Steel Dynamics, Inc. (NASDAQ: STLD) — The Better-Run Twin, Carrying an Aluminum Lottery Ticket at a Richest-Ever Price

Independent fundamental research. Report date: 2026-06-26.

⚡ Claude’s Take

This block is the author’s own subjective opinion and general information only — not investment advice. The analysis that follows (sections 1–15) takes no position and carries no price target; it discusses valuation only as embedded expectations and scenarios.

Call: HOLD / accumulate-on-weakness (toward ~$190–220) / not-a-short. Medium conviction. The best-run house on a tariff-rented street, with a half-built second house thrown in — priced at the top of its own history.

Steel Dynamics is, by the evidence, the best-managed company in North American steel — arguably a notch ahead of even Nucor on the two things that actually compound shareholder value: capital allocation and incentive design. The buybacks are genuinely counter-cyclical (heavy at ~$80–130 in 2022–23, throttled to a ~$115M trickle at ~$251 in Q1-26 — the opposite of the buy-high reflex that defines most of this richest-ever cohort), the share count is down ~31.5% in five years, the dividend has risen 14 straight years, and — rare anywhere, almost unheard-of in a commodity — the long-term incentive plan scores management on after-tax ROIC, operating margin, cash conversion and revenue growth, measured relative to CLF/CMC/NUE/Metallus, with CEO target pay below the 25th percentile of peers. Founder Mark Millett still runs it. Management walked away from the BlueScope JV in February rather than overpay. This is as good as steel management gets.

The catch is the same one that defines its twin Nucor, plus one of its own. First, the same one: steel is a fungible, globally over-supplied, deeply cyclical commodity, and STLD’s entire current earnings strength rests on a reversible policy artifact — the 50% Section 232 tariff and the April-2026 derivative-loophole closure that pushed HRC from ~$850 to over $1,000 and let STLD run 89% utilization against a 77% industry. Strip the tariff and US prices revert toward the China-set world level. Its through-cycle ROIC still swings from ~43% (2021) to ~10% at the 2025 trough — barely its cost of capital — which tells you that even the best EAF operator is a price-taker on the spread. Second, its own: the ~$2.7B Aluminum Dynamics mill is a genuine new earnings leg Nucor doesn’t have (target ~$650–700M of through-cycle EBITDA, a ~30% uplift to the base) — but it is losing money today (−$173M in FY25, −$65M in Q1-26 on a startup pause and a quality-stain write-down) and its $650–700M is management’s unproven promise on a first-of-kind automotive/can qualification ramp. It is real optionality and real execution risk in the same package.

At $251 — ~16x trailing EV/EBITDA, ~27x trailing EPS, and the richest valuation in its entire history on book (99th percentile) and sales (99th) — you are paying a peak multiple on tariff-inflated spreads, for a company whose best leg (steel) is at the top of its policy-driven cycle and whose growth leg (aluminum) hasn’t earned a dollar yet. This is not a short: the capital allocation, the fortress balance sheet, the recovering 2026 earnings (Q2 guided to $3.51–3.55), the aluminum option, and a violent momentum tape (+98% in twelve months) make betting against it dangerous. But the margin of safety is gone. I’d happily own this quality cheaper — accumulating on the inevitable cyclical pullback toward the low-$200s and ideally ~$190–220, where a sane through-cycle multiple on normalized (steel mid-cycle + a probability-weighted aluminum) earnings reappears — and otherwise collect the dividend and wait. The stock has already slipped ~11% from its June ATH; that’s a start, not a sale. Framing: a best-in-class quality cyclical with a free-ish call option, on a momentum-and-policy tailwind, at a full price. Conviction medium. Flips bullish if Section 232 is codified into durable law and Aluminum Dynamics hits its $650–700M run-rate on schedule (turning the lottery ticket into a banked third of EBITDA). Flips bearish on a tariff rollback/adverse ruling, the HRC spread mean-reverting as ~9–12M tons of new US sheet capacity lands, or the aluminum ramp slipping materially past exit-2026. Tag: the better twin, the rented street, the unscratched ticket.

📈 Stock Price Action — Five-Year Event Map

Over five years STLD has run roughly 5.7x off its early-2022 low — from a ~$49 trough (January 2022) to a fresh all-time high of $282.76 (June 12, 2026), before pulling back to $251.00 (June 25, 2026). The 52-week range is ~$120–$283; the stock sits ~11.2% off its high. The arc is a textbook commodity cycle bookended by two policy-and-demand surges, with a distinctive STLD overlay: a ~$5B self-funded capacity build (Sinton, coating lines, and the new aluminum mill) compounding through the cycle.

Cycle narrative. (1) The 2021–22 leg was a genuine commodity super-spike — COVID supply disruption plus fiscal stimulus drove US HRC to records and STLD to back-to-back record years (EPS $15.56 then $20.92), even as it spent ~$2B standing up the Sinton, Texas flat-roll mill. (2) The 2022 wobble was a mild cyclical de-rate as the Fed hiked and HRC slipped from the peak — notably shallower than the integrated peers, the first sign of STLD’s through-cycle resilience. (3) A long 2022–24 recovery rode soft-landing and reshoring optimism while management bought back stock hand-over-fist at ~3–5x EV/EBITDA. (4) Through 2024–25 the stock drifted ~12% lower as the cycle turned, EPS troughed at $7.99, and the aluminum build consumed cash with no offsetting profit. (5) The ~135% surge off the August-2025 low is the re-rating that defines today’s setup: the 50% Section 232 tariff (June 2025) and the April-2026 derivative-loophole closure choked imports, lifted HRC past $1,000, drove record Q1-26 steel shipments (3.6M tons at 89% utilization), and — layered with the aluminum-ramp story — carried STLD to a fresh ATH and its richest-ever valuation. (6) The recent ~11% slip from the $282.76 ATH followed a Q2 guide ($3.51–3.55) that landed below the spot-HRC-implied bull hopes and a round of sell-side price-target trims. Every price move here is a Fact; the attributed driver is Interpretation, cross-referenced to earnings prints, 8-K events, trade-policy actions, and the news feed.

1. Executive Summary

Steel Dynamics is the third-largest steel producer in the United States (~10% of domestic output), the largest metals recycler in North America (OmniSource), one of the largest steel fabricators (New Millennium joist & deck), and — as of 2025 — a brand-new entrant in flat-rolled aluminum (Aluminum Dynamics). It is an electric-arc-furnace (EAF) “mini-mill” operator that melts recycled scrap rather than smelting iron ore, giving it a structurally lower and far more variable cost base than the legacy integrated mills. Founded in 1993 by Mark Millett (still Chairman & CEO) and two partners, it has grown into a ~$18B-revenue, four-segment metals platform run with an unusually disciplined, return-anchored, incentive-driven culture.

