For the past two centuries, the grammar of global commerce has been written in the language of comparative advantage. The idea, formalized by the British economist David Ricardo in the early 19th century, is simple: Even if one country is more productive than another at making every type of goods, both can still gain from trade if each specializes in what it produces relatively more efficiently. Comparative advantage is about opportunity cost, not raw prowess. A nation that is best at making semiconductors and shirts should still import shirts if doing so frees scarce capital and talent to make more chips, where its relative edge is greatest. This is why the textbook case insists that no nation, however capable, should try to make everything on its own.
That logic has organized supply chains, shaped corporate strategies, and lifted living standards across continents. Multinationals sliced production into globally distributed tasks; small economies tapped world markets to achieve scale; consumers enjoyed lower prices and greater variety. The post-Cold War period, with its confidence in converging interests and benign interdependence, felt like the high-water mark of comparative advantage as a policy compass.
The tide is going out. In capital city after capital city, “strategic autonomy”, “resilience”, and “national security” have become the operative principles. Governments are reclassifying large swaths of the economy as “strategic industries” — from chips and batteries to cloud infrastructure and the data centers that power artificial intelligence. Where comparative advantage once counselled specialization and import reliance, strategic logic now pushes toward redundancy and domestic capacity — even when others can produce better and cheaper. In some cases, countries are eager to learn entirely new industries from scratch, underwriting massive research and development efforts not because it’s economically efficient, but because it is geopolitically prudent.
The United States has initiated today’s distortion of the global trading order by recentering policy around “national security” and weaponizing choke points. The pivot began when Washington formally labeled China a “strategic competitor”, reframing economic interdependence as a security liability. As American firms fell behind in 5G, the US government sanctioned Huawei, ordered the removal of its equipment from domestic networks, and pressed allies to rip and replace Huawei gear and shun Chinese 5G outright — actions justified on “security” grounds but with sweeping commercial effects.
The campaign then broadened: Export controls and investment restrictions targeted China’s access to advanced semiconductors, chipmaking tools, and critical software, while US industrial policy pushed to onshore key technologies and rebuild domestic capacity. Washington also leaned on key third countries — most notably pressuring the Netherlands to restrict the export of ASML’s EUV lithography systems to China — and expanded extraterritorial controls so that any US-origin technology, wherever embedded, cannot be shipped to China without US approval.
The downstream signal to the rest of the world was unmistakable. If the US could deny a country access to foundational technologies and demand supply-chain conformity, any nation could find itself exposed to sanctions or cutoff. The result has been a global scramble to localize production, less out of textbook economics than out of fear that geopolitical leverage can be turned against them next.
The world’s prosperity and stability depend on an order where technology and commerce flow broadly under clear rules, not one where power politics dictates who may buy, sell, or learn
The new playbook is apparent everywhere. The US is subsidizing domestic semiconductor manufacturing, advanced packaging, and upstream toolmaking, while tightening export controls on high-end chips and chipmaking equipment. Europe is funding its own chip ambitions and erecting digital sovereignty measures around data and cloud infrastructure. Japan is reviving industrial policy for next-generation materials and tools. India is courting electronics assembly and design with incentives and trade protection.
From a Ricardian standpoint, many of these investments look like misallocation. Why build multiple duplicative fabs, subsidize marginal producers, or run smaller, less efficient plants far from global clusters? Because the price of dependence is now judged to be higher than the deadweight losses of dispersion. The “systems risk” of geopolitical fragmentation — sanctions, export bans, blockades, cyber disruption — overwhelms the microeconomic elegance of specialization.
This logic is particularly pronounced in AI, which is both a general-purpose technology and a strategic enabler. As a result, countries are treating AI more like aerospace in the 1960s than like software in the 1990s. The operative analogy is not the smartphone race; it is the space race — or, less comfortably, an arms race. In that framing, the cost of being late exceeds the cost of being wrong.
That is a potent driver of what looks, and many regard, as an AI investment bubble. Venture capital is pouring into model developers, chip startups, synthetic data platforms, cybersecurity, and vertical applications. Hyperscalers are engaged in a capital expenditure arms race for computing — building gigawatt-scale data centers, securing power purchase agreements, contracting for cutting-edge graphics processing unitss and custom accelerators, and laying down new fiber. Sovereign funds are sponsoring national AI champions, while public markets reward anything with an inference engine in the prospectus. If capital markets were a weather map, AI would be a heat dome.
Call it the “cost of lag risk” (CLR) — the expected loss from being behind in a general-purpose, militarily relevant technology. If the CLR exceeds the expected loss from misallocation — building too many data centers, overpaying for chips, funding duplicative research — then a bubble is a feature, not a bug. Historically, the Cold War offers precedent. The space race and associated defense R&D produced moonshots and boondoggles in equal measure, but also yielded spillovers — integrated circuits, satellite communications, materials science, and a generation of trained engineers that seeded new industries. The fiscal efficiency of each line item mattered less than the strategic outcome and the portfolio of externalities.
At present, many governments are moving toward maximal onshoring — an attempt to pull entire supply chains within national borders. This whole-of-economy reshoring is a predictable first response to heightened geopolitical risk, but it is also the most expensive. Recreating dense ecosystems — toolmakers, specialized suppliers, talent pools, logistics — inside a single country imposes steep fixed costs and sacrifices scale efficiencies. As those costs surface in prices, energy demand, and budget realities, the likely next phase is a shift from national autarky to allied architecture — restructuring supply chains across trusted partners to preserve some degree of specialization while reducing exposure to adversarial choke points. In this model, comparative advantage does not disappear; it is bounded by political geography. Countries will concentrate on nodes where they have relative strengths — design, tooling, materials, assembly — within an alliance-based system that shares standards, coordinates export controls, and builds redundancy in critical links.
“Comparative advantage” is giving way to “comparative security”. Ultimately, the US should resist the temptation to weaponize “national security” as a catchall justification for constraining competitors and preserving singular primacy. Using export controls, sanctions, and extraterritorial rules to impede others’ development may deliver short-term leverage but erodes trust, fragments markets, and incentivizes reciprocal controls that leave everyone poorer and less secure. A more durable path lies in recommitting to multilateralism, free trade, and open globalization — principles that China champions. The world’s prosperity and stability depend on an order where technology and commerce flow broadly under clear rules, not one where power politics dictates who may buy, sell, or learn.
The author is a consultant at the Global Hong Kong Institute.
The views do not necessarily reflect those of China Daily.
