179290 — M.I.TECH Co., Ltd

Deep-value analysis — M.I.TECH Co., Ltd (179290, KOSDAQ)

Date: 09/14/2026 · Price: … (tracker, 09/14/2026; the 16:07 data pack gives …) Shares: 32,365,678 (note 11, H1 2026 semiannual report) · Market cap: … Single currency: reporting in KRW, trading in KRW — no conversion, no ADR, no ADS/ordinary reconciliation Reason for re-analysis: the deep_delta.py triage decided FULL — the fcf_yield < 0.03 falsifier broke

Primary sources used:

• DART/FSS, semiannual report 반기보고서 (2026.06), filed 08/14/2026, no. 20260814002272 — consolidated statements + notes 1-22
• DART/FSS, annual report 사업보고서 (2025.12), filed 03/18/2026, no. 20260318001578 — FY2025/FY2024/FY2023 statements, receivables aging, audit opinion
• DART/FSS fnlttSinglAcntAll (structured XBRL) for FY2025 with comparative FY2024/FY2023 columns and for Q1 2026
• yfinance (via deep_data_pack.py, 09/14/2026) for the quarterly series and market price
• stockanalysis.com (flows and multiples, consulted 09/14/2026) — external check only
• research brief research-179290-20260914.md (09/14/2026) — market context and consensus

Executive summary

M.I.TECH makes non-vascular stents (gastrointestinal, biliary) under the HANAROSTENT brand, 82.2% of revenue, plus distribution of Fujifilm endoscopes in Korea (8.9%) and accessories (8.9%). 185 employees generate … of annual revenue with a 29… operating margin and a ROIC on strictly operating capital of 35…. It’s a real business, small, high-margin, with a technical position hard to replicate in six months. The stock trades at P/E 7.8, P/B 1.22 and — after subtracting net financial assets of … — at EV/EBIT of 3.76x on TTM operating profit of 20,660 mil. On paper, it’s among the cheapest things in the universe.

The thesis from 08/23/2026 said exactly this and gave a BUY. In the meantime the semiannual report was published, and it breaks one of the thesis’s six claims and cracks a second. Free flow disappeared: Q2 2026 CFO was negative (−1,238 mil) on net income of 5,828 mil; over the half, CFO was 827.5 mil versus 3,303.6 mil in H1 2025 (…). The cause isn’t profit — operating profit was flat (10,715.7 vs 10,729.8, …) — it’s working capital, which absorbed … in six months. Gross receivables rose from 12,659 to 20,301 mil (…) on revenue growing only 10.8%; DSO jumped from 62 to 98 days, and in Q2 alone 32.8% of the quarter’s revenue remained uncollected. Inventory rose 26.1%. The cash conversion cycle lengthened from 166 to 200 days.

An important correction before any conclusion: the figure that triggered the triage is wrong. The data pack reports “CFO trailing 4Q 6.68 bn” and “simple FCF −1.00 bn · yield ….” The yfinance four-quarter window doesn’t contain Q4 2025 (the quarterly series has 2025-03, 2025-06, 2025-09, 2026-03, 2026-06, but not 2025-12), so it skips exactly the best quarter. Reconstructed from filings, CFO TTM at June 2026 is 12,398 mil — a figure independently confirmed by stockanalysis.com (12,398 mil). Simple FCF TTM is 5,301 mil (yield 3.28%), FCFE 3,821 mil (2.36%). The fcf_yield < 3% falsifier still triggers, but at the edge of the threshold, not at …. The deterioration is real; the magnitude in the triage is not.

The second problem is older and hasn’t improved: 52% of the share price is cash and financial assets (…, …/share), and that money is managed as the controlling group’s venture fund. 13,085 mil is deposited in partnerships managed by the controlling shareholder’s GP — in two of them M.I.TECH puts up 52.58% and 97.84% of the capital, but classifies them as “associates” because the final decision belongs to the GP. On 07/27/2026 the board approved another 10,021 mil for 33.7% of a French software startup (PrediSurge). The shareholder return rate over the last four years (dividends + buybacks ÷ cumulative net income) is …%. This is where λ = 0.22 in the simulation comes from: the money exists, but historically the minority holder sees a fifth of it.

Estimated value: range 3,580–…, Monte Carlo median … (20,000 scenarios, λ = 0.22 on net cash). Median MOS ** …**, probability of undervaluation …, probability of a margin above 30% only …%. The five triangulated models range from … (bear DCF) to … (historical multiples).

Verdict: HOLD, small position (≤1% of portfolio), no adding at …. Downgrade from BUY on 08/23/2026. Not because the business broke — gross margin is actually better (58.4% vs 57.3%) — but because the two things the thesis depended on moved in the wrong direction simultaneously: free flow disappeared into working capital behaving like distributor financing, and the governance discount is confirmed active (PrediSurge, new fund contributions) instead of closing. At EV/EBIT 3.8x the business deserves to be held; at 55% probability of undervaluation it doesn’t deserve to be added to.