Why friends, self-investment, and openness make you safe
We have been looking at the American Nationalist Hawk and his critiques of free trade in general, and trade with China in particular. The first week we looked at how engagement with China did not in fact strengthen the Chinese Communist Part at our expense but strengthened the Chinese economic periphery. The second week we looked at how trade deficits are not like bank accounts where what comes in and what goes out need to match. And today we are going to look at the charge that the perceived loss of manufacturing in the United States is a national security concern.
In recent years, Washington has revived an old war cry: that outsourcing America’s manufacturing base has become a clear and present danger to national security. It’s a charge leveled with increasing volume by senators, trade officials, think tank fellows, and primetime pundits alike. The target is not just China, but the very idea that global supply chains can coexist with national strength.
Former U.S. Trade Representative Robert Lighthizer calls it “economic madness” to rely on adversaries for core manufacturing inputs (Lighthizer, The Return of Mercantilism, 2021). Senator Josh Hawley declared in a Foreign Affairs essay that “American prosperity depends on American industry,” and that we must “end our dependence on Chinese manufacturing, period” (Hawley, The Case for Nationalism, 2020). Peter Navarro, Trump’s trade whisperer and author of Death by China, has gone even further, arguing that “every dollar we spend on Chinese goods is a bullet in the heart of American security” (Coalition for a Prosperous America, 2020).
These are not fringe voices. They shape policy, set headlines, and increasingly dominate the terms of debate. You hear their warnings in Senate Armed Services Committee hearings, on Heritage Foundation panels, and in the editorial pages of The Wall Street Journal. The argument is often cast in moral, even civilizational terms: to manufacture abroad is to surrender sovereignty.
There is some truth here. The COVID-19 pandemic exposed the fragility of certain global supply chains, from PPE to pharmaceuticals. Russia’s war in Ukraine reminded us that energy dependence can warp foreign policy. And China’s growing assertiveness, especially around Taiwan, has injected real urgency into questions about supply resilience.
But for all its patriotic fervor, the hawkish argument is too often built on a misdiagnosis. It assumes that foreign manufacturing equals foreign control, that outsourcing equals vulnerability, and that reshoring alone can fix the problem. Most troubling, it treats tariffs as a cure-all—when in fact, they are a blunt and costly instrument.
So let’s move beyond slogans and break down a few strategic sectors: microchips, steel, and artificial intelligence. What do the hawks get right? Where do they miss the mark? And what would a smarter, more effective national security strategy look like?
Microchips: Designed in America, Built with Allies
Semiconductors are often called the “new oil” of the digital age, and for good reason. They power everything from smartphones to missile defense systems. So it’s not surprising that policymakers are nervous about where they come from.
The hawks say: We rely too much on Taiwan. What if China invades? What if our chip supply gets cut off?
That’s not a crazy fear. But the chip supply chain is more resilient—and more American—than it first appears. (We discussed this at length when talking about trade deficits.)
Yes, Taiwan’s TSMC fabricates most of the world’s advanced chips. But the blueprints? As we’ve discussed, they’re American. NVIDIA, AMD, Qualcomm, and Apple all design their chips in California. The tools to make those chips—EUV lithography machines, wafer inspection tools, etching systems—are mostly made in the U.S., Japan, and the Netherlands. In short, without American IP and allied technology, Taiwan’s fabs are stunted.
The CHIPS Act was an attempt to create a security buffer around the U.S. chip supply: a kind of geopolitical insurance in silicon. It aimed to bring some semiconductor manufacturing back onshore and reduce dependence on Asia.
The motivation is understandable. Losing access—or even just a delay—could cause enormous disruptions to the U.S. economy and military.
Instead, the CHIPS Act—while noble in intent—has become a cautionary tale. Manufacturing chips in the U.S. is proving astronomically expensive. According to a 2023 report from the Semiconductor Industry Association, building and operating a new leading-edge chip fab in the U.S. costs 30–50% more than in Taiwan or South Korea. TSMC’s Arizona plant, heavily subsidized by the CHIPS Act, has faced delays, labor shortages, and costs that could drive wafer production up to 2x that of Taiwan (Bloomberg, 2023).
We’re trying to replicate in five years what Taiwan and South Korea built over the course of three decades. That’s not just ambitious, it’s structurally improbable. TSMC has spent more than thirty years building the world’s most advanced fabrication capacity, backed by a dense local supply chain, robust education pipeline, and steady government support. South Korea’s Samsung followed a similar arc, investing heavily in domestic infrastructure and talent development throughout the 1990s and 2000s.
In contrast, the United States is starting from a far less integrated base. We lack the concentrated supplier networks, the depth of skilled labor, and the institutional muscle memory that have made East Asian chip clusters so efficient. Building that from scratch will not be cheap. While the CHIPS Act earmarks $52 billion in federal subsidies, industry leaders—including Intel CEO Pat Gelsinger—have warned that full-scale reshoring could ultimately require $300 to $400 billion in public and private capital.
And what’s the strategic return on that investment? Chris Miller, author of Chip War, argues that complete self-sufficiency is both unnecessary and inefficient. The goal, he says, should be resilience through diversification, not redundancy at all costs. In other words, some reshoring is wise. But trying to recreate Asia’s fabrication ecosystem wholesale, at triple the cost, risks turning into a strategic misfire: a kind of geopolitical vanity project that drains resources without delivering real security.
