Reshoring Rewired the Boardroom Faster Than It Rewired a Single Factory Floor

Every few months a press release lands: a company is bringing thousands of jobs back to America. Add up enough of these and you get a number with a heartbeat - 287,000 jobs in 2023 alone, more than 2.5 million since 2010.¹² It is a satisfying story: globalization went too far, the pandemic and the trade wars taught everyone a lesson, and now the factories are coming home. The trouble is that the people compiling those numbers - the Reshoring Initiative itself - quietly tell you most of those jobs haven't been filled, and may not be for years.¹ The strategy came home. The work, so far, mostly stayed an announcement.

The official story is that supply-chain risk triggered a manufacturing revival. The truer story is narrower and more interesting: supply-chain risk rewired how executives think about geography - the maps in the boardroom were redrawn fast - while the slow, expensive business of pouring concrete and hiring shifts has barely moved the macro needle. Strategy is cheap to change. A fab is not.

The number that counts promises as if they were hires

Start with the headline figure, because it is where the illusion begins. The 287,000 jobs attributed to 2023 is real, but it is two things stacked into one. First, it is not a record - the Reshoring Initiative labels it the second-highest year on record, a detail that vanishes the moment the number enters the press cycle.¹ Second, and far more important, it is a count of announcements. The Initiative's own methodology states that actual hiring lags announcements by twelve to twenty-four months on average, and that any year's total can fold in current hiring, prior years' hiring, and hiring that has not happened yet.¹ A jobs number, in other words, is a basket of past, present, and future tense, all reported in the present. Treat it as a payroll figure and you are reading a forecast as if it were a receipt.

There is a second seam in the number, and it matters more than the first. 'Reshoring' technically means a U.S.-headquartered company moving its own production back home. But the headline figure bundles that together with foreign direct investment - overseas firms building plants on American soil. These are economically different animals. A German automaker opening a plant in the South is not a domestic company repatriating a supply chain; it is a foreign firm deepening its U.S. footprint. The popular coverage blurs the two relentlessly, which is how a story about American factories coming home ends up being, in large part, a story about foreign capital arriving.

Why the boardroom moved and the factory floor didn't

The mechanism is worth working all the way down, because it explains the gap. Strategy is a decision; capacity is an asset. A board can decide to 'de-risk China' in a single afternoon, write it into the annual report, and announce a plant. That decision costs almost nothing and travels at the speed of a slide deck. Building the plant is a multi-year capital project - permits, land, construction, equipment, a workforce that may not exist locally. So when supply-chain risk became the dominant theme - geopolitical risk now ranks as the top driver of reshoring decisions, and European-origin announced jobs jumped from 13% of the total to 34% in 2023 as the Ukraine war made energy security an executive obsession⁸ - the announcements surged first, exactly as you'd expect. The intention layer responds in months. The concrete layer responds in years, if at all.

And the concrete layer is where the story falls apart. Manufacturing construction spending - the leading indicator of factories that will actually exist - has fallen 21% from its June 2024 peak of $239 billion, with electronics and semiconductor fab spending down 44% from its own 2024 high.⁶ Manufacturing employment, far from booming, slipped 1% in the year after the April 2025 'Liberation Day' tariffs that were supposed to ignite the revival.⁶ That is the tell. If reshoring were rewiring the country at the scale of the headlines, the cranes and the payrolls would say so. They are saying the opposite.

Then there is the baseline that almost no celebratory headline supplies. U.S. manufacturing employment peaked near 19.5 million in 1979. By December 2024 it stood at roughly 12.9 million - about 8% of nonfarm payrolls, down from 22% at the peak, with no full recovery from any recession since.⁵ Against that backdrop, even a genuinely strong reshoring year is a few hundred thousand announced jobs against a structural hole of nearly seven million. Calling that a 'jobs revival' is like calling a teaspoon a rainstorm because the bucket is technically getting fuller.

Where the real money actually went - and what that proves

If you want to see what genuinely moved, follow the capital, not the job announcements. Between October 2024 and April 2025, semiconductor projects were just 5% of FDI announcements by count - but they absorbed $102.6 billion, roughly two-thirds of all foreign capital invested, driven by names like TSMC, Samsung, and ASML.⁷ That tells you something the headline number hides: reshoring is not broad. It is concentrated in a handful of strategically critical, heavily subsidized industries where the government put real money on the table. The CHIPS Act anchored exactly that bet - $52.7 billion appropriated, with about $39 billion in direct chip subsidies and a 25% investment tax credit, plus a clause barring recipients from expanding chip capacity in China for ten years.³⁴ Where the subsidy is concentrated and the geopolitics are existential, the concrete really does get poured. Everywhere else, the announcement is mostly the product.

But isn't a slow start still a real beginning?

The fair objection is that this is too cynical. Of course capacity lags strategy - that's physics, not deception. The boards genuinely did reprice supply-chain risk after the pandemic and the trade wars, the CHIPS money is real and the chip fabs are genuinely rising, and the hiring lag the Initiative warns about cuts both ways: jobs announced in 2023 may simply not have hit the payrolls yet by 2026. Give it a decade, the argument goes, and the concrete catches up to the conviction. That's all true, and it's the honest case for the optimists.

But the macro data isn't measuring a future that hasn't arrived; it's measuring a present that's already reversing. Construction spending isn't merely lagging - it peaked and is now falling, 21% off its high, with fab spending down 44%.⁶ A delayed boom shows flat or rising leading indicators waiting to convert. A boom that already crested and turned shows declining ones. And the deeper point survives either way: even on the optimists' timeline, the revival is narrow - chips and a few subsidized sectors - not the broad reindustrialization the headlines imply. Supply-chain risk really did rewire strategy. It just rewired far less of the factory floor than the press releases promised, and the part it did rewire is the part the government paid for.

Reshoring is best understood not as a manufacturing boom but as a giant correction in how risk is priced - and corrections in pricing happen on paper long before they happen in concrete. The boardroom maps were redrawn in months; the factories take a decade and only get built where a subsidy makes the math work. The honest version of the story keeps the insight and drops the inflation: companies stopped trusting their supply chains, and that mistrust is real and durable. What it has produced so far is a wall of announcements, a few enormous chip fabs, and a construction line that is pointing down. The strategy came home. We're still waiting on the work.