98070 — Hantech Co., Ltd.

Deep-value: 한텍 / Hantech Co., Ltd. (KOSDAQ 098070)

Analysis date: August 24, 2026 · Reference price: … (close 08.24.2026) Market cap: … (~…) · Shares: 11,121,141 · Flow currency = price currency = KRW Exchange: KOSDAQ (South Korea) · Listed: March 20, 2025 · Controlling shareholder: 후성 / Foosung Corp — 69.41%

Source note, read before the numbers. Hantech isn’t SEC-registered and has no CIK: primary filings are with DART (dart.fss.or.kr), in Korean, and the OpenDART API requires a key this session doesn’t have. The consolidated K-IFRS statements used here come via the Refinitiv/Yahoo feed (098070.KQ), NOT directly from the semi-annual report. I treated this as a risk and cross-checked every aggregate against Korean press and data portals. The result, detailed in the “Red flags” chapter: four exact matches, to the decimal — H1 2025 (83.1 bn revenue / 20.3 bn operating profit), fiscal 2025 (38.18 bn OP / 30.77 bn net profit), 2024 revenue derived from the Shintech contract’s 18.21% disclosure (155.9 bn vs. 155.881 bn from the statements), and TTM EPS …. Balance sheet and cash flow figures couldn’t be independently verified and are flagged as such where it matters.

I corrected two errors in the research brief (detail in “Red flags”): the two major contracts’ values were inflated 10x (억원 read as billions), and the “unreconciled” EPS discrepancy between sources wasn’t a discrepancy, just two different TTM windows.

Executive summary

The thesis, in one sentence: Hantech is a 50-year-old heavy metal fabrication workshop in Ulsan that went through a real and verifiable margin expansion (gross margin from 13.4% in 2022 to 30.4% in 2025), has extinguished practically all its debt, sits on net cash of … — 21% of market cap — and trades at 5.9x TTM EV/EBIT, because the market decided in five months that expansion was a cycle peak. The entire investment decision boils down to one question, and the rest of this report is an attempt to answer it: how much of the 19.8 TTM operating-margin points is structural, and how much is the LNG cycle’s peak?

What the company does. Two operating divisions plus a new one. The chemical equipment division (화공기기사업부) makes heat exchangers, pressure vessels, reactors, and distillation columns for refineries, petrochemicals, LNG, and fertilizers, with a niche competency in special materials — titanium and zirconium. The tank division (탱크사업부) designs and field-builds cryogenic tanks for nitrogen, oxygen, hydrogen, ammonia, and CO2. A green-energy division (그린에너지사업부) was added in 2025-2026: pressure vessels for CCUS, CASK containers, and hydrogen tanks. Customers are EPC firms: Samsung E&A, POSCO E&C, GS E&C domestically, Bechtel, KBR, and JGC globally (Company Guide / fnguide, aggregated 2026-08-24).

What happened to the price. Listed at … in March 2025, the stock rose to … intraday on 03/23/2026 (… versus the offer price), on the wave of the revived LNG Alaska project, then fell to … on 07/30/2026. From the peak: …. Today, …. It isn’t a company that broke — it’s a company whose multiple deflated once the thematic narrative cooled.

What the results actually say. The research brief read Q2 2026 as marked deterioration (revenue …, operating profit …, net profit … versus last year) — the figures are correct, the interpretation is wrong. The comparison base, Q2 2025, is the most abnormal quarter in the company’s visible history: gross margin 42.0% and operating margin 36.8%, versus 22… in the other quarters. On a half-year basis, H1 2026 shows revenue … (93.3 vs. 83.1 bn) and net profit … (16.94 vs. 15.66 bn); only operating profit falls 10.7%, exclusively because of that anomaly quarter. Q2 2026, with a 22.7% operating margin, is the second-best quarter in the series. The deterioration is a base-effect illusion.

The balance sheet. The cleanest element of the thesis. Total debt: 60.9 bn (2022) → 56.9 → 49.9 → 13.7 (2025) → 2.1 bn at 03/31/2026. Cash plus short-term investments: 61.6 bn. Net cash 55.4 bn after the declared dividend of 4.17 bn — mechanically verified by mc_dcf.py from the balance sheet at −…, a 0.14% deviation from my figure. Equity 166.3 bn, total assets 227.2 bn. No dilution post-IPO: 11,121,141 shares, constant across all five reported quarters.

Estimated value. Five models, range 17,900 – …, median margin of safety …. Monte Carlo over 20,000 scenarios gives a median intrinsic value of … (MOS …), with P10 at … and P90 at …. Probability of undervaluation: …%. Terminal value weight in the base case: 45% — below the 60% fragility threshold.

Verdict: PARTIAL BUY, small position (1.5… of portfolio), with adding conditional on a single figure. The bull case: you’re paying 5.9x EV/EBIT and 8.4x profit for a business with 22% ROE, no debt, 21% of the price in cash, where the controlling shareholder bought shares on the open market in May. The bear case, equally serious: it’s a cyclical, tender-based subcontractor with margins that were 6… for three years and 20… for two, and if 6… is the normal, the stock is worth …, not 33,000. The two conservative models in the triangulation — DCF bear and EPV Greenwald — say exactly that, both around …. The range is too wide for a full position. The figure that decides: gross margin in Q3 2026 (reporting ~mid-November): above 25% confirms the structure and justifies doubling the position; below 20% invalidates the thesis and means exit.