The investment tension is the same one that defines its closest peer Nucor, with a twist: STLD is an exceptionally well-run company in a structurally unattractive industry, trading at the most expensive valuation in its own history, on the back of a tariff-driven and possibly transient up-cycle — and it is simultaneously betting ~$2.7B on an aluminum business that has yet to earn a profit. The quality is real and, on capital allocation, arguably best-in-class: profitable through every trough, 14 consecutive years of dividend increases, a ~31.5% share-count reduction in five years executed counter-cyclically (heavy when cheap, throttled when dear), the rare commodity producer that ties executive pay to after-tax ROIC and other return metrics measured relative to peers, and a fortress investment-grade balance sheet. But the economics are unavoidably cyclical: diluted EPS has swung from $2.59 (2020) to $20.92 (2022) to $7.99 (2025); through-cycle ROIC from ~43% to ~10% — roughly its cost of capital at the bottom. There is no customer captivity and no pricing power independent of the commodity cycle. The current strength rests on Section 232 tariffs (raised to 50% in June 2025, extended to full derivative value in April 2026) that lifted the domestic HRC price past $1,000/ton and cut import share — a policy premium shared by all US producers, not a proprietary moat.

Earnings are genuinely recovering off the 2025 trough: Q1-2026 delivered record 3.6M-ton steel shipments, 89% utilization (vs. 77% for the industry), net income of $403M ($2.78/sh), and adjusted EBITDA of $700M — despite a −$65M aluminum startup loss. Management guided Q2 to $3.51–3.55/sh. That is the bull’s fuel, alongside the aluminum option: a fully-ramped Aluminum Dynamics targets $650–700M of through-cycle EBITDA (a ~30% uplift to the FY25 EBITDA base), into structurally better end-markets (countercyclical beverage cans, low-carbon automotive). The bear’s fuel is threefold: the tariff is reversible; the industry is over-building (~9–12M tons of new US sheet capacity landing 2026–2029 into flat demand — the textbook Marathon late-cycle signal); and the aluminum ramp is unproven, already showing first-quarter quality stumbles. At $251 (P/B in the 99th percentile of its decade, P/S 99th, ~27x trailing EPS, ~16x trailing EV/EBITDA), the market is capitalizing today’s tariff-inflated steel earnings and crediting a meaningful share of the un-earned aluminum profit. The body that follows takes no position; it lays out why the quality is real, why the industry is not, and what the price already assumes.

2. Business Overview

What STLD does. Steel Dynamics manufactures and sells steel and steel products, processes and trades the scrap that feeds its own and others’ furnaces, fabricates downstream building components, and — newly — produces flat-rolled aluminum. It is among the most vertically integrated and product-diversified metals companies in North America, spanning metallics (scrap) → steel/aluminum → fabricated products. Production is overwhelmingly via the electric-arc-furnace route: recycled scrap (supplemented by pig iron and DRI-substitute liquid iron) is melted in electric furnaces, cast, and rolled into finished shapes — a lower-fixed-cost, more-variable, lower-carbon process than the integrated blast-furnace/BOF route used by Cleveland-Cliffs and the former U.S. Steel.

Four reporting segments (FY2025):

• Steel Operations (~70% of revenue; the core). Hot-rolled, cold-rolled and coated sheet (Butler IN, Columbus MS, and the Sinton TX mill), plus a wide long-products franchise — structural beams (Columbia City IN), railroad rail, merchant bar, reinforcing bar, and special-bar-quality (SBQ) engineered steels (Pittsboro IN, Roanoke VA). Sold to service centers, OEMs, fabricators, and the auto/energy/construction supply chains. The most commodity-exposed, most cyclical segment, but itself diversified across sheet and long products. STLD is the largest steel coater in North America, a value-added position that benefited directly from 2025 trade-case wins.
• Metals Recycling (OmniSource). The largest ferrous and nonferrous scrap processor in North America, moving ~6.2M gross tons of ferrous in FY2025, of which ~65% (~4.0M tons) ships to STLD’s own mills. Ferrous (shredded, busheling, heavy melt) and nonferrous (aluminum, copper, brass, stainless). A US and Mexican footprint that feeds the Columbus/Sinton steel mills and, increasingly, the aluminum mill. This is the vertical-integration engine that internalizes the industry’s single biggest input risk — scrap supply and cost.
• Steel Fabrication (New Millennium). Steel joists, girders, trusses, and deck for non-residential construction — warehouses, data centers, metal buildings, commercial and institutional projects. Higher-margin (~29% segment operating margin in recent periods) and — critically — a captive internal customer for the steel mills that lets STLD run its furnaces hot through soft patches. The countercyclical buffer that distinguishes STLD’s earnings profile.
• Aluminum Operations (Aluminum Dynamics). A ~650,000-metric-ton flat-rolled aluminum mill in Columbus, Mississippi, fed by a recycled-aluminum “cathouse” (slab caster) in Columbus and a satellite in San Luis Potosí, Mexico. Targeting a 45% beverage-can / 35% automotive / 20% industrial product mix by 2027. Began production in 2025 and is mid-ramp — currently a loss-maker, targeted at $650–700M of through-cycle EBITDA when fully ramped.

How it makes money. Like all EAF steelmakers, STLD earns the metal spread — the gap between the selling price of steel and the cost of scrap/metallics plus conversion. Profitability is driven far more by the spread (a function of steel prices, scrap costs, and utilization) than by tonnage alone. Because labor is incentive- and profit-share-based and the cost structure is highly variable, STLD stays cash-generative even at low utilization — it has never posted an annual loss the way integrated peers do. Revenue is transactional (spot plus short contracts), not subscription-like; ~75–80% of flat-rolled sheet is on contracts with a ~2-month price lag, the rest spot.

Scale and concentration. ~$18B revenue; ~13–14M tons of steel shipped annually; ~10% of US output; ~12,000 employees; HQ in Fort Wayne, Indiana. No meaningful customer concentration — the business is a diversified spread of service centers, OEMs and fabricators. Geographically US-centric with a growing Mexican recycling/aluminum footprint.

Verdict. A genuinely integrated, broadly diversified, scale-leading metals franchise with a real captive-demand buffer (fabrication) and a credible new growth leg (aluminum) — but at its steel core a producer of a fungible commodity whose revenue is transactional and whose profit is a spread set by the cycle. The diversification dampens cyclicality more than the integrated peers achieve; it does not change the fundamental commodity character of the earnings.

3. Industry Dynamics

Structure. US apparent steel consumption is ~95–100M tons/yr of finished product, led by construction and infrastructure (~50%+ of consumption), then automotive (~14%), energy/pipe, machinery, appliances and containers. Non-residential construction is the swing demand variable — currently firm in data centers, infrastructure and multifamily, softer elsewhere. The industry is brutally cyclical, capital-intensive, and produces a fungible commodity with no customer captivity and no switching costs. Steel is priced off global benchmarks (HRC, plate, rebar) net of freight and tariffs; producers are price-takers on the commodity.

The EAF transition. EAF mini-mills are now ~70% of US production, with the integrated blast-furnace/BOF route the shrinking ~30%. Nucor (~18% share) and Steel Dynamics (~10%) are the low-cost EAF leaders; Cleveland-Cliffs (CLF) and the former U.S. Steel (now Nippon-owned) are the higher-cost integrated incumbents. The EAF cost advantage — scrap feedstock vs. iron ore + coking coal, variable labor, flexibility, far lower capex intensity and carbon — is the structural reason STLD and Nucor out-earn the integrated mills across the cycle.

