Nucor's $3.23 quarter is built on a 50% tariff wall, not a wider moat

---

There is a particular kind of earnings beat that flatters everyone in the room and explains nothing. Nucor's first quarter of 2026 is that kind. The company reported revenue of $9.50 billion, up 21% from $7.83 billion a year earlier; net earnings attributable to stockholders of $743 million, or $3.23 per diluted share, against $156 million a year ago; gross profit up 148% to roughly $1.50 billion; operating income up 247% to about $1.12 billion; and an operating margin that climbed back to 12% from 4%. Management had guided to $2.70–$2.80 a share. The print landed at $3.23, the stock surged, and the analyst notes wrote themselves around the words "record" and "recovery."

Every one of those numbers is real. Nucor is the most capital-disciplined, most diversified, most operationally serious steelmaker in North America, and it is not pretending otherwise. The question this piece asks is narrower and more uncomfortable: when you decompose where the $3.23 came from, how much of it is Nucor doing something only Nucor can do — and how much is a price spike, on a tariff-protected commodity, against a demolished year-ago comparison, that any minimill with the same product mix would have printed too? Because the answer determines whether you are buying a through-cycle compounder or renting a cyclical at the exact moment it is least likely to repeat.

The comp that did the heavy lifting

Start with the denominator, because the denominator is doing most of the persuading. The headline that NUE earnings "jumped 264%" is arithmetically true and analytically hollow. A year-ago quarter of $156 million in net earnings — $0.77 per share — was one of the weakest non-recession prints Nucor has posted in a decade. When the base is on the floor, almost any normalization looks like a moonshot in percentage terms. Grow $156 million to $743 million and you have a 376% increase in net earnings and a +264% EPS move, and neither figure tells you anything about the level of profitability, only about the violence of the bounce off the bottom.

Look at the level instead, across the actual cycle, and the story inverts. In 2023, Nucor earned $18.00 per diluted share on $4.53 billion of net earnings. In 2024, that fell to $8.46 per share on $2.03 billion. In 2025 — the trough — it fell again to $7.52 reported ($7.71 adjusted) on roughly $1.7 billion. The first-quarter-2026 run-rate of $3.23, annualized naively, lands you somewhere in the low teens per share: better than the 2025 trough, nowhere near the 2023 peak, and squarely inside the band a cyclical traverses every few years. The 264% headline is a story about how bad early 2025 was, dressed up as a story about how good Nucor is. The forensic move is simply to refuse the comp and ask what the earnings power actually is. It is the earnings power of a company whose best year was twice this run-rate and whose worst recent year was barely below it. That is the signature of a commodity, not a compounder.

Fourteen percent more per ton — and who set the price

Now the mechanism. Nucor itself is admirably clear about what drove the quarter: the steel mills segment, where average sales price per ton rose 14% and shipments hit a record 7 million tons across all product groups. Price times volume. That is the entire P&L of a steelmaker, and both terms went up at once. The seductive reading is that Nucor earned that price through brand, service, and quality. The forensic reading asks who actually set it.

The hot-rolled coil price Nucor realized in the first quarter is a domestic price floated on top of a 50% Section 232 tariff. In June 2025, the federal government doubled the steel tariff from 25% to 50% for essentially every trading partner except the United Kingdom. The mechanical effect is to lift the landed cost of imported steel by the full tariff and let domestic mills price into the gap underneath it. Nucor's own commentary confirms the channel: import share fell from 22% in early 2025 to about 15% in the first quarter of 2026, and the company expects imports to stay at or below that level so long as the 50% wall holds. A 14% ASP increase is not a moat widening. It is a tariff doing exactly what a tariff is designed to do — transfer margin from foreign producers and domestic buyers to domestic mills — and Nucor is the largest domestic mill.

This is the distinction between a moat and a loophole, and it matters because the two have opposite durability. A moat is something a competitor cannot replicate and a regulator cannot revoke. A tariff is the precise opposite: every domestic competitor — Steel Dynamics, Cleveland-Cliffs, the reorganized U.S. Steel — sits behind the identical wall and prices into the identical gap, and the wall itself exists at the pleasure of an administration that doubled it once and could halve it with the same pen. The 14% per ton is not Nucor's to keep. It is Washington's to lend, and the lease has no stated term.

Record shipments into a wall, not into demand

The volume side deserves the same skepticism the price side gets. Record shipments of 7 million tons sounds like demand. Some of it is — data-center construction, infrastructure spending, and service-center restocking are genuinely real. But a meaningful slice of "record shipments" is the mechanical flip side of collapsing imports. When import share drops from 22% to 15% of a roughly flat domestic market, those tons do not vanish; they get re-sourced from domestic mills. Nucor's record volume is partly demand and partly substitution — the same buyers buying the same steel from a domestic supplier because the imported alternative now carries a 50% surcharge.

