Geopolitical fragmentation, the rise of AI, and growing concern about supply chain resilience are putting US manufacturing back in the spotlight. But strengthening America’s industrial base isn’t simply a matter of bringing factories home. The US imports roughly $3 trillion in manufactured goods each year, and some of its most critical supply chains would require a dramatic expansion of domestic capacity.
On this episode of The McKinsey Podcast, McKinsey Global Institute (MGI) Senior Fellow Rebecca J. Anderson and MGI Chair and Senior Partner Shubham Singhal join Editorial Director Roberta Fusaro to explore where the United States is most vulnerable, what it could cost to rebuild critical manufacturing capacity, and how AI, investment, and American innovation could change the equation.
The McKinsey Podcast is cohosted by Lucia Rahilly and Roberta Fusaro.
The following transcript has been edited for clarity and length.
Two forces driving the reshoring debate
Roberta Fusaro: There’s been a lot of conversation about bringing manufacturing back to America. Why is this such an important question right now?
Rebecca J. Anderson: We see two major trends underway in the global economy. One is growing geopolitical fragmentation, which is putting a spotlight on trade dependencies and supply chain resilience—really bringing into focus the need to build capacity, especially in critical industries. The other important dimension is the rise of AI and other critical technologies. There’s been a recognition over the past couple of decades that as the US has lost some of its manufacturing prowess in the global economy, we’re missing out on a lot of the innovation that takes place at the production level. By not having a lot of that manufacturing domestically, we’re missing out on a large part of innovation.
Roberta Fusaro: The US imports about $3 trillion in manufactured goods each year. What’s the impact of all that importation?
Rebecca J. Anderson: Of the $3 trillion worth of manufactured goods imported into the US each year, nearly a third are electronics products. Many electronics are also considered critical to national security, and they’re often from a concentrated set of countries. Take semiconductors: The vast majority of semiconductors imported into the United States come from Taiwan and South Korea. Electronics are from what we call a geopolitically distant trading partner.
The geopolitical distance metric—something we at the McKinsey Global Institute developed by analyzing UN voting records—assigns each country a score based on its geopolitical alignment. The US is at one end of that spectrum, China and Russia at the other, and every other country falls somewhere in between. Electronics is probably the biggest part of the story here, but pharmaceutical products, certain chemicals, and rare earth magnets are also important parts of it.
Roberta Fusaro: The report also says that about 25 percent of imports are potential “Achilles’ heels.” Shubham, what makes a product an Achilles’ heel?
Shubham Singhal: We looked at three things. One is how critical the product or category is. The second is the supply source—for example, does it come from just one country? That makes a big difference in the risk profile. The third is whether the supply source sits in a geopolitically distant country, meaning the two countries’ foreign policy objectives don’t align. When we look at those three factors across all product categories, if any two of the three are present, we consider that an Achilles’ heel. It’s important to note that if all three are present, that’s only 5 percent of total imports—that’s the most important to address from a resilience and national security standpoint.
Where America still leads, and where it doesn’t
Roberta Fusaro: The United States is still a major manufacturing power, but it produces far less than China at this point. Where does the US still have strength, and where has it become more vulnerable?
Shubham Singhal: We make planes in this country. We make motor vehicles and trucks—or at least the supply chains run within North America—so that whole transportation arena still has a fairly significant amount of production, which is good news because it’s quite advanced manufacturing. Second, as much as electronics is a problem for us, there are arenas where we still produce a fair amount—not whole assembled products, since assembly isn’t the biggest thing we do, but some of the advanced elements that go into it.
Rebecca J. Anderson: It’s important to emphasize that while the US has lost a large share of global manufacturing output, it’s still the second-largest manufacturer in the world, and the second-largest exporter. By no means has the US lost its entire manufacturing sector. The story of the past several decades has really been one of moving up the value chain—toward products that are higher value and more design- and IP [intellectual property]-oriented, and toward more advanced manufacturing, as opposed to labor-intensive manufacturing like textiles.
What’s also interesting is that, looking back over the past several years, computer and electronic products are actually the fastest-growing category of output in the United States. The Federal Reserve’s industrial production index, which tracks the volume of production across manufacturing industries, is up almost 40 percent over the last decade specifically in computer and electronic products.
