Export Upcycle Value Sector IC Memo

Executive Summary & Action Plan

Verdict

OVERWEIGHT (selective) — the four-way intersection of improving exports, improving earnings, relative cheapness, and a credible dream currently lives in Korea. Ten screened candidates went through independent adversarial verification; only the four Korean names passed all four criteria outright, while the three U.S. survivors passed on exports, earnings, and dream but hold their valuation leg only conditionally. Own the basket with a Korea tilt and half-size the U.S. sleeve.

Loading chart

Abstract

This screen asked for companies with improving exports, improving earnings, relative undervaluation, and a dream, across Korea and the U.S. Korea's export tape is historic — July exports +62.8% YoY and August 1–20 +56.0%, with semiconductors +198.8%, cosmetics on a nine-month streak, and bio-health at a record half — and it feeds directly into four verified mid-caps: Silicon2 (257720.KQ), ST Pharm (237690.KQ), Binggrae (005180.KS), and HD Hyundai Marine Engine (071970.KS), each with Q2'26 revenue or operating income accelerating and primary multiples below peers or their own history. The U.S. export winners (LNG, aerospace, nitrogen) mostly failed the cheapness test after re-rating; Embraer (EMBJ), Venture Global (VG), and CF Industries (CF) survive as Tier B with documented valuation caveats. Three names were removed by verification for stale data, misclassified China exposure, and a looming earnings air pocket. The frozen 7-name basket carries a probability-weighted +16.3% 12-month path; the main falsifier is the September 1 export release.

Forecast Path

The scoreable target is the frozen equal-weight basket of the seven listed names (local-currency price returns, entry = 100). Base assumes August exports confirm momentum and at least three of four Korean Q3 prints deliver; Bear combines an export-momentum break with the USTR port-fee reimposition and a VG legal setback; Bull requires the Korean rally to broaden beyond semiconductors while the U.S. sleeve clears its catalysts. Probabilities vary by horizon in the prediction sidecar; the chart shows the 12M terminal vector.

Loading chart

Action Plan

Overweight the intersection with a 60/40 Korea/U.S. split: full-size the four Tier A Korean names, half-size the three Tier B U.S. names until each clears its named valuation catalyst.
Stage entries in thirds — now, after the September 1 export release, and after the late-October-to-mid-November Q3 prints — rather than chasing the Korean names that have already re-rated (Silicon2 +16% in six sessions; Binggrae at a 52-week high).
Prefer the drawdown entries where verification found intact fundamentals: HD Hyundai Marine Engine (071970.KS) −54% from its April high and ST Pharm (237690.KQ) −31% from February, both with unchanged-or-raised broker estimates.
Review triggers: USTR port-fee decision 2026-11-09 (HD Marine Engine), VG's BP damages ruling and Plaquemines COD, urea holding above ~$330/t pre-war levels (CF), and E2 delivery cadence into the Embraer Q3 print.
Re-underwrite immediately if the basket hits the 1M Bear level (90) or if August exports break the momentum thresholds below.
Prob-weighted 12M basket return +16.3% | P(thesis wrong) 35% | Confidence: Medium

State of the Sector

Korea is running the strongest export tape in its history: total exports crossed $100B in a month for the first time in June (+70.9%), July held at $98.9B (+62.8%, second-highest ever), and August 1–20 ran +56.0% with semiconductors at 47.2% of all exports — a record share (MOTIE 2026-08-01; Korea Customs 2026-08-21). But the tape is two-track. Semiconductors (+178.8% in July), AI-server SSDs (computers +404.0%), cosmetics (+37.8%, ninth straight month), bio-health (+30.4%), and power equipment are accelerating; autos, steel, petrochemicals (volume −9.1% despite value +10.3%), and displays are flat. The screen therefore hunted the accelerating categories' second-derivative names rather than the crowded leaders.

