HD Hyundai Electric

Phase A — Understand the business

Lens 1 · Company Overview

What it actually is. HD Hyundai Electric (formerly Hyundai Electric & Energy Systems, spun out of Hyundai Heavy Industries in 2017) is a heavy electrical-equipment manufacturer — it builds the iron-and-copper hardware that moves high-voltage electricity around a grid. It is not a utility, not a renewables developer, and not a "clean energy" name in the ESG sense. It sits inside the "energy" coverage bucket, but functionally it is a grid-infrastructure capital-goods company whose fortunes are levered to one thing: how much new high-voltage power capacity the world is building. Right now, that is a lot — AI data centres + aging-grid replacement in North America.

Three product lines (≈100% of revenue is electrical equipment — a rare pure play):

• Power Equipment (the crown jewel) — ultra-high-voltage (UHV/EHV) power transformers up to 765kV, gas-insulated switchgear (GIS), energy-storage/solutions. The 765kV transformer is a ~$7M+ per-unit, engineering-gated product only a handful of firms on earth can build. This is where the margin and the moat live.
• Distribution Equipment (switchgear) — distribution transformers, low/medium-voltage circuit breakers. More commoditised, more cyclical; fell −24.2% YoY in Q1 2026 on a tough comp + Middle-East delivery delays.
• Rotating Machinery — motors and generators for marine (shipbuilding), petrochemical, and power-gen customers. Mature, stable-margin; +10.8% YoY in Q1 2026 on strong marine demand.

Customers. US investor-owned utilities and grid operators, hyperscaler / "Big Tech" data-centre builders (a July 2026 framework deal is explicitly with a "global tech giant"), Middle-East national power projects, European TSOs, and marine/shipyard buyers (partly captive to the HD Hyundai group's own shipbuilding arm). North America is now the largest single market at ~47% of FY2025 revenue.

Contract structure — the key insight. HD Hyundai Electric runs a "Slot Reservation" (a.k.a. selective-ordering) system: because transformer capacity is the binding constraint, customers pre-book production slots and pay a premium for guaranteed delivery timelines, while the company declines lower-margin work. This is take-or-pay-like pricing power — the seller, not the buyer, sets terms. It is the single most important structural fact about the company and the reason margins have gone from low-single-digits to ~25%+.

Corporate parent. ~35.6–37.2% owned by holdco HD Hyundai Co. (KRX 267250), itself controlled by the founding Chung family (Chung Mong-joon / Chung Ki-sun). So minority holders sit under a chaebol holdco — governance and capital-return decisions are made with the group in mind (see Lens 9/13).

Lens 2 · Supply Chain

Upstream inputs → HD Hyundai Electric → end customer, named at each node:

Upstream (inputs / chokepoints):

• Grain-Oriented Electrical Steel (GOES / CRGO) — the transformer-core material and the system-wide bottleneck. Global GOES production is limited and concentrated; the US has only one domestic producer (Cleveland-Cliffs, at Butler PA), with additional supply from Nippon Steel (Japan), POSCO (Korea), thyssenkrupp (Germany), Baowu/China, and AK/others. HD Hyundai Electric is explicitly named as a supplier "constrained by GOES steel and limited global capacity". Korea link: Hyundai Steel (a related Hyundai-family entity) supplies special steels — a partial in-group hedge, but GOES specifically is globally tight.
• Copper (windings), insulating oil / bushings / tap-changers (specialised sub-components, e.g. Reinhausen for OLTCs), cooling and monitoring electronics.
• Labour — skilled transformer-winding labour is itself scarce and a gating input at the new US plant.

The company (transformation):

• Ulsan, Korea — flagship EHV transformer plant; capacity being lifted ~300 → ~360 units.
• Alabama, USA (Montgomery) — the North American EHV plant; capacity 100 → 150 units (Phase 1), then a 2nd plant broke ground Mar 2026 (+50% capacity, adding 765kV-class manufacturing + testing) — full local production ~2028.
• Additional plants in Korea (Cheonan for switchgear/rotating machinery) and a US sales/service subsidiary ("Atlanta subsidiary" referenced in FY25 results).