The Marathon capital cycle — a late-cycle over-build warning. This is the single most important industry signal, and it is flashing. Record metal spreads plus record sell-side optimism are coinciding with a wave of new US flat-rolled capacity being committed into the high prices — exactly the supply-side over-investment that capital-cycle analysis warns mean-reverts returns. Conservatively ~9–12M tons of new US sheet capacity is landing 2026–2029 (Nucor’s West Virginia mill, U.S. Steel/Nippon’s “Big River 2,” Hyundai/POSCO’s Louisiana mill, plus the continued ramp of STLD’s own Sinton mill) into a market management itself guides to “flat to +2%.” Industry utilization sits at ~77% — below the ~80% needed for healthy margins, before the new tons arrive. The mitigant the bulls cite: net US capacity has been roughly flat because old blast furnaces are retiring as new EAF tons come on. The net-vs-gross capacity question is the most important open variable for the forward spread. STLD’s relative positioning here is a positive: it is past its own capex peak (FY24 $1.87B → FY25 $948M → 2026 guide ~$600M) while peers spend, so it absorbs the wave from a position of completed investment and falling capital intensity.

China / global overhang. China’s crude steel output has fallen but Chinese exports remain near record levels (~130M tons) as weak domestic demand pushes tonnage abroad; global excess capacity is ~640M tons and rising. This permanent overhang caps world steel prices and is precisely why US prices hold up only behind the tariff wall.

Section 232 — the policy prop. The Section 232 steel tariff was doubled to 50% (June 2025) and extended to the full customs value of derivative products (April 2026), closing the undervaluation loophole. Import share fell and the domestic HRC price climbed past $1,000/ton in 2026. This premium is a reversible policy artifact — its durability is the central fragility of the entire bull case, and it benefits all US producers, including the high-cost integrated mills it keeps alive (blunting the relative-advantage angle). Management’s own framing — a “paradigm shift” toward durable mercantilist trade protection — is a hypothesis the price is treating as fact.

The aluminum sub-industry (relevant to STLD specifically). North America runs a structural flat-rolled aluminum deficit of ~1.4M+ tons, historically filled by high-cost imports now carrying a 50% tariff. This is the demand gap Aluminum Dynamics is built to fill. Aluminum sheet (especially can stock) is a more consolidated, more contract-oriented, somewhat less violently cyclical market than commodity steel — which is exactly why management is attracted to it. But it is a market STLD has never operated in, with its own metallurgical and qualification complexities.

Verdict: structurally BAD core industry, artificially propped; a structurally better aluminum adjacency, unproven for STLD. Global steel is a fungible-commodity, chronically over-supplied, deeply cyclical, price-taking business; the US sub-market is attractive right now only because of policy plus a demand mini-boom, into which the industry is over-investing. STLD is the best house in a bad neighborhood, and the neighborhood’s current desirability is rented from Washington. The aluminum bet is a deliberate attempt to diversify into a better neighborhood — sound in logic, unproven in execution.

4. Competitive Position

The moat, named precisely (Greenwald taxonomy): a cost advantage plus economies of scale plus a fabrication-driven utilization edge — NOT customer captivity. STLD’s genuine, financially-evidenced advantage is being the lowest-cost, most-flexible, most-vertically-integrated EAF operator in North America, supported by three mechanisms:

• Feedstock cost and integration (cost advantage). OmniSource — the largest scrap processor in North America — internalizes metallics. In FY2025 it moved ~6.2M gross tons of ferrous scrap and shipped ~65% to STLD’s own mills, giving controlled, lower-cost, quality-managed feedstock that integrated peers (buying iron ore and coke) and non-integrated EAF peers (buying scrap on the open market) cannot match. The Butler mill makes ~90% of its own liquid iron from recycled iron-oxide; the Mexican recycling footprint feeds Columbus and Sinton. This integration directly addresses the industry’s #2 risk factor (scrap supply/cost).
• Through-cycle utilization (the cleanest STLD-specific edge). Because Steel Fabrication (New Millennium) is a captive internal consumer of STLD’s own steel, the mills run hot even when external demand softens. In Q1-2026 STLD ran 89% utilization against a 77% industry average — a ~12-point gap that converts directly into lower unit costs and higher through-cycle returns. This downstream buffer, proportionally larger than Nucor’s, is a genuine structural advantage that protects relative profitability in downturns.
• Scale, breadth, and balance sheet. ~10% of US tons across the widest EAF product range (sheet, structural, rail, SBQ, merchant bar); the largest North American coater; an investment-grade balance sheet that funds counter-cyclical investment and survives troughs that wound peers.

What it is NOT. There is no customer captivity (steel is fungible, switching costs ~zero), no pricing power independent of the cycle (it is a commodity price-taker), and no network effect. The “moat” is operational excellence, cost leadership, and a captive-demand buffer in a commodity — a real but narrow advantage that protects relative profitability, not absolute pricing.

The financial proof — and its limit. The through-cycle ROIC dispersion both proves the quality and exposes the cyclicality: ~43% (2021) → ~42% (2022) → ~23% (2023) → ~14% (2024) → ~10% (2025). A trough ROIC of ~10% is roughly STLD’s cost of capital — meaning at the bottom, the best operator in US steel, running flat-out at record utilization, barely clears its hurdle rate, because the swing factor is the metal spread, not volume or operating skill. The durable edge is survivability and relative outperformance: STLD earns a profit every year through every trough (FY25 NI still $1.19B) while CLF and U.S. Steel bleed at the bottom. That through-cycle resilience is real and rare. But it is the resilience of the best commodity operator, not the pricing power of a branded franchise.

Versus peers. Nucor is the closest comparable-quality EAF operator (factor-similarity 0.957) and the natural benchmark. NUE edges STLD on scale (~18% vs. ~10% share), credit rating, and the 53-year dividend record; STLD edges NUE on proportional fabrication buffer / through-cycle utilization and — decisively — on the aluminum growth optionality NUE lacks. The two share the low-cost EAF crown, confirming the edge is relative cost/scale/integration, not uniqueness. CLF and U.S. Steel (Nippon) are higher-cost integrated producers that benefit disproportionately from the tariff (it props up their uneconomic tons), narrowing STLD’s relative advantage in the current regime. CMC is a smaller rebar/long-products EAF player; Ternium (TX) is the Mexican/LatAm integrated exposure.

Verdict: a durable cost-scale-and-utilization advantage in a commodity business — operational excellence, not a wide moat. If you removed STLD’s cost leadership and captive-demand buffer, its economics would deteriorate toward the integrated-peer level — so it is a real, financially-evidenced advantage. But it is a relative edge inside a price-taking commodity, not a source of pricing power, and it is partially neutralized by a tariff that helps every US producer. Best operator (with Nucor); structurally challenged industry.