Without careful calibration, we may end up with slower, more expensive, and last-generation chips while our allies continue to push ahead. The smarter move is to double down on our true advantage: design leadership, global alliances, and the ability to set the pace of innovation from the front, not the middle of the pack.
Steel: National Security or Infrastructure Deficit?
Steel is an area I know very well. And it is a topic where many hawks go off the rails. You’ll hear Josh Hawley at rallies say, “America doesn’t make anything anymore. What if we need to re-arm like it’s 1941?”
It sounds serious. But let’s take a step back and look at the numbers.
As of 2023, America still produces 80 million metric tons of steel per year, according to the World Steel Association—putting us 4th globally, behind China (~1 billion tons), India (~125 million), and ahead of Russia and Germany. That’s way more than our military needs. So no, the steel mills haven’t vanished.
The problem isn’t that we can’t make steel, it’s that we don’t use that much of it here in the U.S anymore.
Here’s a forgotten fact: U.S. steel production peaked in 1969, hitting 141 million tons. While outsourcing is often the blamed culprit, what in fact happened was that peak coincided with the tail end of the Eisenhower Highway System, a mega-project that consumed huge amounts of domestic steel. As our infrastructure spending slowed in the 1970s, our demand for steel dropped along with it.
Meanwhile, Japan ramped up. By 1973, it was producing 119 million tons of steel, and 85% of it was for domestic consumption—ports, bullet trains, bridges. We imported their surplus not because they stole jobs, but because they were out-scaling us as they invested to keep up with their nation’s infrastructure boom. They were scaling up as we were scaling down.
Then came China, repeating the playbook on an even grander scale. From 2000 to 2015, China’s fixed asset investment grew over 20% annually, driving the largest infrastructure boom in history. In 2023, China produced over 1 billion tons of steel and 90% of that was consumed domestically (China NBS, 2023). Their exports? Just 7% of total output.
That doesn’t sound like an invasion to me. It’s true that 7% is still a lot of steel at the scale China operates, but it’s important to see China’s steel mills have never existed with the American import buyer in mind; it has only ever existed for its own nation building.
The pattern is clear: countries that build more, make more. We didn’t see a fall in steel production because of foreign competition. We lost it because we stopped building our own infrastructure.
And that’s what the hawks miss. You can slap on tariffs, but if you’re not laying American track or raising American towers, American mills will stay quiet anyway.
In 2021, infrastructure made up 1.2% of U.S. GDP, compared to 5.5% in China and 3.5% in Japan (World Bank, 2023).
Double that investment in the U.S., and American steel will roar back to life.
AI: National Security Secured Through Openness
Artificial intelligence is the most unpredictable—and arguably the most consequential—technology shaping national security today. It’s not just about chatbots or deepfakes. It’s about which country builds the thinking infrastructure that will drive every other field: defense, biotech, cyberwarfare, logistics, diplomacy.
The fears are not entirely wrong. AI can reshape military decision-making, power autonomous weapons, and manage real-time cyber operations faster than any human ever could. It can also supercharge disinformation campaigns, creating realistic fake videos or manipulating online sentiment at scale. These are not science fiction scenarios, they’re tools already being tested on today’s geopolitical front lines.
And yet, for all the headlines about an AI arms race, the real question isn’t who has AI. It’s how they use it, and what kind of society shapes it.
While China is investing heavily in AI, its models are shaped by a system that prioritizes control over open inquiry. Restrictions on speech, limited access to diverse global datasets, and political red lines inevitably shape the development of Chinese AI. These systems may become highly capable in surveillance and command functions, but they will struggle to match the creativity, adaptability, and open-ended reasoning that emerge from a freer research environment.
In contrast, America’s AI ecosystem thrives on academic freedom, intellectual exchange, and open-source innovation: the same conditions that have driven our scientific leadership for a century.
So yes, while it’s understandable that the U.S. sees this as a national security issue, the most advanced foundational models—OpenAI, Anthropic, Google DeepMind—are still built here, and that’s no accident. It’s the product of open universities, global talent flows, scientific freedom, and a culture that allows dissent, experimentation, and even failure.
If the U.S. wants to lead in AI—and make sure AI reflects democratic values—it can’t do it by closing its doors.
Cutting off visas, attacking academic exchange, or choking off global collaboration in the name of “security” won’t make us safer. It just makes it more likely the next great breakthrough happens somewhere else.
The Real Security Strategy: Build, Partner, Attract
The hawks aren’t wrong to be worried. But they’re wrong about the solution.
Tariffs don’t build resilience. They don’t create new tech, fix infrastructure, or attract talent. They’re the economic equivalent of a drunken bar fight: emotionally satisfying after five shots of nationalism, but a disaster in the morning.
A real national security strategy would be to:
- Reinvest in infrastructure that uses domestic steel and materials;
- Deepen alliances with chipmakers like Japan, Taiwan, the Netherlands;
- Fund education and research to lead the next AI leap;
- Welcome global talent, especially in STEM.
Trade isn’t the security threat. Shortsightedness is.
And if we want to reclaim leadership, we must revisit the verbs that fortified US security in the first place: invest, partner, and attract.