Governments around the world are throwing extraordinary sums of money at semiconductor manufacturing, and the consequences — intended and otherwise — are beginning to ripple through the global economy in ways that few policymakers fully anticipated. What started as a post-pandemic panic over chip shortages has metastasized into something far more ambitious and far more dangerous: an industrial policy arms race with no clear finish line.
The numbers are staggering. The United States alone has committed over $52 billion through the CHIPS and Science Act. The European Union has its own €43 billion European Chips Act. Japan has pledged billions to lure TSMC to Kumamoto. South Korea announced a $470 billion semiconductor investment package. And China, which has been at this longer than anyone, continues to pour tens of billions into its domestic chip industry through state-backed funds and provincial subsidies.
Every major economy now has the same idea: onshore chip production, reduce dependence on Taiwan, and secure the industrial base for artificial intelligence. The problem is that everyone can’t win this race simultaneously.
The Subsidy Spiral and Its Discontents
As the Financial Times has reported, the scale of government intervention in the semiconductor sector has reached levels not seen since the space race. Public money is flowing to chipmakers at a pace that distorts normal market signals, creating a subsidy spiral where each country’s incentive package must be larger than the last to remain competitive for factory siting decisions.
Intel has been a primary beneficiary. The company secured $8.5 billion in direct grants from the U.S. Commerce Department under the CHIPS Act, plus billions more in loans and tax credits. It’s building or expanding fabs in Arizona, Ohio, and New Mexico. But Intel’s own financial struggles — declining revenue in its core businesses, a stock price that has cratered — raise uncomfortable questions about whether public money is propping up a company the market has judged harshly.
TSMC, meanwhile, is constructing a massive facility in Phoenix, Arizona, with substantial CHIPS Act support. Samsung is building in Taylor, Texas. These are real factories producing real chips. But the economics only work because of government subsidies. Manufacturing chips in the United States costs roughly 30-50% more than in Taiwan or South Korea, according to multiple industry analyses. The subsidies close that gap — for now.
What happens when the money runs out?
That’s the question nobody in Washington, Brussels, or Tokyo wants to answer directly. Semiconductor fabs cost $15-20 billion to build and require continuous reinvestment. A single round of subsidies doesn’t create a self-sustaining industry. It creates a dependency. And the political appetite for repeated multi-billion-dollar appropriations is far from guaranteed, especially as deficit concerns mount in the U.S. and fiscal pressures grow across Europe.
The Biden administration structured its CHIPS Act disbursements with guardrails — restrictions on stock buybacks, requirements for childcare facilities at fab sites, limits on expanding advanced manufacturing in China for a decade. The Trump administration, now back in power, has sent mixed signals about whether it will honor the spirit of those commitments or redirect funds toward different priorities. Commerce Secretary Howard Lutnick has indicated the CHIPS program will continue but with a more transactional approach, focused on extracting maximum concessions from recipients.
Industry executives are watching closely. “The worst thing for investment decisions is uncertainty,” one senior semiconductor executive told analysts during a recent earnings call. Fabs take years to build. Committing billions requires confidence that the policy environment won’t shift midstream.
China’s Shadow Over Everything
The entire subsidy competition is, at its core, about China. Every government justifies its spending by pointing to Beijing’s relentless push to build an indigenous semiconductor capability. And China’s progress, while uneven, has been real.
SMIC, China’s most advanced foundry, has managed to produce chips at the 7-nanometer node — a feat that surprised Western intelligence agencies, given the export controls designed to prevent exactly that outcome. Huawei’s Mate 60 Pro smartphone, which appeared in late 2023 with a SMIC-fabricated chip, was a wake-up call. The controls hadn’t failed entirely, but they hadn’t succeeded entirely either.
The U.S. has since tightened restrictions further, limiting exports of advanced chipmaking equipment from ASML, Tokyo Electron, and Applied Materials. The Dutch and Japanese governments, after sustained American pressure, agreed to align their export controls with Washington’s. But enforcement remains a persistent challenge. Equipment and components find their way to China through intermediaries. And China is investing heavily in developing its own equipment supply chain, even if the technology lags by years.
So the subsidy race has a dual logic. Build capacity at home. Deny it to China. Both objectives are expensive, and they sometimes conflict. Chipmakers that accept CHIPS Act money face restrictions on their China operations — a meaningful sacrifice for companies like Intel and Samsung that have significant existing investments there.
The geopolitical stakes extend beyond economics. Taiwan Semiconductor Manufacturing Company produces roughly 90% of the world’s most advanced chips. A Chinese blockade or invasion of Taiwan would create an immediate and catastrophic disruption to global technology supply chains — everything from iPhones to F-35 fighter jets. Diversifying production away from that single point of failure is a legitimate national security imperative. But the cost of that diversification is enormous, and it’s being borne largely by taxpayers rather than shareholders.