5. Growth History and Forward Opportunities

History — cyclical, with a genuine secular-capacity overlay. Revenue grew from ~$9.6B (2020) to a $22.3B peak (2022) and back to ~$18.2B (2025) — the headline swing is overwhelmingly price/spread cyclicality, not secular volume. But underneath, STLD has genuinely added capacity: steel shipments have grown structurally as the ~3M-ton Sinton (TX) flat-roll mill ramped to full rates, lifting STLD to record 13.7M-ton shipments in FY2025. The honest read: top-line swings are dominated by commodity price, but volume growth is real and self-funded, and the next leg of volume comes from aluminum.

The organic growth engine (~$5B+ since ~2020). STLD has self-funded the largest organic expansion in its history across three projects management pegs at ~$1.4B of combined through-cycle EBITDA:

• Sinton, Texas flat-roll mill (~$2B) — ~3M tons/yr of advanced sheet, now ramped; opened STLD to the Southern/Mexican markets and the southwest auto/energy supply chain. Largely complete and contributing.
• Four value-added flat-roll coating lines — galvanizing/paint capacity (Sinton + Heartland/Terre Haute) that lifts STLD’s value-added mix; “operating at full capacity,” directly leveraged to the 2025 trade-case wins (STLD is the largest NA coater).
• Aluminum Dynamics (~$2.7B) — the marquee bet and the key differentiator vs. Nucor. A ~650k-metric-ton Columbus, MS flat-rolled aluminum mill plus recycled-slab “cathouses” in Columbus and San Luis Potosí, Mexico. Target mix 45% can sheet / 35% automotive / 20% industrial by 2027; target $650–700M through-cycle EBITDA (plus $40–50M of related recycling) — a ~30% uplift to the FY25 ~$2.2B EBITDA base, into structurally better end-markets. Ramp status: began production 2025; Q4-25 ~14kt shipped → Q1-26 ~22kt → Q2-26 guided 60–70kt; management targets exiting 2026 at a 90% monthly capacity run-rate. The hot mill runs at rated capacity; two of three cold mills and the first automotive CASH line are operational; multiple can/industrial certifications achieved and automotive qualifications “in coming weeks.”

The critical caveat on aluminum. This is a promise, not a banked result. Aluminum Operations lost ~$173M (FY25) and ~$65M (Q1-26) — the latter on a January production pause and a quality-stain inventory write-down on a first-of-kind facility. Flat-rolled aluminum (especially automotive-grade and can stock) is metallurgically demanding and qualification-gated; STLD has never operated in it. The $650–700M must be probability-weighted, not capitalized at face value, and exit-2026 at 90% capacity is an aggressive bar given the quality stumbles already visible. The offset: current aluminum spot spreads are “significantly above” the assumptions underpinning that $650–700M (a 50% import tariff plus an Iran-war-disrupted global market), so if the ramp executes, near-term economics could exceed the through-cycle target.

Forward demand drivers being priced in. Management frames the steel demand story around data centers, reshoring/onshoring of manufacturing, infrastructure (IIJA), grid/electrification, multifamily construction, energy (oil & gas pipe, solar), and — distinctively — low-carbon auto, where STLD has become a “supplier of choice” for European and Asian transplant automakers that value its lower embedded carbon. Long products (structural, rail, SBQ) are described as sold out / record. The aluminum leg adds beverage-can (countercyclical) and automotive demand.

Verdict: mixed-to-higher-quality growth — a notch above Nucor’s, if aluminum delivers. The historical “growth” is mostly cyclical price oscillation, but the self-funded volume build is real and disciplined, and the aluminum leg is a credible move into better-structured end-markets that Nucor cannot match. Until aluminum proves out, though, a meaningful slice of STLD’s growth premium is an unproven option, not earnings.

6. Financial Quality

Revenue composition & cyclicality. FY2025 revenue of ~$18.2B is ~70% Steel Operations, with Metals Recycling, Steel Fabrication and (newly) Aluminum the balance. The top line is dominated by the metal spread; the same tonnage produces wildly different revenue and profit depending on where HRC and scrap sit. Revenue swung from $9.6B (2020) to $22.3B (2022) to $18.2B (2025) on roughly stable-to-rising volume — proof that price, not volume, drives the income statement.

Margin structure and trajectory. Gross margin has swung from ~13% (2020, 2025 troughs) to ~29% (2021 peak); operating margin from ~9% to ~25%; EBITDA margin from ~12% to ~27%. The driver is the metal spread (steel price − scrap cost), amplified by operating leverage on a high-utilization, variable-cost base. FY2025 compressed to a 13.2% gross / 8.8% operating margin — a cyclical trough — before the 2026 tariff-driven spread recovery (Q1-26 operating income jumped 73% sequentially in the steel segment).

Through-cycle returns — the cyclicality, quantified. ROIC: ~10.5% (2020) → ~42.9% (2021) → ~41.6% (2022) → ~22.7% (2023) → ~13.7% (2024) → ~10.0% (2025). ROE follows: ~12% → ~51% → ~40% → ~20% → ~11% → ~8%. The 2025 trough ROIC of ~10% ≈ STLD’s WACC — the single most important fact about the business’s quality: the best EAF operator in North America, running at record 86%+ utilization, barely cleared its cost of capital at the bottom. This is a best-in-class cyclical, not a secular compounder. Returns improve dramatically with the spread, not durably with scale.

Quality of earnings — a two-sided distortion that partly cancels. TTM EPS of ~$9.34 (and the Q2 guide of $3.51–3.55) is simultaneously:

• DEPRESSED by the aluminum startup: −$173M operating loss in FY25 and −$65M in Q1-26, sitting on ~$4.1B of aluminum assets earning a negative return today. Strip aluminum and “Other,” and the steel/recycling/fab core earned ~$692M of operating income in Q1-26 vs. the $403M headline net income — the steel cycle runs hotter than the headline shows.
• FLATTERED by the tariff-propped spread: HRC ran $850 → $975 → >$1,000, and with ~75–80% of flat-rolled on ~2-month lagging contracts, the Q2 guide mechanically inherits peak spot pricing. A large share of that spread is a Section 232 policy prop, not durable earnings. Do not capitalize the Q2 guide as run-rate.
• The net: current earnings understate the steel-cycle strength but overstate the durable earnings power. Normalized (steel mid-cycle + a probability-weighted aluminum) EPS is plausibly in the low-teens, not the ~$14 the Q2 guide annualizes to.

Free cash flow — working capital is a battery, and accounting is clean. This is the most important nuance. A near-peak HRC print is the worst time for cash conversion: as prices rise, receivables and inventory swell and absorb cash. Q1-26 turned $403M of net income into only ~$10M of FCF (CFO just $148M; CFO/NI 0.37x) because working capital built ~$459M as prices rose (plus a $120M annual profit-sharing funding). FY25 FCF was just $501M (CFO $1,449M − capex $948M, with a ~$442M working-capital drag); FY24 FCF was negative (−$23M) during the capex super-cycle. The flip side: when the cycle rolls over, working capital liquefies and FCF can exceed net income (FY23 CFO was $3.5B). With capex falling from $1.87B (FY24) to a ~$600M 2026 guide as the aluminum build completes, and aluminum ramping, management cites a normalized free-cash profile of ~$2.4B/yr (~$3.2B ex-growth) — credible directionally, though cycle-dependent. Net income tracks CFO well over the cycle (cash_flow_to_net_inc ~1.2x); the divergence is timing, not quality. Accounting is high-quality: no goodwill impairments FY23–25, a stable ~20–22% tax rate, modest ~$69M SBC, and a profit-sharing scheme that auto-flexes labor cost down with earnings.