Downstream (buyers / distribution):

• US IOUs & grid operators (transmission upgrades), hyperscalers / data-centre developers (the July-2026 "global tech giant" framework, up to $720M / KRW 1.12T through 2028), Middle-East national projects (Saudi/GCC), European TSOs (EU sales +38% in FY25), marine/shipyards (incl. the group's own HD Korea Shipbuilding).

Single-source / chokepoint verdict: the chain is input-constrained, not demand-constrained. The scarce nodes are (a) GOES steel and (b) the engineering/testing capability to build 765kV units — both of which favour HD Hyundai Electric because it already holds slots on the former and is one of few globally with the latter. The Middle-East low-voltage-breaker delays in Q1 2026 show the distribution segment's chain is looser and more exposed to geopolitics.

Lens 3 · Competitive Advantages (moats)

The moat is real, narrow, and time-boxed to the shortage — that combination is the whole thesis.

1. Scarce EHV/765kV capability + capacity (the core moat). Only a handful of firms globally build 765kV-class transformers: Hitachi Energy, Siemens Energy, GE Vernova (Prolec), Mitsubishi Electric, Toshiba, WEG, Hyosung Heavy, LS Electric, TBEA/China XD (China) — and HD Hyundai Electric is #1 in US EHV/UHV transformer market share. This is a certification-, testing-, and reference-list-gated business (a utility will not put an uncertified 765kV unit on its grid), so incumbency compounds.
2. Switching costs / lead-time lock-in. With industry lead times at ~128 weeks and up to 4–5 years for high-capacity units, a buyer who has a slot cannot casually switch suppliers. The "Slot Reservation" system monetises this — buyers pre-pay for certainty. That is genuine pricing power, evidenced by the Q4 2025 operating margin of 27.6%.
3. Local-manufacturing / trade moat (emerging). The Alabama build converts a tariff threat into a barrier — US-made 765kV capacity that Korean/Chinese exporters can't match and that satisfies "buy-domestic" utility procurement. Full benefit ~2028.
4. Cost + input access. In-group steel (Hyundai Steel) and a low-cost Korean manufacturing base give a structural cost edge over Hitachi/Siemens on comparable units.

Bargaining power (who needs whom). Today, decisively the seller's. Utilities and hyperscalers with 18-month AI build cycles are colliding with 4-year transformer lead times — they need the slot far more than HD Hyundai Electric needs any one order (hence it can decline low-margin work). Over suppliers, power is more balanced: GOES steel is the one input where HD Hyundai Electric is a price-taker.

Moat durability caveat (critical). This is a cyclical/shortage-driven moat, not a Coca-Cola brand moat. It is wide while the shortage lasts and capacity is short. The bear case (Lens 12/13) is entirely about when the shortage normalises and whether the 765kV certification barrier is enough to hold margins once Hitachi/Siemens/GE/Chinese capacity floods in (2027–2029 waves).

Lens 4 · Segments

By product line (revenue growth, latest data) — throughout; no segment file exists:

Segment revenue split in exact KRW is not cleanly disclosed in English sources — n/a at the dollar level; power equipment is understood to be the clear majority of revenue and the overwhelming majority of incremental profit.

By geography (the more decision-relevant cut):

• North America ~47% of FY2025 revenue (up from ~40% in H1 2025) — the single most important number in the file; the company's fortunes are now a US grid-capex bet.
• Europe: sales +38.3% in FY2025, now >10% of revenue — a genuine second leg emerging.
• Middle East: meaningful (Saudi/GCC projects), but the source of the Q1-2026 distribution delays; diversifying the backlog.
• Korea (domestic): rising with the newly-announced KRW 1,500T national mega-project + HVDC "energy expressway" grid plan (see Lens 8) — a future tailwind more than a current driver.

Trend + cause: the mix is accelerating toward high-margin power equipment and toward North America, which is exactly why blended operating margin re-rated from low-teens to ~24–25% full-year (Q4'25 spot 27.6%). The risk embedded in the same fact: revenue concentration in one product line and one geography (Lens 13).

---

Phase B — Measure performance

Lens 5 · Earnings Result (latest print: Q1 2026, reported 2026-04-28)

`` unless noted.