Computer and electronic products are actually the fastest-growing category of output in the United States.
The ramp-up factor
Roberta Fusaro: In the report, you also introduce the concept of the ramp-up factor, which manufacturers and others can use to measure what it would take to meet domestic demand. What does that metric reveal, and how can manufacturers use it?
Rebecca J. Anderson: With the ramp-up factor, we looked at a granular industry level across 300 or so manufacturing subsectors in the United States and calculated the extent of domestic productive capacity that currently exists, compared with the amount we’ve imported in recent years—essentially, how much would domestic productive capacity have to scale up to fully meet domestic demand. Across all products, that ramp-up factor is about 1.3—domestic productive capacity would have to grow by about 30 percent to fully meet domestic demand.
However, for those Achilles’ heel goods—the 25 percent of total imports with multiple overlapping trade dependencies—capacity would have to double on average, and over half of them would need a ramp-up factor greater than five, meaning domestic productive capacity would need to quintuple. Our ambition in developing this metric was to bring dimension and context to the current discussion around reindustrialization.
Roberta Fusaro: Which products or sectors would be hardest to ramp up?
Rebecca J. Anderson: I’d say textiles and electronics, though they have different orders of magnitude in terms of criticality. Textiles aren’t quite as critical to national security and aren’t necessarily from a concentrated set of countries, so that’s not really where we focused in the report. But we have limited domestic production capacity for critical items like laptops, server equipment for data centers, and rare earth magnets.
Roberta Fusaro: The report suggests that running today’s existing capacity at peak could generate about $660 billion in additional output, but that it wouldn’t make a difference in the areas of largest exposure. Why isn’t existing capacity enough?
Shubham Singhal: Good question. First, if we were to ramp up to peak capacity, you’d get that additional $660 billion of output—though there are reasons we may not have enough orders to fill that space, including cost differentials. Unfortunately, that $660 billion of ramp-up doesn’t actually solve the Achilles’ heel problem—the 25 percent we spoke about earlier—because there’s a mismatch between where that Achilles’ heel sits and where the extra capacity is available. If it’s not enough to meet the Achilles’ heel need, we have to think about solving it differently from just running existing factories at full capacity—that will take real investment to build out.
Roberta Fusaro: According to the report, building that capacity could cost up to $2 trillion. What are the biggest obstacles to generating that investment, beyond capital?
Rebecca J. Anderson: Specialized skills—having a talent pipeline of people with the right training immediately available to take these positions as industries ramp up—are a significant challenge. So is having the energy and infrastructure in place. Those will have to be grappled with alongside any large-scale reindustrialization effort. It’s also important to note that the $2 trillion figure includes upstream manufacturing supply chains for all the Achilles’ heel products, but not the energy and infrastructure build-out that would have to go with it—so realistically, that number is probably going to be higher. And there need to be shovel-ready projects: permitting processes can take a long time, and there’s often a big gap between the intention to do a project and when it’s actually ready to go, with machines running and people working.
Shubham Singhal: We’ve talked to investors who ask: What is my return on invested capital here versus the many other arenas in which I can invest? MGI has published on the next arenas of competition where value creation is happening: Significant investment is going into AI and related infrastructure, information and communication technologies, and so on. So the question becomes, if you’re going to make $2 trillion of investment in the private sector, while it’s feasible, what’s the reason to do so relative to other opportunities where significant investment is being made? That’s a practical reality that has to be thought through as we try to incentivize the growth of production capacity in the United States.
Can AI make reshoring pay off?
Roberta Fusaro: Could AI help make reshoring more economically viable in ways that weren’t possible even ten years ago?
Shubham Singhal: How much we can move the needle in manufacturing with AI and robotics—on a fully loaded basis—to make the humans working there far more productive is a question. Can we better optimize production in plants and networks using the latest technologies? It’s not a silver bullet. It’s a series of applications across various areas where the US has a higher cost structure or factors that reduce our returns, and in each one, you’d have to ask: How would we deploy AI to advance them?
Rebecca J. Anderson: Fundamentally, any manufacturing of the future is going to have to look quite different, especially in a place like the United States, where labor costs are higher. Greater AI adoption and greater use of robotics will have to be part of the business case equation.
