The U.S.-China rivalry is often narrated through the language of tariffs, sanctions, and export controls—discrete policy tools deployed by states against one another. Missing from this picture is a deeper structural question: what happens when the corporations these policies target have interests, dependencies, and bargaining power of their own? What happens when the global economy itself becomes the terrain on which geopolitical competition unfolds, rather than simply its backdrop?
Steve Rolf brings that structural perspective into focus. A Principal Research Fellow at the University of Sussex Business School, Dr. Rolf is a political economist whose research examines the digitalization of economies and its consequences for workers, states, and regulators. He recently published an article, “Spatial Limits to Weaponizing Interdependence: TSMC in the U.S.–China Chip War,” in a theme issue of the journal Economic Geography titled ‘Geopolitical Rivalry and the New Global Economic Geography’, which he also guest-edited. The article examines why the world’s leading chipmaker has retained strategic agency despite mounting pressure from both Washington and Beijing.
Rolf’s broader work, developed with colleagues at the Second Cold War Observatory, argues that U.S.-China rivalry is increasingly becoming a structuring force in the global economy—reshaping trade, technology, corporate strategy, and even the context in which other international conflicts unfold. Rather than treating markets and geopolitics as separate spheres, his research shows how deeply intertwined they have become, offering a new framework for understanding globalization’s next chapter.
Vimi Wang: You argue we are not just watching a trade dispute, but the reshaping of the entire global economy. For someone who has never thought about “economic geography,” what is one thing about their daily life—a phone, a car, or even a grocery bill—that is already being reshaped by the U.S.-China rivalry?
Steve Rolf: Cars and phones are both great examples, highlighting different aspects of current transformations. The kind of car you are able to buy nowadays is very contingent on where in the world you sit. The United States currently has a rule coming from the Bureau of Industry and Security (BIS) that seeks to ban connected vehicles—particularly electric vehicles, but even conventional vehicles that connect to the internet for autonomous-driving purposes—if they use Chinese or Russian software. That ban comes into force next year, alongside related hardware regulations and tariffs on products like Chinese electric vehicles. So the kind of car you are able to buy depends on where you are in the world, not just because of market differentiation, but because states are trying to carve out markets for their own security initiatives.
The same is true of mobile phones. We have seen a real divergence of technology ecosystems between China and the United States over the past decade through a gradual change in legislation. The US Huawei ban was important, and ZTE too, going back nearly a decade to the first Trump administration. But China has also responded to tensions by limiting engagement with certain American technologies (such as banning Google mobile services) and supporting domestic “national champions”—most notably Huawei, which has developed its own mobile operating system, HarmonyOS, and has since taken it off Android compatibility entirely.
These dynamics extend overseas as well, where the US and China are engaged in a rivalry to secure technological dominance in third country markets. However, they cannot easily reproduce the same degree of separation abroad as at home. In the African continent, for example, there’s sharp competition between the U.S. and China. Chinese-subsidiary firms like Transsion now hold nearly 50% of the African smartphone market, with its own app layer. These phones also offer payments services running on Huawei’s mobile money platform rather than U.S.-connected financial infrastructure. At the same time, many such phones continue to operate using Android (Google), which shows how messy and interconnected superpower competition is within third countries. As you can see, regulatory walls are emerging between the two superpowers, reshaping daily life at home while also extending into third-country markets and influencing how competition unfolds abroad. Although it clearly depends on where you are, geopolitics is beginning to structure daily life in meaningful, material ways for large numbers of people.
VW: Huawei is a fascinating example of how U.S.-China competition can shape the decisions of other countries. Building on that idea, you and your co-authors argue that the “Second Cold War” is not simply a U.S.-China rivalry. Rather, it has become an organizing framework for conflicts that predate it, including Russia’s war in Ukraine. Can you unpack that? How does one rivalry come to reshape so many others?
SR: What we want to avoid is viewing every global incident simply through the lens of U.S.-China rivalry, even though that tendency is present in some foreign policy discussions. To take a step back, many geopolitical conflicts are unfolding simultaneously today, from Ukraine to Gaza, Lebanon, and Iran. The argument is not that these conflicts are simply extensions of U.S.-China rivalry. Rather, the rivalry is becoming an organizing force that shapes the broader environment in which these localized conflicts and competitions take place. That distinction is crucial.