• Revenue: KRW 1.0365T (~$0.75B ), +2.1% YoY. Deliberately soft top-line growth — the selective-ordering system is prioritising margin over volume.
• Operating profit: KRW 258.3B, +18.4% YoY — profit growing ~9x faster than revenue = the mix/pricing story in one line.
• Operating margin: 24.9% (vs ~21–22% a year earlier; vs 27.6% spot in Q4'25 — so Q1 is a step down sequentially from the record, worth watching).
• Order intake: $1.797B, +34.6% YoY — a record Q1, hitting 42.6% of the full-year $4.222B order target in one quarter. Of this, $1.315B was North America. (A separate source cites KRW 2.646T new orders — the largest quarterly total in company history; the $1.797B and KRW 2.646T are the same event in different bases.)
• Backlog: $7.888B, +17.2% vs end-2025 (and vs ~$6.55B a year earlier) — ~2x annual revenue of visible, pre-priced work.
• Segment drivers: power equipment +21.6% (NA transformers), rotating machinery +10.8% (marine), distribution −24.2% (comp + ME delays).
• Balance-sheet flags: company is in a net-cash position; no distress signals surfaced. (Detailed receivables/inventory lines not available in English web sources — n/a; verify in DART.)
• Market reaction / what was priced in: the read-across is that the market now demands not just growth but acceleration — the Q2 2025 modest miss triggered a sharp selloff, and the FY guide of KRW 4.35T once sent shares −6%. At ~31–40x, the bar is high and the stock is fragile to any shortfall.
• Unusual vs own history: a 24.9% operating margin on a transformer maker is historically extraordinary — pre-2021 this business ran low-single-digit margins. The abnormality is the thesis, and the risk.

Lens 6 · Earnings Calls (sentiment trend)

No transcripts/ on the shelf (transcripts=0) and Korean IR calls are sparsely covered in English — so this is ``-inferred from results commentary and management statements across the last ~4 quarters, not a verbatim transcript read (flagged).

• Consistent management focus (last ~4 quarters): (1) selectivity over volume — "prioritise high-margin projects, decline low-margin work" recurs every quarter; (2) North American capacity — the Alabama expansion is the strategic centrepiece; (3) 765kV as the differentiator; (4) backlog quality / duration over headline size.
• Tone shift: from "riding the AI/data-centre wave" (2024–early 2025, opportunistic) → "managing a multi-year, capacity-constrained order book and localising to defend against tariffs" (late 2025–2026, structural/confident). The framing has moved from cyclical upside to "regime change, not a cycle" — bullish, but exactly the kind of confidence that peaks near cycle tops.
• What they've started saying: tariffs (explicit acknowledgement of KRW 10–20B absorbed cost — see Lens 12), local content, 765kV testing capability, framework/master-supply agreements with Big Tech.
• What to watch for them to stop saying: "declining low-margin orders." The day selectivity ends and they start chasing volume is the day the shortage is over and the margin thesis breaks. Not yet in evidence.

Lens 7 · Comps

Peer table — HD Hyundai Electric vs global grid-equipment peers.

Honest limitation: I could not source a clean, same-date multiples grid for the peers in web-only mode, so most cells are n/a rather than guessed. This is the single biggest gap in the dossier and the first thing to fill from a terminal (Bloomberg/CapIQ/MarketScreener valuation tabs) before pricing a trade.

What is sourced on relative valuation:

• HD Hyundai Electric trades at ~31–40x trailing, ~25–28x forward FY2026 consensus, described as "roughly in line with global power-equipment peers" on forward but at a premium on trailing. So it is not obviously cheap and not obviously expensive vs peers on forward — the debate is entirely about whether the forward E is real.
• The Korean trio (HD Hyundai Electric + Hyosung Heavy + LS Electric) combined: FY2026e revenue KRW 17.96T (+19.6%), OP KRW 2.99T (+38%), closing on KRW 3T.
• Positioning vs peers: HD Hyundai Electric is the purest transformer play (~100% electrical equipment) — max torque to the upcycle, min downside cushion. Hyosung Heavy carries a construction segment (floor in a downturn, less torque). This purity is why it earns the highest multiple and the sharpest drawdowns.

