Broad, strings-free subsidization of U.S. semiconductor manufacturing is costly, unnecessary and perhaps even harmful to the industry itself.

First, the U.S. semiconductor industry is healthy and expanding. While the U.S. share of global chip production has fallen since the 1990s, the industry’s research-and-development spending, capital expenditures and inflation-adjusted output—in terms of value and wafer capacity—have increased substantially over the same period. American semiconductor firms also still produce 44% of their wafer supply domestically and lead the world in chip design and innovation.

The most advanced chips are today imported from East Asia, but U.S. powerhouse Intel and other semiconductor manufacturers—enjoying astronomical profits due to intense global demand—are planning future investments in production of advanced and legacy chips in the U.S., with or without subsidies. They have gone on a spending spree in the U.S. and other locations outside of China or Taiwan.

TSMC’s Arizona facility will open in 2024, and it will be the most advanced in the country. Samsung is expanding in Texas, as is Intel in Arizona, Oregon and Ohio. Around $80 billion of private investment in American chip manufacturing is forecast through 2024, and experts agree it will happen regardless of federal support, because chip makers covet the U.S. workforce and proximity to specialized equipment manufacturers. These are simply not companies that need taxpayer help or the government’s encouragement to produce far more in the U.S.

Second, China-related foreign-policy concerns are overblown. Multinational sourcing decisions always consider geopolitical risks, and this is certainly the case for semiconductors. Many large chip-consuming companies are already adjusting their supply chains to account for geopolitical tensions. Semiconductor manufacturers are doing the same: Expansion plans of Samsung, Intel, TSMC, GlobalFoundries and others are all motivated, at least in part, by geopolitics. If giant multinationals deem Taiwan or even Asia to be too risky and want chip production elsewhere, they can pay for it—and that’s exactly what they’re doing.

Besides, the federal subsidies under consideration are in no way sufficient to replace supplies from Taiwan if a Chinese invasion removed those chips from the global market, nor would they act quickly to have any effect. Any new subsidized U.S. production would at best be online in 2025 and more likely in 2026–27. Fortunately, the worst-case scenario of a Chinese invasion is a remote risk.

Subsidies, on the other hand, present far likelier risks—as China itself shows. Its industry remains years behind top chip makers and riddled with problems despite the Chinese government spending decades and billions of dollars to achieve national semiconductor greatness. U.S. subsidies, rather than producing an efficient, competitive domestic industry, could instead make it bloated, dependent on federal assistance and globally uncompetitive.

They could also contribute to a global semiconductor glut and create new and costly trade conflicts.

Chip making is notoriously cyclical, with a history of strong capital spending followed by overcapacity, price crashes and struggling firms. Current investment is, by some accounts, already at levels that typically lead to oversupply, and previous demand forecasts may have been too optimistic. Many analysts are thus worried about a global glut in 2023 that would put U.S. and foreign chip companies in financial distress. U.S. subsidies would exacerbate these unstable market dynamics.

Subsidies would also foment trade disputes as nations move to protect struggling domestic chip makers from subsidized import competition—precisely what happened decades ago when U.S. tariffs targeted Japanese and Korean memory chips, harming American computer companies and consumers in the process.

Other common justifications for subsidies also fall flat. They won’t alleviate the current chip shortage, which should end long before subsidized U.S. production arrives. Furthermore, the proposed funds aren’t limited to supporting the most advanced semiconductors or national-security-related ones, undermining arguments that these subsidies are needed because free markets can’t fully address risky R&D or national-security concerns. Indeed, the House and Senate bills earmark billions for older commercial chips because Detroit auto makers use them.

Subsidies for today’s leading technologies might not solve tomorrow’s problems, either, as attention and resources are already shifting from even the most advanced current technologies to different models like quantum and neuromorphic computing.

Ultimately, the economic and national-security justifications for throwing taxpayer billions at domestic chip makers are weak. Self-interested semiconductor firms may claim that their situation is dire and that only subsidies can save them and the country. But Congress shouldn’t play along.