Take the Russia-Ukraine conflict. This remains a conflict between the parties directly involved, but it has also drawn in Europe—given its deepening relationship with Ukraine over recent decades—and increasingly the United States. Ukraine was one of the core foreign policy areas for the Biden administration, and despite the rhetoric, I think it has also remained important to the second Trump administration’s foreign policy strategy. On one hand, the U.S. has interests in some of these lower-tier conflicts that aren’t directly about superpower rivalry. On the other hand, China has become a critical economic partner for Russia, providing a major market for Russian energy as Russia faces growing challenges accessing other markets.
Arguably, those dynamics wouldn’t have unfolded the same way without this pre-existing U.S.-China rivalry. The fact that China is now engaged in a prolonged strategic competition with the United States makes it more likely to preserve and even deepen its relationship with Russia—not as a formal treaty ally, but as an important informal partner within this broader geopolitical competition. For the United States, there has also been growing concern about China’s role in this conflict, a concern that has surfaced at several key moments.
You can see the same dynamic in other conflicts—for instance, Chinese-flagged vessels being able to traverse the Strait of Hormuz, while vessels connected to the U.S., Israel, or the West more broadly can’t. The point is that this broader superpower rivalry doesn’t determine everything, and we should not reduce every micro- or medium-level conflict to a simple extension of that competition. Any given conflict probably has some relationship with this broader U.S.-China rivalry—though that relationship can be complicated. The Chinese have interests on both sides of the current Middle East conflict: connections with the Gulf states, with Iran, and with Israel as well. This is a reminder that there are no clear territorial blocs here; states do not fall neatly into one camp or another.
That is one of the key arguments my colleagues and I at the Second Cold War Observatory have been making. We have moved from a First Cold War world, where states aligned with one of the two superpowers—the Soviet Union or the United States—to a world where globalization, networked integration, and deep trade, financial, and investment flows make that separation much harder. States no longer neatly belong to one camp or another. Instead, they are embedded in distinct networks, and competition often unfolds within them. Geopolitical positioning today is therefore much more complex than it was during the First Cold War.
VW: Many headlines claim globalization is ending as countries talk about decoupling and reshoring. Yet your work suggests something much more complicated is happening. Are we witnessing the end of globalization, or simply a new version of it?
SR: We have argued the latter. Seth Schindler and I coined the term “geostrategic globalization” in an article last year, capturing what we see as the key transformation of the global political economy over the past five or ten years. If you define globalization in strict terms—as the extension of economic activity across national borders through international trade, investment, financial flows, and global value chains (where goods cross borders many times before reaching consumers)—we don’t think that is going away. You just need to look at the data: measures of international economic activity remain at all-time highs. While they have stopped growing since around 2008, they have plateaued at a very high level. The world’s economies remain deeply interconnected, and we do not see that changing.
So does that mean nothing has changed—that all these geopolitics is just surface-level froth, and the deep structures of interdependence rule out real conflict? Unfortunately, as time goes on, it is becoming increasingly clear that deep global economic integration is not a barrier to conflict. We see conflicts escalating in many places today, from actual military conflicts to the trade, investment, and financial conflicts we will discuss shortly.
Those are difficult facts to square, and that is exactly what “geostrategic globalization” is meant to capture. The degree of global economic integration is likely to remain high, but states are also playing a much more active role in shaping those international economic connections—channeling them towards where they want and pulling them back from where they do not. That is what I mentioned earlier: the creation of distinct regional markets for smartphones and cars, but the same is true for biotechnology, AI, and other strategic sectors. We are seeing real bifurcations in the world economy, driven by states creating barriers to certain forms of integration while building new connections elsewhere. Friend-shoring, reshoring, and similar strategies are becoming increasingly widespread.