Balance sheet — a genuine fortress. Q1-26: ~$4.2B total debt, ~$0.56B cash → ~$3.64B net debt; ~$9.13B equity; net debt ~1.5–1.7x trough EBITDA; ~$2.0B liquidity ($0.8B cash/investments + $1.2B undrawn revolver); current ratio ~3.1x; investment-grade; no maturity wall until ~$350M in October 2027. This is what lets STLD buy back stock counter-cyclically and finish a ~$2.7B aluminum build through a downturn without strain.

Verdict: a best-in-class cyclical with fortress finances and clean accounting — but economics that improve with the spread, not durably with scale. The quality is in the survivability, the cost position, the capital discipline and the balance sheet — not in any escape from the commodity cycle. At the 2025 trough, scale and excellence bought a ~10% ROIC, no more.

7. Capital Allocation

This is STLD’s strongest dimension, and the one that most justifies a quality premium — pressure-tested, the Millett reputation holds.

Buybacks — genuinely counter-cyclical, the rarest and most valuable trait. Share count fell ~31.5%, from 211M (2020) to 144.4M (Q1-26). Crucially, the pace tracks value, not momentum: heavy repurchases when the stock was cheap (FY22 ~$1.80B, FY23 ~$1.45B at ~$80–130 and 3–5x EV/EBITDA) and a deliberate throttle to ~$115M in Q1-26 (vs. ~$250M in Q1-25) at ~$251 and a richest-ever multiple. Management explicitly attributed the Q1-26 slowdown to working-capital needs and the rich price. This is the opposite of the buy-high reflex that defines most of the richest-ever cohort (e.g., HUBB, WWD) — STLD bought the most stock when it was cheapest. $687M remained authorized at March 2026.

Dividend — 14 consecutive years of increases at a ~13% CAGR and a conservative ~24% payout, complemented by the variable buyback. A “progressively positive” base dividend by design, leaving ample room for growth investment.

M&A and organic investment — disciplined, return-anchored, and willing to walk. STLD’s history is built on smart deals (OmniSource recycling, New Millennium fabrication, Roanoke/structural, Mexican recycling) and, more recently, on ~$5B+ of self-funded organic capacity (Sinton, coating lines, Aluminum Dynamics) plus small accretive tuck-ins (New Process Steel, Dec-2025). Decisively, management walked away from the BlueScope JV in February 2026 after a “best and final” offer was rejected — refusing to overpay. The only fair caveat: the ~$5B build’s through-cycle return targets are management’s own and unproven (aluminum is mid-startup; consolidated ROIC is at a cycle-trough ~10%), so the capital-allocation grade rests partly on a promise.

Incentive alignment — a standout, and unusual for the cohort. The proxy reveals genuinely return-anchored, peer-relative pay: the annual cash bonus funds only on adjusted net income above a threshold return on average stockholders’ equity (an ROE gate), and the 3-year LTIP scores management on four equally-weighted measures — Revenue Growth, Operating Margin, Cash-Flow-from-Operations/Revenue, and after-tax ROIC — measured RELATIVE to CLF/CMC/NUE/Metallus. Three of four are quality/return metrics, benchmarked to peers, not absolute tonnage. CEO target pay sits below the 25th percentile of the peer group; governance is clean (hedging banned, clawback in place, no excise-tax gross-ups); say-on-pay passed at ~92% in 2025. For a commodity producer, this is best-in-class design.

The one offsetting tell — insiders never buy. Across the entire 506-filing, five-year Form 4 corpus, there is not a single open-market purchase (code P) by anyone — every transaction is a grant (A), tax-withholding (F), planned sale (S), or gift (G), all 10b5-1. The reassuring half: the three top officers (Millett, Wagler, Schneider) did zero discretionary open-market selling — no top-ticking into the ATH — and insider ownership is meaningful (Millett ~2.1%, all D&O ~6.6%, co-founder Teets ~3.6%). The unreassuring half: the people who know the business best have never once chosen to buy it personally with cash, even when the stock traded sub-$130. Read it as conviction-neutral, not conviction-positive.

Verdict: top-tier capital allocation — arguably the best in North American steel. Counter-cyclical buybacks, a long dividend record, return-anchored peer-relative pay, willingness to walk from a deal, and a fortress balance sheet. The only asterisks are the unproven returns on the ~$5B build and the absence of any personal insider buying. Management has earned the benefit of the doubt; the price has already priced in the doubt being resolved favorably.

8. Changes and Headwinds — Last Two Years

Strategic / operational.

• Aluminum Dynamics startup (the dominant change). Columbus, MS flat-roll aluminum mill began production in 2025; mid-ramp through 2026 with a −$173M FY25 / −$65M Q1-26 loss, a January production pause, and a quality-stain inventory write-down — alongside genuine progress (multiple certifications, automotive qualifications imminent, 90%-capacity exit-2026 target).
• Section 232 escalation (the dominant tailwind). Tariff doubled to 50% (June 2025) and extended to full derivative value (April 2026); 2025 trade-case wins on coated products directly lifted STLD’s value-added spreads (it is the largest NA coater).
• Sinton mill fully ramped; four value-added coating lines at full capacity — lifting volume and value-added mix.
• BlueScope JV pursued and abandoned (Feb 2026) — best-and-final offer rejected, no engagement since; management refused to overpay.
• Capex peak passed: FY24 $1.87B → FY25 $948M → 2026 guide ~$600M, an inflection toward materially higher free cash flow.

Leadership / board. Co-founder Richard Teets and director Shaheen retiring at the 2026 AGM (board shrinking 9→7); SVP Pushis retirement with Alvarez promoted to lead aluminum. Founder Mark Millett remains Chairman & CEO; the highly regarded Theresa Wagler remains CFO — succession depth is a watch-item but not an immediate concern.

Headwinds. A reversible tariff regime; a ~9–12M-ton new-capacity wave landing into flat demand; near-record Chinese steel exports; pig-iron/scrap input-cost pressure (partly mitigated by OmniSource); aluminum startup execution risk; and the cyclical reality that 2026’s recovering earnings sit on a policy-propped spread.

Verdict: the changes net to thesis-strengthening operationally (capex peak passed, value-add mix up, aluminum optionality live) but thesis-fragile on durability (the earnings recovery is tariff-dependent and the aluminum payoff is unproven). The next 12–18 months — does aluminum ramp on schedule, does the tariff hold — will resolve most of the open questions.