Lens 8 · Stock-Price Catalysts (moves >5%, last ~5 years — pattern read)

`` throughout. The name is a momentum vehicle for the "electrification/AI-power" trade and moves on order-flow and sector sentiment more than on classic earnings surprises.

• The regime move (2021→2026): operating profit grew ~100x in four years; 3-year TSR >1000% (vs Schneider ~80%); +~146% over the trailing 365 days to ~KRW 950,000. This is a structural re-rating, not a trade.
• Up-catalysts (what the market rewards): (1) record order prints / big NA awards — e.g. the July 2026 KRW 1.12T ($720M) Big Tech data-centre framework; (2) US capacity announcements (Alabama 2nd plant, 765kV); (3) Korea mega-project headlines — power-equipment names surged on the KRW 1,500T semiconductor/AI/physical-AI investment plan; LS Electric +12% / Hyosung +8% in a single session on grid news.
• Down-catalysts (what the market punishes): (1) any earnings shortfall — the Q2 2025 modest miss produced a sharp selloff; (2) conservative guidance — the KRW 4.35T FY guide sent shares −6%; (3) tariff / trade-policy scares; (4) Korean grid-project delays (Uljin–Hanam HVDC stalled for years on local opposition), which cap the domestic leg.
• Pattern verdict: the market reacts most to order momentum and capacity/geopolitics, and is asymmetrically unforgiving of margin/guidance disappointment because the multiple prices perfection. Beta to the whole "power-shortage" basket is high — it will not decouple from Hitachi/Siemens/Hyosung on a sector wobble.

---

Phase C — Judge people & books

Lens 9 · Management

• Track record. The operating story is genuinely elite: revenue and OP up five consecutive years, OP ~100x in four years, margins from low-single-digits to ~25%, and a net-debt → net-cash transition. Whether that is management skill or riding the biggest grid-equipment upcycle in a generation is the honest question — probably both: the "selective ordering / Slot Reservation" discipline is a real, deliberate margin choice, not just tailwind.
• Leadership (note the churn). English sources show a leadership transition: Cho Seok, the CEO who ran the turnaround, was promoted to vice-chairman, with Kim Young-ki (a VP) stepping up as CEO. Group-level, Chung Ki-sun (3rd-gen founding family, grandson of Chung Ju-yung) is now HD Hyundai chairman/vice-chair driving the group. Verify the exact current CEO and appointment date in DART/IR — English sources are slightly stale and conflicting.
• Tenure & skin in the game. This is a chaebol subsidiary: the aligned owner is HD Hyundai Co. (~35.6–37.2%), controlled by the Chung family — so insiders' skin is at the holdco level, not personal exec stakes in 267260. Professional-manager CEOs run the sub; capital-return and strategy answer to the group. insider-transactions.csv absent → individual-exec ownership n/a.
• Capital allocation. Reinvestment-first and, so far, disciplined and value-accretive: KRW 400B (~$274–290M) into Ulsan+Alabama capacity, then ~$200M into the 2nd US plant — i.e. plough profits into the scarce asset (capacity) at the top of a shortage. The dividend is tiny (~0.47% yield, KRW 3,800/sh) — this is unambiguously a growth/reinvestment name, not a payout story. Reported ROE ~38.9% (FY2024, subsidiary) — if accurate, capital is being deployed at very high returns.
• Red flags (governance). The chaebol structure is the flag: (1) minority holders sit under a holdco whose interests (holdco-discount reduction, group-wide capital moves) may not align with 267260 minorities; (2) related-party flows exist by design (in-group steel from Hyundai Steel, marine equipment sold to the group's own shipbuilder) — normal for a chaebol but worth monitoring for transfer-pricing; (3) leadership churn at the top. No accounting or promotional red flags surfaced.
• Archetype. Professional-manager-run subsidiary of a family-controlled group — implies operational competence + strategic capital access, but minority-alignment risk and decisions taken through a group lens. Right archetype for a capacity build-out; watch it at capital-return time.

Lens 10 · Forensic Red Flags

Accounting-risk scan (income statement / balance sheet / cash flow). Because this is web-only with no filings on the shelf, this is a risk map to verify in DART, not a forensic finding from primary documents — stated plainly.