There was an interesting article in the Financial Times recently reporting on a study by EY-Parthenon estimating it would cost nearly $24 trillion for the U.S. and Europe to fully decouple from China over the next 25 years. That’s an extraordinary degree of investment—realistically, it can’t happen. We do not have the money to fix roads, let alone invest at that scale on a problem that doesn’t necessarily exist. It does not fundamentally hurt to import clothes or basic consumer electronics from China. Instead, we see a much more patchwork effort by states to identify what is and isn’t strategic—which areas are sensitive enough that they can no longer be integrated purely through markets and require greater state oversight, versus areas where states can remain more relaxed. That is what we mean by geostrategic globalization: preserving deep international economic integration while states play a much more active role in managing and redirecting capital flows.
VW: There is a striking argument in your work: the global economy many people viewed as “neutral” or “rules-based” was never actually separate from politics—it was a political project all along. If that is true, what does that mean for how we should think about free markets today?
SR: There is a misconception when we make the argument that globalization is increasingly geopolitical. The pushback we often get is that neoliberal globalization was also a form of geopolitics—and I completely agree, it was. We make that argument both in the piece on geostrategic globalization and in a more recent one written with Tim Zajontz that came out in the same journal, Globalizations, about a month ago, called “Structural Power Redux: Geoeconomic Rivalry in a Networked World.”
In both pieces, we look at how the U.S. constructed its power during the period of mainly market-based neoliberal globalization, from the late 1970s onward. Before that, in the aftermath of the Second World War, we had a much more state-managed and territorially bounded political economy, where states had managerial capacity over their own economies, and capital flows, trade, and investment were much lower – particularly in the immediate postwar period. Utilities and key industries were broadly nationally owned or managed, and industrial policy was widespread. This was the era of big states and Keynesian macroeconomic management.
But from the 1970s onward, that order was transformed by the neoliberal revolution—pushed first and foremost by the United States, both through its own bilateral relations with other countries and through changes within international organizations like the IMF and World Bank. The policy programs they promoted after the debt crises of the 1980s and ’90s helped create a much more market-based international economy built around ostensibly neutral rules, opening economies and privatizing national utilities and industries across much of the world.
The big winners of this process were U.S. multinational corporations. They were increasingly able to operate across the globe—bidding for construction contracts in former Soviet states, acquiring oil interests across the Middle East, and establishing manufacturing plants in Mexico, Korea, Taiwan, and, from the 1990s onward, China. While this was formally a neutral, market-based system, the de facto beneficiaries of marketization and globalization were U.S. multinationals and, ultimately, the U.S. state itself.
This state of affairs persisted until it stopped working for the United States—above all, once China emerged as a serious contender in a number of areas. First, through its major infrastructure investments in the 2010s, particularly the Belt and Road Initiative, a vast program of investment across the Global South and Europe that intensified U.S. concerns about China. Then, increasingly, through initiatives like Made in China 2025 and the industrial strategy introduced in 2015, which aimed to move China closer to the technological frontier across strategic industries. The results have been remarkable, from industrial robotics to semiconductor technology. The short story is that China became, if not fully at the technological frontier, much closer to it than the United States was comfortable with.
The U.S. sees this order it constructed—based on free markets where anyone can compete, but with the implicit assumption that it would outcompete everyone because it had the largest, most powerful, and most competitive corporations—suddenly becoming a potential threat as Chinese firms emerge as serious rivals. This prompts a shift away from neoliberal globalization toward something much more overtly geopolitical, or geostrategic, where investments and international economic linkages are increasingly scrutinized through a political lens. ‘Does this system still serve us?’ ‘Is it in our national interest to keep our borders open to the world’s cars?’ Increasingly, the answer has been no—we do not want Chinese EVs entering the U.S. market because they are cheaper and potentially more competitive than what domestic automakers can produce. Add in broader security concerns, and the landscape looks very different from what it once was.
Taking one step back: markets cannot exist without states—they are completely intertwined and codependent institutions. The key patron of the global free-market system, the United States, has decided that free markets no longer work in its favor and is now rewriting the rules of that system to alter its own position. Whether that will succeed remains an open question. In the short term, it may bring certain benefits—you protect domestic automakers from Chinese competition, for example. But in the long run, whether it works as a broader strategic move is far less clear. Whatever your judgment, the fact remains: there has been a concerted shift away from the Washington consensus that free markets serve U.S. interests.