The U.S. mirror image is real but differently priced. The same AI-memory cycle that powers Korean chip exports shows up as record quarters at the U.S. equipment makers; defense backlogs, LNG cargoes, gas-turbine orders, and Boeing's best first-half deliveries since 2018 all confirm a U.S. export upcycle (BEA: YTD exports +11.7%, 2026-08-04). The difference is valuation: verification found most U.S. export winners already re-rated — which is why the U.S. side of this list is thinner and carries caveats, and agriculture (Deere's −15–20% North American large-ag outlook) is the one cleanly deteriorating export theme.

Forecast Drivers & Cycle Map

The table maps each basket sleeve to the export data that feeds it, the next confirmation, and how the thesis would fail.

Sleeve	Export evidence (as of)	Next confirmation	Principal failure mode
K-beauty distribution — Silicon2	Cosmetics +37.8% Jul, 9-mo streak; company EU +62%, N.America +79% YoY (Q2'26)	Sep 1 exports; Q3 print (~Nov)	U.S. 15% tariff bites; brand disintermediation
Oligo CDMO — ST Pharm	Bio-health +30.4% Jul; record $353.9M backlog (Jul)	2nd oligo plant onstream Q4'26; Q3 print	Commercial ramp (RYTELO) undershoots; Section 232 pharma tariff
K-food — Binggrae	Ice-cream exports record 2025, Jan–Apr +14.1%; US unit +30.4% H1	Q3 (peak-summer) print ~mid-Nov	Weather-normalized 2027 comps; peak-multiple trap
Ship engines — HD Hyundai Marine Engine	Ship exports +46.9% Jul; H1 orders ₩672.7B > FY25 revenue	Q3 print late Oct; USTR 11/9	China-yard exposure (12/15 contracts) vs port fees
U.S. LNG — Venture Global	127 cargoes +42% YoY Q2; 91% of 2026 contracted	Plaquemines COD; BP damages	$42.4B debt + arbitration + 2027–28 LNG supply wave
Nitrogen — CF Industries	~5Mt Mideast supply offline; India import pull	Fall application season; Q3 (~Nov)	Urea back below pre-war $330/t = mid-cycle at median multiple
Aerospace — Embraer	Backlog $34.5B record; C-390 7+ European customers	Dubai Airshow Nov; Q3 print	H2 E2 cadence miss at a re-rated multiple

Selection & Dispersion

Prices below are verified closes from the authoritative quote source (KRX 2026-08-26; U.S. 2026-08-25). Scores are the adversarial verifier's 0–10 marks on the four criteria (Exports / Earnings / Valuation / Dream); every claimed number was independently re-searched, which is also how three original candidates were removed. No fair values were built for this screen — cheapness is argued on primary multiples (EV/EBITDA, EV/FCF, FCF yield, PEG; P/S only for pre-profit), never on forward P/E.

