Financebotresearch desk研究台

KT › analyze

KT · Analyze

WATCH Infrastructure

2026-08-11 · $18.55 ADR · Mkt cap $8.83B · Communication Services / Telecom (South Korea)

ADR structure: 1 ADR = 0.5 common share. Financials report in KRW; the ADR trades in USD. Every per-share figure below states its unit explicitly. See §0.

Verdict up front

WATCH — conviction 5.5 / 10.

This is the most straightforward of the seven names and the one with the least interesting business. KT is a low-growth Korean telecom trading at 0.73× book with a board-approved, legally-reinforced commitment to return 50% of adjusted standalone net income for 2026–2028. The re-rating catalyst is real, verifiable, dated, and already partly in the price. What is not in the price yet is how ugly 2026's reported numbers will look, because 2024 was artificially depressed by a ~KRW 1tn restructuring charge and 2025 was artificially inflated by ~KRW 500bn of non-repeating property income. Buy the yield on the comp-driven weakness, not before it.

0. Data integrity — read this before any number

Three traps fire on this name, all recorded in the playbook:

  1. pitfall-adr-book-value-corrupts-price-to-book — here the vendor gets it right. Equity KRW 17.61tn ÷ 241.1M common = KRW 73,040/common → KRW 36,520/ADR ≈ $26.5, matching the reported BVPS of $25.54. P/B 0.73 is trustworthy.
  2. pitfall-vendor-ev-inverts-net-cash / mixed-currency feeds — the snapshot reports EV of "9.75T", which is KRW, alongside a USD market cap of $8.83B, and P/S of 0.00. These fields are unusable. EV/EBITDA of 1.58 is a currency-mixing artefact, not a valuation. Ignore all four.
  3. The base-year trap (new, and the most important). Neither 2024 nor 2025 is a clean comparison base. Detail in §1.

1. Fundamentals — and the base-year problem

All figures KRW unless noted. USD conversion at ~1,380 KRW/USD.

KRW bn FY22 FY23 FY24 FY25
Revenue 25,640 26,290 26,380 28,250
Gross profit 15,900 16,720 17,020 18,100
Operating income 1,790 1,690 874 2,590
Net income 1,260 993 460 1,720
EPS (dil, KRW/common) 2,602 2,019 953 3,557
OCF 3,600 5,500 5,070 4,940
Capex (3,990) (4,170) (3,360) (4,040)
FCF (390) 1,330 1,700 901
Buybacks 0 (300) (27) (250)
Dividends (477) (527) (872) (578)

The base-year problem, stated plainly.

  • 2024 was artificially LOW. Operating profit fell 50.9% to KRW 809.5bn on a one-off labour restructuring of roughly KRW 1.0tn — 4,400 staff (2,700 retired, 1,700 transferred at reduced wages), severance up to KRW 430m per head. Ex-charge, 2024 operating profit would have been ~KRW 1.81tn (+9.8%).
  • 2025 was artificially HIGH. FY25 operating profit +205% to KRW 2,469bn — but roughly KRW 1tn of real-estate development revenue and ~KRW 500bn of operating profit from the Gangbuk HQ redevelopment landed across Q1–Q3 2025 and does not repeat.

Normalised core operating profit is roughly KRW 1.9–2.0tn, not 2.47tn. Normalised net income is roughly KRW 1.2–1.3tn, not 1.72tn. Every valuation below uses the normalised figure. This is a live instance of Knowledge/Playbook/pitfall-divested-segment-corrupts-multiyear-cagr in its one-off variant — a non-recurring item corrupting both the base and the growth rate in opposite directions.

CAGRs (3yr, distorted by the above): Revenue +3.3% · Net income +10.9% (meaningless — measures 2022 against an inflated 2025) · OCF +11.2% · Shares +0.1%. Treat revenue's +3.3% as the only honest growth number here, and even that is flattered by property.

Free cash flow. FY25 FCF KRW 901bn ≈ $653M on an $8.83B market cap = 7.4% FCF yield. Capex is heavy and structural (KRW 4.04tn, 14% of revenue) — this is a telecom, the network is the business, and the capex never stops. FCF is genuinely volatile (−390 → 1,330 → 1,700 → 901), which is normal for the sector but limits how much weight the FCF yield can carry.

Balance sheet. Assets KRW 43.05tn, debt KRW 10.79tn, cash KRW 3.51tn → net debt ~KRW 7.28tn ≈ $5.3B. Debt/assets 25.1%, D/E 63%. Interest expense KRW 366bn against normalised operating profit ~KRW 1.95tn = 5.3× coverage. Unremarkable and safe for a regulated carrier. Debt issued and repaid are near-identical each year (5.75tn / 5.74tn in FY25) — this is commercial-paper roll, not accumulating leverage.