That distinction is invisible in the shipment line and decisive for the thesis, because substitution volume is a one-time level shift, not a growth rate. You capture the import share once. After imports fall from 22% to 15%, there is no second 7-point gift unless the tariff rises again. The bull who annualizes record Q1 shipments as a new baseline is double-counting a non-recurring re-sourcing as if it were structural demand growth. Strip the substitution and the underlying end-market signal is far more muted: Nucor's own guidance language is that 2026 demand picks up only "slightly" from 2025, dragged by lagging sectors, with nonresidential construction customers reporting improving second-half outlooks — the language of hope deferred, not demand delivered.

"Harvest phase" — the tell in management's own words

On the first-quarter call, Nucor's leadership framed the moment as the beginning of a "harvest phase," after years of building four new facilities and pouring capital into the West Virginia sheet mill. It is an honest and even admirable framing — and it is also, read coldly, an extraordinary thing for a self-styled secular compounder to say out loud. You harvest a crop once it is ripe; you harvest it because the growing is done and what remains is to collect. "Harvest" is cyclical vocabulary. It is what you say at the top of the field, not the bottom.

The forensic point is not that management is being deceptive — it is being unusually candid. The point is that the bull narrative and management's own metaphor are in quiet tension. The bull buys NUE as a machine that compounds through cycles. Management describes a company that spent years planting and is now gathering — which is to say, a company whose near-term earnings are front-loaded by a capital cycle that is ending and a price cycle that is cresting. When the people running the company reach for harvest imagery in the same quarter they print a 264% EPS jump on a tariff-fueled price spike, the disciplined reader hears a peak being narrated in real time, however unintentionally.

Cyclical priced as secular: the $6.7 billion mirage

Here is where the valuation does its sleight of hand. The bull anchor is Nucor's November 2022 Investor Day "through-cycle" EBITDA target of $6.7 billion, which the West Virginia mill, four ramping facilities, and a projected $500 million EBITDA uplift from completed projects are meant to deliver at full run-rate. Set against full-year 2025 EBITDA of roughly $4.2 billion, that target implies a $2.5 billion step-up, and the bull capitalizes the gap as if it were contracted and certain.

Two problems. First, "through-cycle" is a phrase that quietly assumes a normalized steel price — and the entire reason the first quarter looked spectacular is that the current steel price is anything but normalized; it is tariff-inflated. You cannot use a tariff-spiked spot environment to underwrite confidence in a through-cycle EBITDA figure that, by construction, is supposed to average across good prices and bad. If you believe the $6.7 billion, you are implicitly assuming the 50% wall is permanent — at which point it is not "through-cycle" at all, it is "through-tariff." Second, the target was set in late 2022, in the long shadow of the 2021–2023 super-spike when Nucor was earning $18 a share; "through-cycle" framed against that backdrop is structurally optimistic. The market is being asked to pay a mid-cycle multiple on a tariff-peak earnings stream toward a normalized-EBITDA target conceived at the prior cyclical high. That is three layers of optimism stacked into one number, and each layer is a place the thesis can break.

The capital is going out the door regardless

While earnings ride the tariff, the cash commitments are fixed and large. Nucor spent $661 million on capital expenditures in the first quarter alone — roughly 40% of it on the West Virginia Apple Grove sheet mill — against full-year capex guidance of about $2.5 billion, with roughly $950 million earmarked for that single mill in 2026. The mill is targeted for completion late in 2026. None of that spending flexes with the steel price. It is committed concrete and committed steel-fabrication equipment, and it lands whether hot-rolled coil holds at tariff-supported levels or reverts.

This is the cyclical's oldest trap, and Nucor is disciplined enough to usually avoid it: peak cash flow funds capacity that arrives into a softer price. The new West Virginia tons come online precisely as the company itself guides to only slight demand improvement and a second-half-weighted nonresidential recovery. If the tariff is trimmed, or if buyers who pulled forward orders ahead of price increases stop reordering, the incremental tons meet a thinner price than the one that justified them. Nucor returned $254 million to shareholders in the quarter — a respectable but not aggressive payout against $743 million of net earnings, because the capital program has first claim on the cash. The buyback authorization is large; the actual repurchase pace is governed by how much the West Virginia mill and the four ramping plants demand first.

Adjusted, GAAP, and the quality of the headline number

Quality-of-earnings discipline requires one more pass. Nucor's first-quarter print is relatively clean — $743 million GAAP, $3.23 GAAP diluted, with no heroic adjustments papering over the headline. That is to Nucor's credit and worth saying plainly. But the cleanliness of the GAAP number is exactly what makes the composition of it so important. There is no accounting trick here; the entire question is economic, not accounting. The earnings are real cash earnings — they are simply earnings produced by a price and a volume that are both sitting at policy-supported, comp-flattered highs.

Note also the cadence beneath the annual figures: 2025 was not a smooth trough but a deteriorating one, with the fourth quarter of 2025 missing at $1.73 against a $1.86 expectation on $7.69 billion of revenue, below the $7.87 billion expected. The first quarter of 2026 then snapped sharply higher. A clean GAAP number that swings from a Q4 miss to a Q4-trouncing Q1 beat in a single quarter is, itself, the tell: this is an earnings stream with enormous quarter-to-quarter operating leverage to a price the company does not control. High operating leverage cuts both ways. The same sensitivity that turned a 14% ASP gain into a 247% jump in operating income will turn a 14% ASP give-back into a brutal contraction. The quality of the earnings is fine. The stability of them is the exposure.