Tier	Company	Close (verified)	Latest quarter (calendar Q2'26)	Primary valuation evidence	The dream	E/E/V/D
A · Top-Pick	HD Hyundai Marine Engine (071970.KS)	₩52,900 · ₩1.79T	Rev ₩128.1B +29.1%, OPM 24.4% (+6.9%p)	2026E EV/EBITDA ~8x vs engine peers 15–22x; net cash ~₩364B; PEG deep sub-1	Engine shortage + China-yard penetration: H1 orders ₩672.7B beat FY25 full-year revenue	8/8/8/8
A	Silicon2 (257720.KQ)	₩51,800 · ₩3.40T	Rev ₩402.6B +52%, OP ₩83.0B +59%, OPM 20.6%	2026E EV/EBITDA ~11x vs own 15–16x history; PEG ~0.5	K-beauty's global logistics/retail layer: Douglas access, own "moida" stores, MOTIE cites offline expansion it enables	9/9/7/8
A	ST Pharm (237690.KQ)	₩110,000 · ₩2.29T	Rev ₩108.5B +58.9%, OP ₩18.3B +41.7% (one-off cost drag)	Fwd EV/EBITDA ~17–20x vs own 30–50x history and 33–35x broker target multiples; PEG <1	Top-3 global oligo CDMO; record $353.9M backlog; capacity 3x from Q4'26	8/7/7/8
A	Binggrae (005180.KS)	₩94,300 · ₩834B	Rev ₩442.1B +7.9%, OP ₩69.7B +159.9%	EV/EBITDA 2025A 4.5x / 2026E 2.8x vs peer Lotte Wellfood 7.1x/5.3x; net cash ~25% of mcap	Melona at ~70% of Korean ice-cream share in the U.S.; nationwide Costco; Australia launch	8/9/7/7
B	Embraer (EMBJ)	$74.63 · $13.3B	Rev $2.24B +23% (record Q2), adj EPS +102% vs consensus	EV/FCF 14x, FCF yield 7.6% TTM, PEG <1 — but EV/EBITDA 15.4x ≈ Airbus, above own history	C-390 becoming NATO's C-130 successor; $34.5B record backlog	8/8.5/6/7.5
B	Venture Global (VG)	$14.07 · ~$35.2B	Rev $4.58B +48%, adj EBITDA $2.5B record, FY guide raised	Fwd EV/EBITDA ~8x vs Cheniere 10.5–13.8x — but FCF deeply negative, EV half debt	Contracted LNG staircase: Plaquemines → CP2 (H2'27) → CP2 Phase 2 → CP3	9/8/6/8
B	CF Industries (CF)	$127.26 · $19.3B	Adj EBITDA $1.19B +56% (EPS/rev missed consensus)	EV/EBITDA 5.3x LTM, FCF yield 9.9% (~8.8% even on mid-cycle FCF) — but ~7x on mid-cycle EBITDA ≈ own median	Blue Point: world's largest low-carbon ammonia plant (FID'd, 2029), Japanese offtake	7/7/6/8

Tier B is included deliberately, not carelessly: each passed exports, earnings, and dream decisively, and each remains cheap on at least one primary cash-flow metric — but the verifier could construct an honest argument that the discount is cyclical (CF), partially spent (EMBJ), or a leverage artifact (VG). Size accordingly.

Top Pick — HD Hyundai Marine Engine (071970.KS): the only name scoring ≥8 on all four criteria. The market sold it −54% from April on sector flows while H1 orders exceeded last year's full revenue, margins reached 24.4%, and brokers held targets of ₩100,000–144,000 (Sangsangin 2026-07-15 "extreme undervaluation"; Daol 2026-07-21). Its one concentrated risk — Chinese-yard orders vs the 2026-11-09 USTR decision — is dated, watchable, and priced.

What to Watch Now

2026-09-01 — MOTIE August full-month exports: the single highest-information release for this basket. Thresholds: total ≥+30% and semis ≥+100% keep Base; total <+10% or semis <+50% starts the Bear clock.
Late Oct–mid Nov (est.) — Q3'26 prints for all four Korean names: the seasonally decisive quarter for Binggrae (summer) and Silicon2 (peak season; a broker flags ₩1T quarterly revenue in reach).
2026-11-09 — USTR decision on suspended China-built-vessel port fees: the direct swing factor for HD Hyundai Marine Engine's Chinese-yard order book.
H2'26 — Venture Global Plaquemines COD declaration and the BP ICC damages ruling; CF's fall ammonia season and India urea tenders; Embraer's Dubai Airshow (C-390 pipeline: Lithuania, Slovakia, Greece).

What to Watch Next (Structure)

The 2027 base-effect cliff. Semiconductor exports cannot lap +178% forever; from early 2027 the Korean export complex faces triple-digit comps. The screen's exit discipline matters more than its entry: this is a 6–12M momentum-plus-value trade, not a buy-and-forget list.
Disintermediation vs the platform. Silicon2's dream survives only while independent K-beauty brands need a consolidated overseas layer; Douglas/Amazon direct deals are the structural counter-case.
Bio-manufacturing re-shoring. U.S. biosecure-act politics and the 15% Section 232 pharma tariff cut both ways for ST Pharm — penalty on Korean API pricing, but a moat against Chinese CDMO rivals.
LNG supply wave 2027–28. VG's equity is a leveraged bet that contracted fees outrun a well-telegraphed global supply expansion; watch charter/spot spreads, not just cargo counts.