• Revenue recognition (highest-priority item to verify). Long-cycle transformer orders + a "pre-pay for slots" model raise the classic percentage-of-completion vs point-in-time question, plus treatment of customer pre-payments/deposits (contract liabilities). With backlog ~2x revenue and a record order quarter, how and when revenue and margin are booked across multi-year contracts materially affects reported profit. Verify POC policy + contract-liability balances in the DART financials. n/a from web.
• Cash flow vs earnings. With reported net-cash and heavy capex, watch FCF vs OP (capacity build absorbs cash) and working capital — a book that grows this fast can see receivables/inventory outrun revenue. No divergence evidenced in web sources, but also not disclosed at line-item level → explicit gap. n/a.
• Segment reporting. Product-line revenue split is disclosed only qualitatively in English (growth rates, not clean KRW splits) — limits outside forensic checks. Verify in DART.
• Related parties. In-group steel purchase (Hyundai Steel) and in-group marine sales (HD shipbuilding) = ongoing related-party lines to check for arm's-length pricing.
• SBC / non-GAAP. Korean reporting is IFRS with statutory operating profit; SBC-flattered non-GAAP is less of a risk than for US names, but confirm.
• Tariff cost treatment. KRW 10–20B tariff cost absorbed H2 2025 — confirm it's expensed, not capitalised/deferred.

Regulatory findings (required sub-section). Per regulatory/regulatory-findings.md (generated 2026-07-07):

• SEC (EDGAR LR + AAER): none possible — HD Hyundai Electric has no CIK and does not file with the SEC. total_sec_findings: 0.
• Non-SEC / web search ("HD Hyundai Electric" (FTC OR DOJ OR FDA OR CFPB OR consent decree OR settlement OR fine OR penalty)): no material enforcement actions, consent decrees, fines, or penalties surfaced against the entity in web search as of 2026-07-07 ``. (The only "tariff cost" is ordinary trade duty, not an enforcement action.)
• Item 3 / Legal Proceedings: N/A — no US filing exists; the equivalent Korean disclosure lives in the DART business report and was not on the shelf / not fetched → n/a; verify in DART before relying on a clean-legal conclusion.
• Conclusion: No material regulatory or legal findings surfaced — verified via SEC EDGAR EFTS (LR/AAER, returned zero — no CIK) and web search as of 2026-07-07. Korean DART legal-proceedings disclosure was NOT reviewed and remains an open verification item.

---

Phase D — Project & stress-test

Lens 11 · Forward Projection (EPS, next 3 fiscal years — FY2026 / FY2027 / FY2028)

Anchors (sourced):

• FY2025 (completed): revenue KRW 4.0795T, OP KRW 995.3B, OP margin ~24.4% full-year.
• Consensus next-FY EPS KRW ~27,192; consensus target KRW 1,398,235 (Buy/Strong-Buy, 19 buys / 0 sells).
• Q1 2026: rev KRW 1.0365T (+2.1%), OP KRW 258.3B (+18.4%), backlog $7.888B (~2x sales), FY2026 order target $4.222B (42.6% booked in Q1).

Base case (most likely). Backlog ~2x revenue + record orders give unusually high visibility. Assume top-line +12–18% (constrained by choice/capacity, not demand) and operating margin holding ~24–26% as NA/765kV mix offsets some tariff/steel cost:

• FY2026 OP ≈ KRW 995.3B × ~1.20 ≈ KRW ~1.19T; margin ~25% on rev ~KRW 4.7–4.8T.
• FY2027 OP ≈ KRW ~1.35T; margin ~24–25%.
• FY2028 OP ≈ KRW ~1.45–1.55T.
• Base EPS path anchors on consensus ~KRW 27,000+ (FY2026) and grows mid-teens → roughly KRW ~31–34k (FY2027) and KRW ~34–38k (FY2028). EPS is `` — the sourced consensus figure is FY2026 only.

Bull case. Shortage runs hotter/longer (lead times stay 3–4yr through 2028+), Korea's KRW 1,500T mega-project adds a domestic leg, Europe passes 15% of revenue, margins hold ≥26%: OP compounds ~18–20%/yr → FY2028 OP KRW ~1.7T+, EPS KRW ~40k+, and the multiple holds on "regime, not cycle." Upside to KRW 1.4M+ (consensus target) and beyond.