There’s a tension here that the industry prefers not to discuss openly. Chipmakers benefit from the subsidies. They also benefit from the threat that justifies the subsidies. Every escalation in U.S.-China tensions strengthens the case for more government money. This doesn’t mean companies are manufacturing the threat — the risks around Taiwan are genuine and well-documented — but the incentive structure is worth noting.
Meanwhile, the AI boom has added another accelerant. Demand for advanced chips from Nvidia, AMD, and increasingly from hyperscalers designing their own silicon has created a gold-rush mentality. Governments now view semiconductor capacity not just as a defense concern but as the foundation of economic competitiveness in the AI era. If your country doesn’t have fabs, the argument goes, you’ll be dependent on others for the most important technology of the next decade.
This argument has some merit. And some holes. Most countries subsidizing chip manufacturing aren’t building the kind of leading-edge capacity that produces AI accelerators. They’re building trailing-edge or mid-range fabs that produce automotive chips, industrial sensors, and power management components. Important, yes. But not the strategic chokepoint that dominates the policy rhetoric.
The distinction matters because it reveals a gap between the political narrative and the industrial reality. Politicians talk about AI and national security. The factories being built often serve more mundane markets. That’s not necessarily bad — the automotive chip shortage of 2021-2022 demonstrated the real economic damage caused by insufficient capacity at mature nodes. But it does suggest the subsidy programs are solving a different problem than the one being advertised.
Europe offers the clearest illustration. The EU Chips Act aims to double Europe’s share of global semiconductor production to 20% by 2030. That target was always ambitious; most analysts consider it unrealistic. Europe’s strengths lie in chip design (ARM, ASML’s lithography equipment, NXP’s automotive chips) rather than high-volume manufacturing. Building fabs in Europe means contending with higher energy costs, stricter environmental regulations, and a thinner talent pool for semiconductor manufacturing than exists in East Asia.
Intel’s planned fab in Magdeburg, Germany — a cornerstone of the EU’s strategy — has faced repeated delays and uncertainty about its future, particularly as Intel restructures its business. The German government committed nearly €10 billion in subsidies for the project, making it one of the most heavily subsidized fab investments in history on a per-facility basis.
Japan has had more success, at least in execution. TSMC’s Kumamoto fab is operational, and a second facility is planned. The Japanese government has been pragmatic, focusing on attracting established manufacturers rather than trying to build national champions from scratch. But Japan’s approach also highlights the limits of subsidy-driven strategy: the country is essentially paying a Taiwanese company to do what it already does best, just on Japanese soil.
What Comes Next
The semiconductor subsidy race isn’t slowing down. If anything, it’s intensifying. India has entered the competition with its own incentive program, attracting commitments from Micron and others. Saudi Arabia and the UAE are exploring semiconductor investments as part of their economic diversification strategies. Vietnam, Malaysia, and Singapore continue to compete aggressively for packaging and testing operations.
But the industry itself is sending cautionary signals. Semiconductor manufacturing is cyclical. The current AI-driven boom in demand for advanced chips coexists with oversupply in memory and certain logic segments. Building massive new capacity with public money during a boom risks creating a glut when the cycle turns. And cycles always turn.
There’s also the question of whether subsidies are actually changing the competitive dynamics or merely transferring wealth from taxpayers to some of the most profitable companies on Earth. TSMC’s gross margins exceed 50%. Nvidia’s are even higher. These aren’t struggling industries in need of life support. They’re enormously profitable enterprises that have learned to extract public money by wrapping their investment decisions in national security language.
That’s a cynical reading, and it’s not entirely fair. The market failures in semiconductor manufacturing are real — the massive capital requirements, the long investment horizons, the strategic vulnerability of concentrated production. Government intervention can be justified on those grounds. But the scale of current subsidies, the lack of coordination between allied governments, and the absence of clear metrics for success suggest that the policy response has outrun the policy thinking.
The most likely outcome is a world with more semiconductor manufacturing capacity, spread across more countries, at higher cost. Chips won’t get cheaper because of these investments. They’ll get more expensive, or at minimum, the cost reductions that would have come from continued concentration in East Asia won’t materialize. Consumers and businesses will bear that cost, invisibly, embedded in the price of every electronic device.
Some of that cost is worth paying for genuine supply chain resilience. Some of it is pure political theater — ribbon cuttings and press releases that make for good campaign material but questionable industrial strategy.
Sorting one from the other is the hard part. And right now, nobody’s doing it particularly well.
WebProNews is an iEntry Publication