Allocation & Selection Notes

Verification removed three candidates the initial screens liked, and the removals are part of the product. Dongsung FineTec (033500.KQ) fell when its headline "Q2 +40%" turned out to be the year-ago quarter — the real story is two quarters of revenue misses, a 2025 accounting sanction (audit-designation, trading halt), and a cheap multiple that may be pricing distrust, not opportunity. ACM Research (ACMR) posted a genuine blowout quarter but earns 99.6% of revenue in mainland China with Entity-Listed subsidiaries — it is an import-substitution play against this screen's export frame, and its cash-flow metrics fail the primary tests. Golar LNG (GLNG) has a $17B contracted backlog but enters a confirmed four-quarter EBITDA air pocket (Hilli off-hire to H2'27) at 21x trailing EV/EBITDA — the acceleration is 2027's story, not this screen's. Watchlist for the next refresh: Golar on charter #4, Cosmax (192820.KS) and Samyang Foods (003230.KS) on any pullback, Elbit Systems (ESLT) and Moog (MOG.A) if defense multiples compress, Excelerate Energy (EE) on FCF inflection.

Catalysts, Leading Indicators & Risks

Settlement thresholds for every catalyst and indicator are frozen in the prediction sidecar; the table shows the decision-relevant pairs.

Catalyst	Timing	Impact	·	Risk	Prob	Impact
MOTIE August exports	9/1	Confirms/breaks the export leg for all four KR names	·	KR export momentum stalls (semis <+50%)	Low-Medium	Basket-wide de-rating of the core thesis
KR Q3'26 prints (4 names)	late Oct–mid Nov (E)	Converts export data into EPS; entry point #3	·	≥2 of 4 KR names miss revenue	Medium	F2 fires; earnings leg failed
USTR China-vessel port fees	11/9	Binary for HD Marine Engine's China orders	·	Fees reimposed at original scope	Medium	HD Marine Engine growth channel impaired
VG Plaquemines COD + BP damages	H2'26	Resets fee economics; prices legal overhang	·	Aggregate damages ≥$2B on levered balance sheet	Medium	VG equity stub compresses hard
CF fall season / India tenders	Sep–Nov	Tests whether nitrogen holds above mid-cycle	·	Urea sustained <$330/t	Medium	CF's LTM cheapness evaporates
Embraer Dubai Airshow + Q3	Nov	C-390 orders extend the defense dream	·	H2 E2 delivery cadence miss	Medium	Guidance cut at a re-rated multiple

Falsification & Review Rules

F1 — Export leg: two consecutive months with total exports <+10% YoY and semis <+50% → cut stance to Neutral, halve the Korean sleeve, re-underwrite on standalone earnings.
F2 — Earnings leg: ≥2 of 4 Korean names miss Q3 revenue consensus (by 2026-11-30) → basket-level earnings leg failed; rebuild probabilities before the 6M horizon.
F3 — Policy: USTR reimposes port fees at scope on 11/9 → HD Marine Engine leg to Bear; +5pts to 3M/6M bear weight.
F4 — Legal: VG aggregate damages ≥$2B → remove VG from recommended exposure at the next refresh (frozen basket stays for scoring) and disclose the miss.
F5 — Price: basket ≤90 at the 1M mark (2026-09-26) → stance to Bear, halt adds, rebuild from September data.
Scheduled reviews: after 9/1 exports, after the Q3 season, and at each horizon date. The basket itself is frozen — names are never swapped retroactively.

Export Upcycle Value Screen — Rising Exports, Rising Earnings, Still Cheap, With a Dream (Export Upcycle Value Sector)