.

After Trump’s drawdown of foreign aid, China is also spending much more than the United States to win the hearts and minds of other countries. It doesn’t take a weather person to know which way the wind blows: east. As Xi recently wrote in an article for a Russian newspaper, “unilateralism, hegemony and acts of bullying are inflicting severe harm” around the world. Only Trump could make China look like the defender of free trade and solidarity.

To be clear, Trump needed to go to (trade) war with nations not playing fair. The top offender is China, which started the global trade war around 2006 and has ramped it up since Xi took office. Along with some other countries, Beijing has engaged in systemic mercantilist practices, helping to hollow out U.S. manufacturing and leading to massive U.S. trade deficits.

China conducted massive intellectual property theft. It bludgeoned foreign firms, forcing them to produce in China and transfer technology. Then, it closed its markets once it gained production capabilities in a particular industry.

U.S. attempts to address China’s cheating through the World Trade Organization (WTO) have been few and far between, in part because the WTO is structured in ways that make effectively prosecuting such actions nearly impossible. U.S. firms also largely refused to cooperate with the WTO for fear of Chinese government retribution. The same goes for the EU.

The relatively few cases the United States brought and won provided mostly pyrrhic victories, such as quotas for rare-earth exports and wind power subsidies, neither of which changed competitive realities. And in most cases, China filed counter suits to U.S. suits. Trump was right: It was time to declare war. But not, as Macdonald would say, on the world all at once.

So, is there any hope? Probably not. But if Trump is one thing, it is unpredictable.

Here’s what Trump should have done: first focus on a few of the worst offenders of global trading rules, such as Vietnam, Indonesia, and India. Before imposing tariffs, invite those countries to talk, list key U.S. demands, and give them 90 days to fix them. Only impose tariffs if they refuse to respond. And while doing that, focus on the trade barriers and irritants that are most important to U.S. competitiveness in advanced industries.

If these nations want to expand their shrimp exports, then who cares? If they want to close their markets to U.S. whiskey, that’s not worth fighting about. But attacking U.S. tech firms and limiting access to advanced U.S. goods and services—that’s worth going to the mattresses over.

Then, pivot to Europe. After that, pivot to Japan, South Korea, and Taiwan. You get the picture. But don’t go to war with the whole world at the same time. All that does is create a worldwide anti-American alliance.

When China conducts its trade wars, it has enough sense to not attack every country all at once. Beijing launches a trade strike on a particular nation and then waits for a bit, gauging the foreign response; when outrage has died down, it launches another. And like a frog in boiling water, other nations have taken China’s blows with barely a whimper.

China remains the only country with the will and the means to destroy U.S.—and Western—advanced technology industries. Trump’s trade war should be designed to keep Beijing from winning. Above all, that means working with U.S. allies. But in the face of the United States’ increasing hostility and declining power, many nations find it easier to cut deals with Beijing.

Without allies, any war is lost. Unless Trump shows a willingness to negotiate with U.S. allies—something he refused to do when he met with Canadian Prime Minister Mark Carney in May—the United States will be just as globally isolated as it was before World War I. However, the big difference between then and now is that technology requires U.S. firms to attain global scale to survive and that China has emerged to outcompete them.

We can only hope that the United States emerges from its own multifront conflict better than Germany did.