The Onshoring Story Is Pointing Investors at the Wrong Companies

If you've been told the tariff-driven reshoring boom is creating the next wave of industrial stock winners, you're probably looking at the wrong companies.

The narrative sounds right on the surface. Tariffs go up on Chinese, Canadian, and Mexican imports. Companies announce they're moving production home. ISM manufacturing PMI hits 54.6 — expansion territory. The White House calls it the "largest reshoring wave in American history". It should be a straightforward play: buy domestic manufacturers.

But manufacturing construction spending tells a different story. It peaked at $239 billion in June/July 2024 and has since declined 21%. Factory building didn't accelerate after tariffs. It was already slowing down.

That doesn't mean there's no industrial boom happening. It means the boom is somewhere else entirely.

The real construction story

While traditional factory construction has been falling, data center spending grew from about $9.5 billion annualized in early 2020 to $47 billion by early 2026, with year-over-year growth above 30%. And data centers don't just need servers. They need the entire physical infrastructure stack: substations, transformers, switchgear, busways, power distribution units, and liquid cooling systems.

S&P Global forecasts $1.3 trillion in aggregate U.S. energy utility CapEx between 2026 and 2030; Morningstar DBRS projects $1.4 trillion in electricity infrastructure investment from 2025 to 2030 — double the prior decade.

This is the infrastructure buildout. It's not about bringing garment factories back to the Midwest. It's about building the power delivery layer for an economy that is fundamentally more electric, more digital, and more capital-intensive than the one that existed five years ago. And the companies sitting at the junction of the grid and the chip don't need reshoring to justify their orders. They just need demand.

Eaton and the toll road on data center power

Eaton (NYSE: ETN) makes the equipment that sits between the utility grid and every major piece of commercial and industrial equipment. Electrical panels, switchgear, circuit breakers, transformers, UPS systems, power distribution for data centers. It is, in the clearest possible sense, a toll road business: if you're building a data center, you need this equipment, and there are only a handful of suppliers who deliver at scale and reliability.

The numbers from the last two quarters of 2026 show the toll road is busy. Second-quarter revenue hit $8.5 billion, up 21%, with organic growth of 14%. Twelve-month rolling average orders in Electrical Americas accelerated 41%. Total backlog in the Electrical sector grew 48% year over year. The company's second-quarter segment margin hit 24.9%, at the high end of its guidance range.

That margin figure is the pricing power test. Eaton isn't just selling more — it's selling at better economics. In a competitive market with weak demand, volume growth comes at the expense of margin. Here, both are expanding simultaneously. That's the pattern you look for when asking whether a business can grow dividends through any cycle.

On the dividend side, Eaton has raised its payout for 12 consecutive years. The current quarterly dividend is $1.10 per share, yielding about 1% at the current price. The payout ratio sits near 41%, leaving substantial room for the dividend to grow alongside earnings. Free cash flow over the trailing twelve months was $3.9 billion, up roughly 20% year over year. The dividend yield isn't attractive on its own — but the equity yield curve approach says you don't buy dividend growers for their current yield. You buy them so the compounding works over time. A 1% yield growing at 12% a year becomes a 60% yield on cost after three decades. The question is whether the business can sustain that growth rate.

The valuation question

Here's where the story gets harder. Eaton trades at roughly 42 times trailing earnings, 27.6 times EV/EBITDA, and 5.3 times sales. Net debt stands near $20 billion against a $160 billion market cap. These aren't cheap multiples. They reflect a market that has already priced in the data center tailwind and is asking whether the growth can sustain this valuation.

For comparison, nVent Electric (NVT) — another electrical infrastructure company benefiting from the same buildout — trades at 42 times earnings with a slightly lower EV/EBITDA of 25. Caterpillar (CAT) sits at 34.5 times earnings. Vertiv (VRT), which supplies data center cooling and power management, is even richer at 62 times earnings. Eaton sits in the middle of this pack, which tells you the market is willing to pay up for this category of business.

The valuation isn't a reason to avoid Eaton. It's a reason to check whether the growth cycle is real, durable, and priced in already. The backlog growth of 48% and order acceleration of 41% suggest the growth is real and not yet priced out — backlog is a forward-looking measure that has to play out over quarters, not a backtested consensus. But the multiple also means the stock is vulnerable if the data center buildout slows, if utilities delay capital programs, or if competition erodes margins.

What would change the case

The Eaton story depends on three things continuing: data center demand, pricing power, and utility willingness to spend. Each has a visible failure mode.

Data center demand is the most uncertain variable. The buildout is driven by hyperscaler capital expenditure — U.S. hyperscalers collectively spent over $410 billion on CapEx in 2025, up from $200 billion in 2024. If that spending decelerates, the order flow slows. But this isn't a discretionary consumer trend. It's driven by the structural demand for compute capacity, which has been growing faster than either cloud providers or chip manufacturers anticipated. The risk isn't that AI demand disappears — it's that it matures and the growth rate normalizes.

Pricing power has held through 24.9% margins, but it's a function of current capacity constraints. If competitors expand capacity faster than demand, the margin advantage fades. Eaton's scale and brand position in critical power infrastructure make this less likely than in a commoditized market, but it's not immune.

Utility capital spending is the most durable of the three drivers. The $1.4 trillion grid investment cycle is driven by aging infrastructure, reliability requirements, and regulatory mandates that aren't cyclical. This is the floor under Eaton's growth, even if the data center ceiling proves lower than the current run rate suggests.

What the onshoring noise obscures

The reshoring narrative isn't harmless — it just points investors at the wrong businesses. The fastest-growing manufacturing sectors — computers/electronics and aerospace — face the lowest tariffs, while heavily protected industries remain stagnant or declining. Tariffs didn't create a manufacturing boom. Tariffs compressed gross margins as input prices rose more than output prices, according to Federal Reserve regional surveys.

The real structural shift isn't about where goods are made. It's about what kind of physical infrastructure the economy requires. Power capacity. Grid reliability. Electrical distribution for facilities that didn't exist a decade ago. The companies making that infrastructure don't need tariffs to create demand. They need the economy to keep becoming more electric and more digital.

Eaton is the clearest public company playing that role with a proven dividend track record and a balance sheet that can sustain the investment cycle. The valuation is rich, the margin expansion is real, and the backlog growth suggests the growth story isn't finished. If you're looking for income growth — not current yield — in a real-economy business with pricing power, this is the kind of setup the equity yield curve approach identifies. The price you pay for it is accepting that the market already believes the story and is asking you to be right about what comes next.