9. Risk Analysis (Risk Matrix)

Master risk read. Risks 1–3 and 5–6 are tightly correlated and would fire together in a classic Marathon top-of-capital-cycle unwind (tariff softens → imports return → spread compresses → the new tons hit → utilization and ROIC fall → the multiple de-rates). The one idiosyncratic, steel-cycle-independent risk is aluminum execution (#4) — the only place STLD can disappoint while steel is fine, and the only one with ~$4.1B of dedicated assets behind it. Genuine financial-distress risk (#11) and catastrophic-loss risk are low; the realistic bad case is a cyclical de-rating, not impairment.

10. Valuation Discussion (Embedded Expectations)

No price target, no recommendation — embedded-expectations and scenario framing only.

Where the multiple sits. At $251 (June 25, 2026), on 144.4M shares, market cap is ~$36.3B and — with ~$3.64B net debt — enterprise value is ~$39.9B. Against TTM figures (sales $19.0B, EBITDA $2.458B, EPS $9.34), that is ~16.2x EV/EBITDA, ~2.1x EV/sales, ~27x trailing EPS, ~4.0x book. On AZI’s own-history percentiles, STLD trades at the 99th percentile on P/B, 99th on P/S, 94th on P/E — a ~97th-percentile composite, i.e., the richest valuation in its own decade-plus history. (Note: a third-party EV snapshot struck at the March-31 quarter close understates this — at the live price the multiples are materially higher; figures above are re-derived at $251.)

Why the trailing multiple overstates the richness — and why it still matters. The 16x trailing EV/EBITDA is inflated because TTM EBITDA is trough-ish (2025 was the cyclical bottom) and dragged by ~$173M of aluminum startup losses. On a forward basis — FY26 recovering steel earnings (the Q2 guide annualizes toward ~$14, EBITDA plausibly ~$3.0–3.3B) — the multiple compresses to ~12–13x EV/EBITDA and ~18x earnings. On a fully-ramped normalized basis (steel mid-cycle + Aluminum Dynamics at its $650–700M target, EBITDA ~$3.5–3.8B) it compresses further toward ~10.5–11x. So the bull is not paying 16x for steady-state; the bull is paying ~16x trailing for a business about to (a) recover off a trough and (b) light up a new ~30%-of-EBITDA aluminum leg.

What the price embeds. To justify $251, the market must underwrite roughly: (1) the Section 232 spread holds at or near current levels long enough for FY26–27 steel earnings to validate a mid-teens-or-better normalized EPS; (2) Aluminum Dynamics ramps to its $650–700M EBITDA target on roughly management’s schedule, with the current above-plan spot spreads sweetening the early years; and (3) the ~9–12M-ton capacity wave is absorbed by data-center/reshoring/onshoring demand without collapsing the spread. That is a coherent, internally-consistent bull case — but every leg is a forward assumption, and two of the three (tariff durability, aluminum execution) are outside management’s control or unproven.

Scenario sketch (illustrative, not targets).

• Bear — tariff softens or an adverse ruling lands, the spread mean-reverts as new tons hit, aluminum slips past 2026; normalized EPS settles ~$8–10 and the multiple de-rates toward a through-cycle ~10–12x earnings / ~8x EV/EBITDA. The realistic downside is a cyclical de-rate, valuing the equity well below today.
• Base — tariff broadly holds, steel earns a recovering mid-cycle, aluminum reaches ~80–90% of target by 2027; normalized EPS ~$12–14, EV/EBITDA settles ~10–12x — roughly fair-to-full around today’s price, with the aluminum ramp the swing.
• Bull — Section 232 codified into durable law, aluminum hits $650–700M on schedule with above-plan spreads, demand absorbs the new capacity; normalized EPS pushes toward ~$15–17 and a structurally-higher-spread “paradigm shift” earns a sustained premium multiple — the case the current price and the ~$262–291 sell-side targets implicitly back.

Comp context. STLD screens optically richer than Nucor on trailing EV/EBITDA (~16x vs. ~9x), but that gap is largely the aluminum-loss-and-trough-EBITDA artifact; on forward/normalized EBITDA the two converge, both at the rich end of their own histories. CLF/U.S. Steel (integrated) and CMC (long products) trade lower on quality, not just cycle. Versus its own history, STLD has never been this expensive on book or sales. The embedded-expectations conclusion: the market is correctly pricing STLD as the best-in-class operator with a credible growth leg — and is generously crediting both the tariff’s durability and the aluminum payoff before either is proven. No price target; the body takes no position.

11. Variant Perception

Consensus belief. STLD is a best-in-class, well-managed EAF steelmaker enjoying a tariff-driven earnings recovery, with a transformational aluminum growth leg about to inflect — a “quality compounder in metals,” worth a premium multiple. Sell-side is clustered Neutral-to-Overweight with price targets of ~$262–291 (above the current $251), i.e., a modestly constructive consensus that already embeds the recovery and much of the aluminum option.

The strongest bull case. This is the best capital allocator in North American steel (counter-cyclical buybacks, ROIC-anchored pay, fortress balance sheet) catching three tailwinds at once: a durable “paradigm shift” toward protected US steel pricing, a completed ~$5B capacity build that lifts volume and value-add just as capex falls (free-cash inflection), and a ~$650–700M aluminum EBITDA leg into structurally better end-markets that no peer can match. If the tariff holds and aluminum delivers, normalized earnings step structurally higher and the premium multiple is earned, not lent.

The strongest bear case. Steel is a fungible, over-supplied, price-taking commodity; STLD’s entire current earnings strength is a reversible policy artifact, and its through-cycle ROIC still bottoms at ~10% (its WACC) despite record utilization. The industry is over-building into flat demand (Marathon late-cycle), the aluminum leg is losing money and unproven (already stumbling on quality), and the stock is at its richest-ever valuation on book and sales with a high-beta (1.16) momentum profile that crashes at cyclical turns. You are paying a peak multiple on tariff-inflated, trough-recovering earnings, plus a premium for an un-earned aluminum profit — the textbook setup for a cyclical de-rate.

The 3–5 assumptions that matter most:

1. Tariff durability. Does Section 232 (and the derivative closure) hold long enough to make today’s spread a durable baseline? (Outside management’s control.)
2. Aluminum execution. Does Aluminum Dynamics reach ~$650–700M EBITDA on schedule, or does the ramp slip and the quality issues recur? (The single idiosyncratic swing factor.)
3. Net-vs-gross capacity. Do retiring blast furnaces offset the ~9–12M new EAF tons, keeping the spread from collapsing?
4. Demand absorption. Do data centers / reshoring / low-carbon auto genuinely absorb the new tonnage?
5. Normalized earnings power. Is mid-cycle EPS low-teens (base) or mid-teens-plus (bull) — i.e., is today’s price fair or cheap on the right normalized number?