VW: The world’s most important chipmaker hires PhDs in international relations while navigating pressure from both Washington and Beijing. Given America’s central role in semiconductor technology, why hasn’t the U.S. been able to fully weaponize TSMC?
SR: We (Joseph Baines, Julian Germann, and myself) wanted to look at TSMC because of an interesting news story. The U.S. put in place a range of controls on semiconductor technologies and how they could be imported into China—this happened in a few iterations, but became particularly intense from October 2022, when the Biden administration instantiated a range of technology controls unilaterally, not coordinated with allies.
TSMC has two fabrication facilities in mainland China, one in Nanjing and one in Shanghai. In 2022, the former became subject to U.S. licensing requirements to import the machinery needed to keep it running. We noticed the U.S. granted these licenses—first on a 12-month basis, then indefinitely. That seemed strange: why impose unilateral technology restrictions, only to immediately grant exceptions to the main firms importing the targeted equipment? We started with that puzzle.
TSMC is technically a prime candidate for the U.S. to weaponize against China. It is the world’s largest semiconductor manufacturer, the only company able to produce advanced nodes at scale, and likely to remain at the technological frontier for years. Although it is headquartered in Taiwan rather than the U.S., TSMC is deeply dependent on American firms: on the supply side, through equipment makers and design-tool companies like Applied Materials and Cadence, and on the demand side, through major U.S. customers like Apple, Nvidia, and Meta. The U.S. has shown it can use this leverage—after Huawei was added to the entity list in 2019, TSMC cut off supply. The question is not whether the U.S. has leverage over the world’s key semiconductor manufacturer, but why it has not used that leverage more.
Quite the opposite: TSMC considerably expanded its Nanjing fab from 2021-4, investing over $2 billion. It is doubling down in China, not pulling back. We explain this through “territorial embeddedness”—a concept geographers use to describe economic activity that depends on place-based relationships. TSMC relies heavily on the world-class, deeply integrated electronics and machine tools manufacturing ecosystem which has formed across the Taiwan straits. Mainland Chinese suppliers play a hugely outsized role in supplying key equipment and manufacturing inputs to TSMC’s Taiwan and mainland China fabs, with dense personnel flows between leading Chinese firms and TSMC and proximity to consumers (firms which physically assemble chips into devices, such as Foxconn and Luxshare, rather than their nominal US customers) are also key factors. Building and operating a semiconductor fab is extremely difficult without access to mainland China’s dense industrial ecosystem.
On top of that, the United States and its major tech firms are now tied together in the AI race with China. For tech firms, it is a key economic goal; for the U.S. government, it is a geopolitical one. But both share the aim of staying ahead of Chinese competitors. To do that, U.S. firms’ access to TSMC’s manufacturing capabilities is an absolute precondition. There is no realistic scenario in which that can be sacrificed. TSMC’s capacity is already under enormous demand, with orders reportedly booked years in advance. Anything that destabilized TSMC’s operations would threaten U.S. tech companies and broader geopolitical goals alike—especially given that many of TSMC’s critical inputs, including machinery and components, come from mainland China.
You get a complex picture. As one 2022 report put it, the U.S.—particularly under Biden—wanted to “strangle” China’s chip and AI sector, freezing its development in place, and TSMC appeared to be a key lever. But the U.S. has been reluctant to use it because American tech firms are deeply dependent on TSMC to manufacture the chips and AI accelerators needed to compete with China. The result is a relationship of mutual dependency: the U.S. needs TSMC, while TSMC relies on its manufacturing ecosystem and suppliers across the Taiwan Strait. That gives TSMC bargaining power, allowing it to deepen its presence in China rather than pull back—and helps explain why the U.S. has been cautious about using TSMC as a weapon. Doing so could ultimately harm its own strategic and economic interests.