Bear case. Competitor capacity (Hitachi $1.5B, Siemens Energy $1B, GE Vernova/Prolec, Chinese TBEA/XD) lands 2027–2029 and normalises pricing; the Slot-Reservation premium compresses; tariffs bite pre-2028; a data-centre pipeline cull (30–50% of 2026 DC projects at risk of delay/cancel ) softens orders. Margin reverts toward high-teens, growth stalls → FY2028 OP flat-to-down vs FY2026, EPS KRW ~22–26k, and a multiple de-rate to 12–18x does most of the damage (a de-rate from 31x→15x on flat EPS ≈ −50% stock).

Forecast (Brier-logged intent — NOT created, per --watchlist rule): base call would be "267260.KS FY2026 operating profit ≥ KRW 1.15T," p≈0.62, resolves 2026-12-31. Logged here as intent only.

Lens 12 · Bull vs Bear

Bull case (narrative). This is the single cleanest listed instrument for the most under-supplied link in the entire AI build-out. Everyone watches the GPU; almost no one can build the 765kV transformer that lets the GPU turn on — and lead times are four years. HD Hyundai Electric is #1 in the US, runs a Slot-Reservation model that lets it set price and decline low-margin work (hence 27.6% spot margins on a metal-bending business), sits on a ~2x-revenue backlog with a record order quarter, is building US capacity that turns the tariff threat into a moat, and has a second demand leg forming in Europe (+38%) and a third in Korea's KRW 1,500T grid plan. It throws off net cash and earns ~38% ROE. The tailwind is measured in years (grid replacement + electrification + AI), not quarters — "regime change, not a cycle."

Bear case (2–3 permanent-impairment risks).

1. Cycle/supply normalisation kills the margin, not the revenue. The moat is the shortage. Hitachi ($1.5B), Siemens Energy ($1B), GE Vernova, and Chinese TBEA/XD are all adding capacity through 2027–2029. When slots stop being scarce, the Slot-Reservation premium evaporates and 25% margins revert to mid-teens — and the multiple de-rates on top of it. Double-whammy.
2. Multiple, not business, is the risk. At ~31–40x trailing, the stock prices continuation of a 100x-in-4-years regime. The Q2 2025 miss selloff and the −6% on a merely-conservative FY guide prove the market will not tolerate any wobble. Most of the downside is de-rating.
3. Concentration. ~47% of revenue is North America and the profit engine is one product line (power equipment). A US data-centre pipeline cull (30–50% at delay/cancel risk) or a US trade/tariff shift hits the exact centre of gravity.

Pre-mortem (18 months out, thesis broke — what happened?). Most likely story: 2027 capacity waves + a data-centre digestion pause turned "sold-out through 2028" into "order growth stalled," a quarter missed on margin as steel/tariff costs bit before Alabama fully ramped, management quietly started chasing volume (stopped "declining low-margin work"), and a 31x multiple compressed to ~16x — the stock halved even though revenue barely fell. Secondary story: a tariff/trade shock landed before 2028 local production, compressing the ~40% NA book.

Are multiples too high? On trailing, yes (priced for perfection). On forward (~25–28x), only if you believe the forward E — which depends entirely on the shortage lasting. This is a duration-of-shortage bet dressed as a growth stock.

Contrarian view (what the market is refusing to see). Bulls treat the ~25% margin as the new normal ("regime change"); the contrarian read is that peak margin ≈ peak shortage, and the very capacity investments the whole industry is celebrating (including HD Hyundai Electric's own Alabama build) are sowing the 2028–2029 normalisation. The most valuable question is not "how big is the backlog" but "what is the through-cycle margin once everyone's new plants are running" — and nobody is modelling that yet.