Factor-positioning read (overlay, not a call). STLD is empirically a high-beta (1.16) Materials / Infrastructure cyclical with a notable DividendYield loading and negative LowVolatility loading — a quality-cyclical, not a value or defensive name. The tape is a violent one-way street up: +98% over twelve months (Sharpe ~2.8), +135% off the August-2025 low to the June ATH, now ~11% off that high. Momentum has positive expectancy on average but crashes at cyclical turns — and this is a stock whose lifetime max drawdown is ~−87%. The factor read says the market has firmly “started to agree,” which sharpens the framing: this is not a falling knife the value buyer is early on — it is a crowded, late-cycle momentum/quality-cyclical trade at a richest-ever multiple, where the risk is buying the top of a policy-and-spread cycle, not catching a falling knife. That is evidence for where consensus may be offsides (over-crediting durability), and it is why Claude's Take frames the name as quality to accumulate cheaper, not chase here.

Where I think consensus is offsides. Consensus is right about the quality and probably right that aluminum eventually works. Where it is likely too generous is in treating the tariff-propped spread as a durable baseline and in capitalizing the aluminum profit before the ramp is proven — paying a richest-ever multiple for both. The variant view is not that STLD is a bad business (it isn’t) but that the price already pays for a benign resolution of two genuinely uncertain, partly-uncontrollable variables.

12. Fact vs. Interpretation Table

13. Open Questions

1. Tariff durability: Will Section 232 (and the derivative-value extension) survive legal challenge and a future administration, or is the >$1,000 HRC a 1–2-year phenomenon?
2. Aluminum ramp: Does Aluminum Dynamics exit 2026 at 90% capacity and reach $650–700M EBITDA — and do the Q1 quality issues recur as the auto/can qualification mix scales?
3. Above-plan aluminum spreads: Management hinted the current aluminum spread implies a “structural shift” above the $650–700M assumptions — how much, and how durable?
4. Net capacity: Do US blast-furnace retirements genuinely offset the ~9–12M new EAF tons, or does net capacity rise and compress the spread?
5. Normalized EPS: Is mid-cycle earnings power low-teens or mid-teens-plus once aluminum contributes — i.e., is today’s price fair or cheap on the right number?
6. Succession: With Millett still CEO and the board shrinking, what is the long-term succession plan beyond Wagler/Schneider?
7. Next growth bet: Management signaled more aluminum (and downstream) investment ahead — at what return and capital intensity, and does it extend or strain the discipline?

14. What Must Be True

Bull case — what must be true:

• The Section 232 regime holds, keeping the US spread structurally above world levels, so FY26–27 steel earnings validate a mid-teens-or-better normalized EPS.
• Aluminum Dynamics ramps to ~$650–700M EBITDA on schedule, with above-plan spot spreads boosting the early years — turning a ~30%-of-EBITDA leg from loss to profit.
• Demand (data centers, reshoring, low-carbon auto) absorbs the new industry capacity without collapsing the spread; capex falls and free cash flow inflects toward ~$2.4–3.2B/yr.
• Falsification test: The HRC metal spread mean-reverts toward world levels (tariff softened or out-supplied), and/or Aluminum Dynamics is still loss-making or sub-50%-of-target EBITDA at the end of 2026. Either breaks the bull.

Bear case — what must be true:

• The tariff is a transient policy artifact; once it softens or imports route around it, the spread compresses toward the China-set world level and steel earnings re-trough.
• The ~9–12M-ton capacity wave lands into flat demand (Marathon late-cycle), pushing utilization and ROIC back toward the ~10% WACC floor.
• Aluminum disappoints — the ramp slips, quality issues recur, and the $650–700M is delayed or impaired — so the growth premium evaporates.
• The richest-ever multiple de-rates toward a normal through-cycle level.
• Falsification test: Section 232 is codified into durable law and Aluminum Dynamics hits its $650–700M run-rate on schedule — together, these convert the policy prop and the lottery ticket into a durably higher, less cyclical earnings base, and the bear is wrong.

15. Source Appendix

See the separate Source Appendix (STLD_source_appendix.md, Appendix B of the combined report) for the full list of primary sources — STLD FY2025 10-K (filed 2026-02-27), Q1-2026 10-Q (filed 2026-04-27), the 2026 proxy (filed 2026-03-27), the 506-filing Form 4 corpus, the Q1-2026 earnings-call transcript (2026-04-21), the Q2-2026 guidance release (2026-06-17), and and third-party quantitative data sources — each with access dates. Every non-obvious fact in this memo traces to a primary source in that appendix.

This analysis (sections 1–15) takes no position and contains no price target or buy/sell recommendation; the only position in this document is the clearly-labeled Claude's Take block at the top, which is the author’s own subjective opinion and general information only — not investment advice.

APPENDIX A — Standard Diligence Questionnaire

Steel Dynamics, Inc. (NASDAQ: STLD) — supplemental to the research memo, report date 2026-06-26.

Grounded in primary filings. Fact / Interpretation / Assumption labeled where it matters.

General

What thoughtful questions have other investors asked about this company? The recurring institutional questions, evident from the Q1-2026 call: (1) the aluminum ramp — startup issues, the inventory write-down, ramp cadence, and how far spot spreads exceed the through-cycle plan; (2) capital allocation now that capex is falling and free cash is inflecting; (3) the sustainability of the steel-price rally and whether weekly, gradual price increases are more durable than past spikes; (4) pig-iron/scrap input costs; (5) the abandoned BlueScope JV and the appetite for further M&A and long-product/downstream expansion; and (6) steel-vs-aluminum substitution now that STLD sits on both sides. The unifying theme: how much of the current earnings power is durable vs. policy/cycle, and how much aluminum upside to credit before it’s proven.

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? Recovering off a recent low toward a tariff-propped high. FY2025 was the cyclical trough (diluted EPS $7.99, ROIC ~10%); 2026 is recovering hard on the Section 232 spread (Q1-26 $2.78; Q2 guide $3.51–3.55). The steel earnings are moving toward a policy-driven high; the consolidated number is held down by the aluminum startup loss. (Interpretation.)

Driven by the external environment or internal actions? Predominantly external (steel price / metal spread / tariff policy), with a meaningful internal overlay (record utilization from the fabrication buffer, the value-add coating-line mix, the self-funded capacity build). Spread, not management, is the swing factor — but STLD out-earns peers on the same spread. (Fact + Interpretation.)

How stable are revenues? Unstable by nature — revenue swung $9.6B → $22.3B → $18.2B (2020→2022→2025) on roughly stable-to-rising volume; price dominates. (Fact.)

Outlook for products/services? Steel demand firm in data centers, infrastructure, reshoring, multifamily, energy, and low-carbon auto; long products (structural, rail, SBQ) described as sold-out/record. Aluminum addresses a ~1.4M-ton structural North American deficit. (Fact = management framing; durability is Interpretation.)

How big is this market — growing, shrinking, domestic, international? US finished-steel consumption ~95–100M tons/yr, mature/cyclical (flat-to-low-single-digit growth). Flat-rolled aluminum is a structurally deficit, somewhat-better-structured adjacency. Predominantly domestic, with growing Mexican exposure. (Fact.)

Business Quality & Competitive Moat

Is the industry getting more or less competitive? More — ~9–12M tons of new US EAF sheet capacity is landing 2026–2029 into flat demand (Marathon late-cycle over-build). The tariff currently masks this by excluding imports. (Fact + Interpretation.)