I think this gives a window into how geopolitics actually operates today. For the most part—certain regional conflicts aside—we’re not talking about shooting wars. Though those still matter, they may be less common and less decisive. Instead, it is about how states try to mobilize corporate operations toward their favored goals, and then run into unexpected barriers because of these complex global entanglements—an iterative process where states try to bring corporations in line with their objectives, while corporations retain some agency over whether they comply. This also explains why—as you alluded to in your question—firms are increasingly being compelled to develop expertise in international relations and geopolitics. This simply wasn’t necessary in the recent past, but is now needed to give firms the best chance of reducing mounting risks and potentially even capitalizing on interstate rivalries in new ways.
VW: Historically, economic restrictions have sometimes pushed countries to become more self-reliant. Do you think today’s export controls could ultimately accelerate China’s technological development rather than slow it? And if so, what would that mean for the future of U.S. strategy?
SR: It’s complex. In AI—the “AI race,” so to speak—it is clear that China is constrained by compute. I have been trying to use Z.ai’s new model, GLM 5.2, and Moonshot’s Kimi K3, over the past few weeks, and it has been difficult to access both because the servers are consistently overloaded. On one hand, that shows there is strong demand for Chinese tools; on the other hand, it highlights the limited availability of inference compute. Both in training and inference, U.S. AI chip restrictions on China have clearly had an impact, showing how economic coercion can function as a meaningful geopolitical strategy. But they have not been enough to severely damage the sector. We are talking about models that may be around six months behind those of leading U.S. labs—a meaningful gap in a fast-moving industry, but not a devastating one. And what is more, China’s catchup has actually coincided with the period of maximum attempted U.S. coercion.
That gets to the other side of your question: efforts to weaponize firms and technologies against other states can clearly backfire. This is especially clear in China’s chip sector. The Chinese state is not naive—it has long recognized potential U.S. leverage in digital technology and hardware, and it has pursued semiconductor industrial policies dating back to the late 1990s and early 2000s. But those efforts were never hugely successful because chips from firms like TSMC remained abundant, high-quality, and affordable. Persuading firms to move away from that supply when alternatives are available is genuinely difficult—it is a collective-action problem. Any individual firm may recognize its vulnerability to U.S. pressure, but what can it do alone? Local governments, which channel much of China’s industrial policy and subsidies, often face the same calculation: why not encourage firms to buy from TSMC, given its decades of specialization and world-class engineering talent, and target other, more comfortable niches?
I think the chip war of the past eight or nine years has resolved that collective-action problem for Chinese firms and the Chinese state. It is now clear that access to Nvidia AI accelerators and other U.S. technologies cannot be taken for granted, so it makes sense for Chinese firms to invest in alternatives. That shift has made a huge difference. Huawei is a clear example. U.S. measures against the company, especially its addition to the entity list, were initially extremely damaging, hurting its smartphone business, access to operating systems, and other key areas. But Huawei has since staged a remarkable resurgence: its revenues have recovered significantly, and it is now producing, with Chinese partners, semiconductor and AI chips that are not far from the technological frontier while partnering successfully with AI model developers.
Arguably, none of this would have happened without U.S. measures. Huawei might have remained a profitable company selling 5G equipment and smartphones in global markets. But stripped of those markets and access to U.S. chips and software, it was forced to build its own technology—and appears to have succeeded. If anything, Huawei is now a more economically and technologically secure firm than it was five or six years ago. The U.S. has essentially exhausted its leverage: it placed Huawei on the entity list, cut off access to U.S. technology, and restricted its market access. Yet the company survived and appears to be thriving.
There’s no single answer as to whether these controls help or hurt—you have to assess them field by field, firm by firm. But at least in chips, and likely in AI as well, the evidence so far suggests they have backfired on the U.S. and the architects of these policies. I don’t think they have stopped Chinese development or innovation.
As for whether this will produce truly indigenized firms, I wouldn’t go that far. China remains a deeply globalized economy, and its firms are genuinely internationalized. The real question is where the boundary gets drawn: what stays outside the firm and the state, and what gets brought in-house. China will continue to depend on external sources for raw materials, battery minerals, rare earths, technologies, and global flows of people and goods. Globalization has changed and become much more shaped by state intervention, but the genie cannot be put back in the bottle. China is unlikely to become a fully self-reliant, insular economy. This means we should expect U.S. attempts to stall its rise to chase those potential friction points around the globe.