Managing choke points
Roberta Fusaro: There are certain products or inputs—rare earth magnets, for example—that are essential for downstream industries but subject to economic choke points. How should manufacturers think about that kind of risk going forward?
Shubham Singhal: We’re seeing this in our work on global trade flows. The geopolitical distance that trade travels is falling, and the geopolitical distance of foreign direct investment is falling even faster, which is a leading indicator. At a minimum, it will require “friendshoring” of entire parts of the supply chain needed for these critical industries.
At a minimum, it will require “friendshoring” of entire parts of the supply chain needed for these critical industries.
Beyond that, because additional choke points can emerge—where flow even between allies gets blocked—you have to think about ways investment can happen onshore, or very close to it. I refer to an interesting playbook from the COVID era: Vaccines were developed very quickly. The government showed up early as a guaranteed buyer for a successful vaccine. That creates a significant incentive for companies to deploy capital fast—pharma companies didn’t wait to develop before setting up manufacturing; they did both in parallel.
Apply that now to something like rare earths: Guarantees on pricing floors reduce uncertainty and increase companies’ willingness to lean in and significantly increase capacity. There will be arenas that have to be addressed both by reducing the geopolitical distance that goods travel and, in some instances, by bringing production onshore—and figuring out what incentives are needed to make that happen.
Next steps for CEOs
Roberta Fusaro: If you were advising a manufacturing CEO on supply chain resilience, what are the first things they should start doing?
Rebecca J. Anderson: A general idea we introduce in this report, and in other recent MGI work, is that when it comes to trade dependencies, we really need an all-of-the-above approach. That could mean ramping up domestic capacity. It could mean rearranging trade—sourcing from a wider set of countries to build in more resilience and redundancy, even if it’s not the most efficient thing from a bottom-line perspective. It could also mean reducing or replacing certain imports by innovating the way things are made. There’s work being done on whether we can innovate our way around a rare earths dependency—are there other materials we could use instead? It’s highly uncertain, and you can’t put a timeline on it, but historically, that’s where the US economy has shone the most.
If we look at the last 250 years of US history, the US has been extremely innovative and entrepreneurial. In some separate work we did looking at the top 100 inventions over that period, about three-quarters had some degree of US involvement. Just because things are made a certain way today doesn’t mean they’ll always be made that way, or that we’ll even be using the same products in a rapidly advancing technological climate. So an all-of-the-above approach really needs to be considered.
Roberta Fusaro: What questions are going unnoticed among the broader business population about manufacturing?
Shubham Singhal: First is optimism—the US has not lost its manufacturing. As Rebecca mentioned, we’re the second-largest manufacturer and second-largest exporter, with some critical industries intact. Second, there’s a lot of focus on the $1.2 trillion trade deficit, but we’d posit that the $750 billion Achilles’ heel is the important issue to focus on, not the gross deficit. Third, we’ve discussed the details of how, where, and whether that $2 trillion in investment can be deployed. But what cuts through it is that if the signal is clear that the United States needs to make something, it will figure out how to make it—really fast, really well—because of innovation.
If the signal is clear that the United States needs to make something, it will figure out how to make it—really fast, really well—because of innovation.
The question I’d want people to consider isn’t “why can’t we do it?” or “why haven’t we already done it?”—it’s whether the signal is clear enough. Is it critical? Is the business case really there, and have we spent the time to make sure it is? If it is, we will. Just look at the rate at which we’re building out the equivalent of roads, dams, and bridges for an AI-driven intelligence economy and at the growth in investment happening there. The question is whether it’s critical enough, and whether the signal is clear enough, for American ingenuity to go to work on solving the problem.
Rebecca J. Anderson: The other thing to keep in mind is the American economy in this context—this intersection of the next era of manufacturing taking place in the United States—has shown remarkable resilience over time, and a great capacity for reinvention.
The US economy has reinvented itself multiple times over, for example, during the Great Depression, during World War II, and after the Civil War. There’s great depth in entrepreneurship and in capital markets willing and able to fund it, which I think is very special about the US economy. So if something is a national priority, and there’s intention and momentum behind it, the US economy has some really great wellsprings of innovation to draw on to make it happen.