The denominator illusion in the multiple

There is a second denominator trap, this one in the valuation rather than the earnings. When a cyclical prints a strong quarter, the trailing and forward P/E both compress mechanically — earnings rise, the multiple "looks cheap." A steelmaker at a single-digit-to-low-teens P/E on peak earnings is the textbook value trap: the multiple is low because the market correctly suspects the E is not sustainable. The disciplined approach normalizes earnings across the cycle before applying a multiple. Do that for Nucor — average across the $18.00 of 2023, the $8.46 of 2024, the $7.52 of 2025, and a tariff-aided 2026 — and the "cheap" multiple on spot earnings becomes a far less obviously cheap multiple on mid-cycle earnings. The cheapness is an artifact of measuring the price against a peak-flattered E. That is the denominator illusion: the same trick the 264% headline plays on the growth rate, the trailing multiple plays on the valuation.

What the bulls genuinely get right

Now concede, specifically and fairly, what is genuinely strong — because the bull case on Nucor is not a fantasy, and treating it as one would be its own form of dishonesty.

First, Nucor is, by a wide margin, the best-run steelmaker in North America, and that is not a slogan. Its electric-arc-furnace minimill model is structurally lower-cost and lower-carbon than blast-furnace integrated production; its profit-sharing culture and variable cost structure let it stay cash-generative deep into troughs where integrated peers bleed. The 2025 "trough" still produced $4.2 billion of EBITDA and $1.7 billion of net earnings — a level of through-cycle profitability that genuinely distinguishes Nucor from a generic commodity producer. Many cyclicals lose money at the bottom. Nucor compounds book value through it.

Second, the balance sheet is a fortress. The company ended the first quarter with $2.48 billion in cash and short-term investments and an undrawn $2.25 billion revolving credit facility that does not expire until 2030. That liquidity means the West Virginia mill and the four ramping facilities will be finished regardless of the price environment — and if the cycle turns down, Nucor will buy back stock and take share while weaker competitors retrench. Financial strength is itself a competitive weapon for a cyclical, and few in the sector have more of it.

Third, the diversification into steel products and downstream fabrication genuinely dampens the pure-commodity swing. The steel products segment showed improved earnings on stable realized pricing in the first quarter — a steadier ballast against the violent mill-segment swings. Nucor is more than a coil price; it is towers, joists, racking, buildings, and data-center infrastructure, and that mix is real and growing.

Fourth — and this is the strongest bull point — the secular demand story is not invented. Data-center construction, grid build-out, infrastructure spending, and reshoring of manufacturing are real, large, multi-year sources of domestic steel demand, and Nucor is positioned squarely in front of them. If those demand vectors hold and the tariff wall stays up, the $6.7 billion EBITDA target is achievable rather than fanciful, and today's price will look like a reasonable entry into a structurally higher earnings plateau. That is a coherent, defensible bull case. The bear case is not that it is impossible. The bear case is that the market is paying for it as though it is already certain, while the single most important variable — the tariff — is a political decision, not an operating result.

The variable nobody underwrites

That is the crux, and it deserves its own frame. Every model of Nucor's forward earnings — bull, base, or bear — silently embeds an assumption about the Section 232 tariff, and almost none of them stress it. The tariff doubled to 50% in mid-2025. It could be trimmed in a trade negotiation, carved out for an ally the way the United Kingdom already was at 25%, challenged, or simply allowed to lapse in priority under a future administration. The moment the wall comes down, import share reverts from 15% toward 22%, the domestic price gap compresses, the 14% ASP premium deflates, and Nucor's enormous operating leverage runs in reverse. None of that requires a recession. It requires only a change in trade policy — the one input to Nucor's earnings that no amount of operational excellence can hedge, and the one the bull case treats as a constant.

A genuine through-cycle compounder is robust to its policy environment. Nucor, on the evidence of this quarter, is unusually dependent on it. That is not a knock on the company. It is a description of where the earnings come from — and a warning that a stock priced for a secular re-rating is exposed to a variable that lives in Washington, not in Charlotte.

The kicker

None of this says Nucor is a bad company. It says the opposite: Nucor is so well run that its operational excellence becomes the perfect camouflage for a quarter whose real engine was a 50% tariff and a crushed comp. The $3.23 is genuine cash, but it is cyclical cash dressed as secular cash, and the gap between those two readings is the entire investment debate. The bull pays a mid-cycle multiple on a tariff-peak number toward a normalized target conceived at the last super-cycle high — three optimisms stacked into one price. The bear simply asks the company to keep doing this after the wall comes down, after the import-substitution gift is fully banked, after the harvest is in. Management told you which season it is; the market is still pricing the spring.

The steel is real, the cash is real, and the moat is mostly a tariff — and tariffs, unlike furnaces, are repealed by signature, not by competition.