Lens 13 · Devil's Advocate (short-seller)

Dismantling the bull case:

• Where revenue is concentrated / what breaks it: ~47% North America, engine = one product line (power equipment), demand = AI-data-centre + US grid capex. If the DC pipeline is even 30% overstated (and 30–50% of 2026 DC projects are flagged for delay/cancel ), the incremental order growth — which is what the multiple pays for — stalls fast.
• Why the moat is weaker than bulls think: the 765kV certification barrier is real but not unique — Hitachi, Siemens, GE Vernova/Prolec, Mitsubishi, Toshiba, Hyosung, LS, and Chinese TBEA/XD all play here, and all are adding capacity. Chinese makers offer "significantly shorter delivery and competitive pricing" and will take the price-sensitive tier. The moat is a temporary capacity gap, not a structural franchise.
• Most dangerous competitor bulls underestimate: Chinese OEMs (TBEA, China XD) on price + Hitachi Energy / GE Vernova on US-local capacity + certified relationships. GE Vernova already has US transformer manufacturing and just won Korea's own HVDC backbone with KEPCO — i.e. foreign majors are winning on HD Hyundai Electric's home turf too.
• Worst capital-allocation / governance angle: chaebol-subsidiary structure — minorities under a ~36% holdco with related-party steel-in / marine-out flows and a holdco-discount-reduction agenda that can prioritise group value over 267260 minorities. Tiny dividend = minorities get re-rating or nothing.
• Assumptions that must hold for today's price: (1) shortage/lead-times stay 3–4yr through ≥2028; (2) margins hold ~25% despite the whole industry adding capacity; (3) NA orders keep compounding and the DC pipeline is real; (4) tariffs don't bite the NA book before 2028; (5) the multiple doesn't de-rate. That's five things that all have to go right.
• If growth disappoints 20–30%: in a premium-multiple name this is reflexive — a growth miss triggers a de-rate, so a ~25% earnings shortfall plausibly maps to a 40–55% drawdown (EPS −25% × multiple 31x→~18x). The Q2 2025 selloff is the template.
• Single scenario that permanently impairs: a structural US localisation/tariff regime that favours domestic-owned (Hitachi/GE) or Chinese-priced supply while global capacity catches up — compressing both the ~47% NA revenue share and the premium margin at the same time. Plausibility: moderate — the Alabama build is precisely the hedge, but it doesn't fully protect until ~2028, leaving a 1–2 year exposure window.

Lens 14 · Management Questions (15, ordered by information value)

1. Through-cycle margin: once your Alabama Phase 2, Hitachi's $1.5B, Siemens' $1B, GE Vernova, and Chinese capacity are all running (2028–2029), what operating margin do you underwrite through the cycle — and how far below today's ~25% could it go?
2. When do you expect transformer lead times to normalise below ~24 months, and what happens to your Slot-Reservation premium on that day?
3. What share of the $7.9B backlog is firm/take-or-pay with penalties vs framework/indicative — and what are the cancellation/deferral terms on the Big-Tech data-centre master agreements?
4. How is revenue and margin recognised across multi-year contracts (POC vs point-in-time), and what is the current contract-liability (customer pre-payment) balance?
5. What is your GOES steel cost and availability outlook, and how much is contractually locked vs spot-exposed over the next 3 years?
6. Quantify remaining tariff exposure: how much of the ~47% NA revenue is still imported, and what is the P&L bridge as Alabama localises through 2028?
7. If the US data-centre pipeline is culled 30–50% as some analysts forecast, what is the demand elasticity of your NA order book?
8. What are your through-2028 capacity utilisation assumptions, and at what utilisation does incremental capacity become margin-dilutive?
9. Capital return: with net cash and ~38% ROE but a ~0.47% yield, what is the buyback/dividend framework — and how do you weigh minority returns against the holdco's agenda?
10. Which related-party flows (Hyundai Steel inputs, in-group marine sales) are material, and how is arm's-length pricing assured?
11. What is your China strategy — do you cede the price-sensitive tier, or defend it, and where?
12. How defensible is 765kV specifically — how many quarters of certification/testing lead do you hold over the #2–#4 competitors?
13. Europe is now >10% of revenue (+38%): is this a durable second leg or project-lumpy, and what's the 3-year target mix?
14. On the leadership transition (Cho Seok → new CEO): what changes vs stays in the selective-ordering strategy under new management?
15. What is the single early-warning metric you watch that would tell you the shortage is turning before it shows up in orders?

---