VW: One assumption many people have is that governments ultimately hold final authority. But your research suggests companies like TSMC can sometimes shape the choices of even the world’s most powerful states. Has globalization created corporations that are becoming geopolitical actors in their own right?
SR: This is a fantastic question, and the answer is basically yes. States are increasingly involved in geopolitical competition, either as protagonists or because they are responding to U.S.-China rivalry and its effects on their economies. But to act geopolitically, states need corporations. They rarely operate as fully self-contained geopolitical actors anymore; much of today’s competition happens through firms and economic networks.
This is visible in the cases of TSMC and Huawei: corporations are often both targets of geopolitics and the actors expected to carry out geopolitical objectives. Sometimes there is little room for bargaining power—when the U.S. sanctions a firm like Huawei, companies like TSMC have little choice but to comply.
But geopolitical competition also operates through softer tools. Conditional subsidies are one example: the CHIPS and Science Act tied access to U.S. subsidies to reducing business with Chinese firms and building fabs in the United States. Controls on non-rivalrous goods, like cloud compute, are another. These softer instruments leave firms more room to decide whether to align with state objectives or hedge against them.
This has become quite acute for many companies, putting them directly in the spotlight. In the geostrategic globalization article, we look at two examples: Apple and Intel—two archetypes of U.S. big tech, but with very different responses. The U.S. has pushed Apple, through a mix of hard and mostly informal pressure, to wind down its China supply chain for close to a decade. That creates a dilemma for Apple: if China were formally sanctioned, the choice would be made for it, but as things stand, Apple still has room to decide how this unfolds. Its response has been instructive—it has begun diversifying a supply chain that was once extremely concentrated in China.
Patrick McGee’s book Apple in China documents this well: Apple has encouraged suppliers like Foxconn to move operations to India, Vietnam, and elsewhere in East, South, and Southeast Asia. But often those sites handle only the final assembly stage, while much of the manufacturing still happens in China beforehand. So this is less a departure from China than a form of hedging against geopolitical risk. Fundamentally, Apple does not want to leave China, and it has continued lobbying against sanctions and tariffs. Last year, tariffs on key iPhone components were rolled back, reportedly after direct negotiations between Tim Cook and Donald Trump led to promises of major investments within the U.S. So Apple’s strategy has been to build some redundancy while preserving business as usual as much as possible.
The opposite approach is exemplified by Intel. In recent decades, Intel’s vertically-integrated business model (encompassing chip design, manufacture, sale, and branding) has struggled, particularly with the rise of pure-play foundry models like TSMC. Its response has been to embrace the competition with China that the U.S. government has driven. It pulled back sharply on expansion plans in China—abandoning plans to acquire a major GlobalFoundries facility in Chengdu around 2021–22—and instead embraced the CHIPS and Science Act subsidies, along with projects such as building secure semiconductors for the Pentagon with supply chains designed to minimize Chinese inputs. Intel has aligned itself closely with U.S. geostrategic goals, and that alignment is likely to deepen, as it offers Intel a strategic opportunity to become a favored player in the U.S. ‘s increasingly securitized chip ecosystem. This culminated with the US government taking a 10% equity stake in the firm last summer. In this way, firms like Intel can be seen as accelerants of geopolitical conflict, insofar as their willingness to embrace state objectives broadens the range of geostrategic possibilities for the latter.
Those are two typical, divergent responses corporations can take. The point is that they have choices to make—when power is exerted in a hard, direct fashion, less so, but in day-to-day operations, corporations have become actors through which geopolitics happens. Geopolitical success or failure is therefore often contingent on decisions made in corporate boardrooms, which differs from older images of geopolitical conflict playing out on a battlefield. At the same time, foreign corporations are increasingly now targets of geopolitical initiatives. The Huawei case was a targeted effort to take down a firm seen as a uniquely important threat to the U.S., yet Huawei’s agility, strategy, and state support allowed it to survive. As the U.S.–China rivalry deepens, the question is no longer whether corporations are participants in geopolitics, but how much they will shape its outcomes.