How profitable is the business (ROIC, ROE)? Through-cycle ROIC ~10% (2025 trough) to ~43% (2021 peak); ROE ~8% to ~51%. At the trough — running flat-out — the best EAF operator barely clears its ~10% WACC. (Fact.)

How profitable is the industry — competitors, barriers to entry? Structurally low through-cycle returns; integrated peers (CLF, U.S. Steel) lose money at the bottom. Barriers are capital (a new mill is $2B+ and years) and scrap access, but the product is fungible — barriers protect incumbency, not pricing. (Fact + Interpretation.)

Can the business be easily understood? Yes — a four-segment metals producer earning a metal spread, with a new aluminum leg. The complexity is in normalizing cyclical earnings and probability-weighting the aluminum ramp, not in the model. (Interpretation.)

Can it be undermined by foreign low-cost labor? Indirectly — via Chinese/global steel over-supply and imports, the permanent ceiling on world prices. The current buffer is entirely the Section 232 tariff. (Fact.)

Do brands matter? No. Steel is a commodity; the “brand” is reliability, supply-chain solutions, low-carbon credentials, and service — real for relative share/utilization, irrelevant to commodity pricing. (Interpretation.)

Nature of competition / customers’ switching costs? Competition is on cost, reliability, product breadth and service; customer switching costs are ~zero (fungible commodity). STLD’s edge is being lowest-cost and most-integrated, plus captive fabrication demand — not customer lock-in. (Fact + Interpretation.)

Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet? The OmniSource scrap network and the through-cycle utilization edge are economic assets not on the balance sheet. Conversely, the ~$4.1B of aluminum assets are on the balance sheet but earning a negative return today. (Interpretation.)

Off-balance-sheet liabilities? None material flagged; standard operating leases and purchase commitments. Accounting is conservative (no impairments FY23–25). (Fact.)

How conservative is the accounting? High quality — no goodwill impairments FY23–25, stable ~20–22% tax rate, modest ~$69M SBC, profit-sharing that auto-flexes labor cost down with earnings, net income tracking CFO ~1.2x over the cycle. (Fact.)

How CapEx-hungry is the business? Very, structurally — but past its peak. Capex ran $1.87B (FY24) during the aluminum/Sinton build, falling to a ~$600M 2026 guide as projects complete. Maintenance capex is far lower; the recent surge was growth. (Fact.)

Capital Allocation & Management

How much FCF, and how is it used? Cyclical and working-capital-sensitive: FY25 FCF $501M (capex- and WC-depressed); management cites a normalized ~$2.4B/yr (~$3.2B ex-growth). Used for counter-cyclical buybacks, a growing dividend, and self-funded growth. (Fact + Interpretation.)

Significant acquisitions recently? Small accretive tuck-ins (New Process Steel, Dec-2025); the larger move was ~$5B+ of organic build. Management walked away from the BlueScope JV (Feb-2026) rather than overpay. (Fact.)

Buying back shares? Yes, aggressively and counter-cyclically — share count −31.5% (211M→144.4M), heavy at $80–130 (2022–23), deliberately throttled to $115M at $251 (Q1-26). The opposite of buy-high. (Fact.)

Issuing shares to insiders? Routine equity comp only (~$69M SBC); no large insider issuance; buybacks vastly outweigh dilution. (Fact.)

Compensation policy of directors/management? Standout. Annual bonus gated on adjusted net income above a threshold ROE; 3-year LTIP on ROIC/operating-margin/cash-conversion/revenue-growth measured relative to CLF/CMC/NUE/Metallus; CEO target pay <25th percentile of peers; clean governance (hedging banned, clawback, no gross-ups); ~92% say-on-pay. (Fact.)

Motivations of management? Founder-led (Millett, co-founder/CEO, ~2.1% owned; D&O ~6.6%); incentives are genuinely return- and peer-relative. The one tell: zero open-market insider purchases in the entire 5-year Form 4 corpus — no top-ticking, but no personal-cash conviction either. (Fact + Interpretation.)

Valuation & Market Data

ADR, MLP, or K-1 issuer? No — a standard US C-corporation common stock (NASDAQ: STLD), 1099 dividends. (Fact.)

Dividend policy? ~14 consecutive years of increases, ~13% CAGR, ~24% payout, ~0.8% yield — a “progressively positive” base dividend complemented by the variable buyback. (Fact.)

How profitable is the business? See ROIC/ROE above — high at the peak, ~WACC at the trough; best-in-class for the industry but capped by the commodity. (Fact.)

Is net income diverging from cash from operations? Only on timing: a near-peak price print converts poorly (Q1-26 CFO/NI 0.37x as working capital builds), and a down-cycle converts richly (FY23 CFO $3.5B). Over the cycle CFO ~1.2x net income — no quality red flag. (Fact.)

Risks & Downside

What factors would cause the stock to decline? A Section 232 rollback/adverse ruling; the HRC spread mean-reverting as the new-capacity wave lands; an aluminum ramp slip or recurring quality issues; a non-residential construction downturn; a broad cyclical de-rate from the richest-ever multiple. (Interpretation.)

Risk of a catastrophic loss? Low. Fortress investment-grade balance sheet (net debt ~1.5–1.7x trough EBITDA, ~$2.0B liquidity, no maturity wall to Oct-2027); STLD has never posted an annual loss. The realistic bad case is a cyclical de-rate, not impairment. (Interpretation.)

Chance of a total loss? Negligible absent an extraordinary, unprecedented event. (Interpretation.)

Recent News & Events

Has the business environment changed recently? Yes — materially. Section 232 to 50% (Jun-2025) plus the derivative-value extension (Apr-2026) drove HRC past $1,000 and a ~135% stock run off the August-2025 low; Aluminum Dynamics began production (mid-ramp, loss-making); 2025 coated-product trade-case wins lifted value-add spreads. (Fact.)

Significant acquisitions? New Process Steel tuck-in (Dec-2025); BlueScope JV pursued and abandoned (Feb-2026). (Fact.)

Change in accounting policies? None material. (Fact.)

Recent changes — new markets, facilities, management? New aluminum business (Columbus MS + San Luis Potosí MX); Sinton mill and four coating lines fully ramped; capex peak passed; board shrinking 9→7 with co-founder Teets and director Shaheen retiring at the 2026 AGM; Alvarez promoted to lead aluminum. Founder Millett remains Chairman & CEO; Wagler remains CFO. (Fact.)

APPENDIX B — Source Appendix

Steel Dynamics, Inc. (NASDAQ: STLD) — research memo dated 2026-06-26. Primary sources prioritized; third-party quantitative sources reconciled to filings. Access date 2026-06-26 unless noted.

Primary — SEC filings (EDGAR, CIK 0001022671)

Primary — company disclosures & transcripts

Third-party — quantitative (reconciled to filings)

Peer reference & analytical frameworks (public)

Every non-obvious fact in the memo traces to a primary source above. Management commentary is treated as a hypothesis and validated against filings and external data.