The United States government handed the semiconductor industry $52 billion in subsidies through the CHIPS and Science Act — and American factories are still getting outcompeted by Taiwan and South Korea on contract after contract.

Let that sink in for a second.

Fifty-two billion dollars. That is more than the entire GDP of Honduras. More than NASA’s budget for the next four years combined. And yet, when major tech companies need cutting-edge chips manufactured at scale, the phone calls are still going to TSMC in Hsinchu and Samsung in Suwon — not to Phoenix or Columbus.

The real story behind the headlines is not about a funding shortfall. It is about something far more uncomfortable — and far more structural — that nobody in Washington wants to say out loud.

The Number Everyone Is Ignoring

Here is what most people do not know: according to a 2024 report by the Boston Consulting Group and the Semiconductor Industry Association, manufacturing a chip in the United States costs approximately 30% to 50% more than manufacturing the same chip in Taiwan or South Korea — even after the subsidies are factored in.

Think of it this way: imagine the government gave your local bakery $10,000 to compete with an industrial bread factory overseas. But that factory overseas still runs its ovens at half the cost, has workers trained specifically for that one task since childhood, and has supplier relationships built over 40 years. Your bakery is still losing. The check helped. It did not fix the problem.

That cost gap is the story. The subsidies are a headline. The structural disadvantage is the reality.

How We Got Here: A 40-Year Slide

The United States was not always playing catch-up. In the early 1990s, American companies produced roughly 37% of the world’s semiconductors, according to the Semiconductor Industry Association’s 2023 State of the Industry report. By 2023, that number had collapsed to 12%.

This did not happen by accident. It happened by deliberate choice.

Through the 1980s and 1990s, Wall Street analysts pushed American manufacturers to focus on chip design — the high-margin, asset-light end of the business — and outsource the messy, capital-intensive fabrication to Asia. Companies like Intel held on longer than most, but the pressure was relentless. Meanwhile, Taiwan and South Korea made different bets. Their governments treated semiconductor manufacturing as a matter of national survival, not a market opportunity. Taiwan’s government helped launch TSMC in 1987 with direct investment and policy protections. South Korea made Samsung’s semiconductor ambitions a literal state priority.

Ask yourself why they do not advertise this part: the very Wall Street logic that gutted American chip manufacturing is the same logic now cheering the CHIPS Act subsidies while those companies collect the checks and still route their most advanced orders to Asia.

Did You Know: TSMC’s Arizona fab, which received CHIPS Act funding, will manufacture chips that are less advanced than what TSMC produces at its flagship plants in Taiwan — a gap TSMC itself acknowledged in a 2023 earnings call.

The Workforce Problem Nobody Wants to Talk About

Throwing money at buildings does not build expertise overnight. Taiwan has over 300,000 skilled semiconductor workers with decades of specialized, hands-on experience. South Korea has built entire university ecosystems — Samsung alone partners with more than 20 Korean universities on semiconductor engineering curricula.

Here is what this actually means for you: when TSMC opened its Arizona facility, it flew in hundreds of Taiwanese engineers to train local staff, because American workers simply did not yet have the specific process knowledge to run the fab independently. This is not an insult to American workers. It is the predictable result of 40 years of offshoring that gutted the vocational and technical training pipeline for semiconductor fabrication in this country.

Intel’s Ohio facility — the largest semiconductor project in American history — has faced repeated construction and hiring delays, partly because the specialized contractor and engineering workforce in central Ohio does not exist at the scale required. Intel disclosed timeline setbacks in 2024 that pushed initial production estimates back by at least one to two years.

Convenient, right? Billions out the door, timelines slipping, and the chips your phone and car and medical device need are still being stamped out of Tainan and Icheon.

Warning: Several analysts, including those at research firm Gartner, have flagged that without a sustained 10-to-15-year commitment to workforce development and supplier ecosystem building, CHIPS Act investments risk producing underutilized fabs that cannot compete on yield or cost — essentially expensive monuments to good intentions.

Multiple Sides of a Real Argument

To be fair — and I always try to be — there are serious people who argue the CHIPS Act is working exactly as designed. Economic Policy Institute senior economist Josh Bivens argued in 2023 that rebuilding domestic capacity is necessarily a long-term project and that expecting five-year results from a 40-year decline is unrealistic. That is a reasonable point.

Others, including analysts at the Cato Institute, take the opposite position: that semiconductor subsidies distort markets, reward politically connected corporations, and create dependency on government favor rather than genuine competitive advantage. Also a reasonable point — just from the other direction.

And then there is the national security framing, which is arguably the strongest argument for the CHIPS Act on its merits. Taiwan sits 100 miles from mainland China. If access to TSMC were disrupted — through conflict, blockade, or political pressure — the United States economy would face a shock that makes the COVID supply chain crisis look minor. The Department of Commerce estimated in 2021 that a six-month shutdown of leading-edge chip supply would cost the U.S. economy $240 billion.

That number is why the subsidies exist. Whether they are being executed effectively is a separate and urgent question.

Pro Tip: If you want to track whether CHIPS Act investments are actually producing results, watch the metric that matters most: not groundbreakings and press conferences, but yield rates — the percentage of functional chips produced per wafer — at domestic fabs versus Taiwan and Korea. When that gap closes, the policy is working. Until then, it is potential, not performance.

Who Actually Benefits Right Now?

I dug into the actual research so you do not have to — here is what I found.

The companies receiving the largest CHIPS Act grants include Intel ($8.5 billion), TSMC’s American subsidiary ($6.6 billion), Samsung’s Texas operation ($6.4 billion), and Micron ($6.1 billion), according to Commerce Department announcements through early 2025. Two of the top four recipients are foreign companies operating American subsidiaries.

And who benefits from you not knowing this? Every stakeholder who wants you to feel like the problem is solved — so the political pressure to actually fix the structural issues fades away.

Action Step: Contact your congressional representative and ask them specifically: what are the measurable yield and employment benchmarks attached to CHIPS Act grants in your state? If they cannot answer, they are not doing oversight. This money has your name on it.

Your Next 3 Steps

1. Follow the fab timelines, not the press releases. Set a Google Alert for “Intel Ohio fab” and “TSMC Arizona production.” When those facilities hit full-scale production of advanced nodes — not just ceremonial ribbon-cuttings — that is the real milestone. Reuters and the Wall Street Journal both track these timelines with specificity.

2. Pressure your pension fund or 401(k) provider to ask hard questions. If you hold U.S. semiconductor ETFs like SOXX or SMH, a significant portion of those holdings still depends on Asian fabrication. Ask your fund manager how they assess geopolitical supply chain risk. The question alone signals to them that their clients are paying attention.

3. Support technical education funding at the state level. The workforce gap is the longest and hardest problem to solve — and it is mostly solved at the state and community college level, not in Washington. Look up whether your state has a semiconductor workforce initiative. If it does not, ask your state legislator why.

The $52 billion headline was always meant to feel like an answer. The real question — whether America can rebuild something it chose to abandon over four decades — is still very much open. And you deserve to know that.

— Nicole Rivera, WolfTrend