Per-share (FY25, per ADR): Revenue ~$21.0 · FCF ~$1.35 → 13.7× P/FCF · BVPS $25.54 → P/B 0.73×

Capital allocation. Dividends KRW 578bn + buybacks KRW 250bn = KRW 828bn against FCF KRW 901bn = 92% of FCF returned. High but covered. Capex takes precedence and always will. M&A is trivial (KRW 21bn). Goodwill is shrinking (708 → 209bn over four years) — write-downs of past acquisitions, a quiet admission that the diversification of the 2010s did not work.

Scorecard: balance sheet B+ · cash conversion C+ (volatile) · capital return A− · growth D · earnings quality C− (two consecutive distorted base years).

2. The Value-Up story — the actual thesis

This is why the name exists on a screen, so verify it concretely.

KT's commitment (board-approved, filed on 6-K):

Element Detail
Payout policy 50% of adjusted standalone net income, FY2026–2028, via cash dividend + treasury buyback-with-cancellation (excludes non-cash/non-recurring items)
2026 dividend Minimum DPS KRW 2,400 per common (Q1 KRW 600 already paid)
2026 buyback KRW 250bn — 4,215,851 shares, board-approved 2026-02-10, trust executed 2026-03-10 → 2026-09-09, then cancelled
Cancellation target KRW 1tn of treasury shares cancelled by 2028
ROE target 9–10% by 2028 (from ~6%)
AI/IT mix target 6% → >19% of sales by 2028

Korean policy backdrop, 2026. The Commercial Act was amended so director fiduciary duty runs to all shareholders, and mandatory treasury-share cancellation was introduced. The National Assembly (Dec 2025) cut dividend-income tax to 14–30% from 45%, with the ≤30% rate conditional on a >40% payout ratio — KT's 50% policy qualifies. An Enforcement Decree revision passed cabinet 2026-02-24 requiring high-dividend firms to disclose value-up plans. The Korea Value-Up Index is +130% since Sept 2024 and the KOSPI crossed 5,500 in February 2026.

Assessment: this is the most credible catalyst in today's batch of seven — dated, board-approved, tax-advantaged, and legally reinforced. It is also the most widely known. KT's P/B has already moved from a historical ~0.56× to 0.73×. You are not early.

3. Moat

Regulated telecom oligopoly. Three carriers (SKT, KT, LG U+), enormous sunk network capex, spectrum licences, and a state-linked ownership structure. The moat is efficient scale plus regulatory barrier, and it is durable — nobody is building a fourth Korean mobile network.

It is also a moat that caps returns rather than compounding them. Normalised ROE is ~7% against a 9–10% target for 2028; ROIC on invested capital of KRW 28.4tn against normalised NOPAT ~KRW 1.5tn is ~5%. Below cost of capital. This is not a compounder and should never be valued as one. It is a bond-like cash stream with an equity option on governance reform.

Adversarial stress-test. The threat is not entry — it is regulatory extraction. The Korean government has repeatedly pressured carriers on retail pricing, and the state-linked governance means KT's board is not fully insulated from political priorities. The second threat is substitution at the value layer: messaging, content and payments migrating to KakaoTalk, Naver and the global platforms, leaving KT with the pipe and the capex.

Evergreen rating: 7/10 as a business (the pipe endures), 4/10 as an investment (the returns are capped and the growth is nil).

The AICT pivot — currently slideware. The Microsoft partnership (Oct 2024, KRW 2.4tn / $1.8bn over five years) targets AX revenue of KRW 269bn in 2025 → KRW 1.4tn by 2029. Against a KRW 28tn revenue base, today's figure is under 1% of revenue. And the direction is wrong: Q1 2026 B2B revenue −2.2%, parent AI/IT revenue −2.3%, KT Cloud revenue flat with net profit down ~80% (KRW 19.3bn → 3.9bn). Critics also note the Microsoft tie-up locks KT to a single stack, undercutting its "open AI" positioning. Treat AICT as free option value, not as an earnings driver — and note it is currently decelerating.

K Bank / BC Card — a defused liability, not crystallised value. K Bank listed on KOSPI 2026-03-05 after three cancelled attempts (60M shares, ~KRW 498bn raised, ~KRW 3.37tn valuation at the KRW 8,300 offer), down from the KRW 5.3tn targeted in 2024. It popped ~16%, closed roughly flat, and later traded below issue. BC Card holds 33.7% ≈ KRW 1.14tn at the offer price. The real benefit was removing drag-along/call options that would have obliged BC Card to repay financial investors had K Bank failed to list by July 2026. The June 2026 lock-up expiry into a weak share price is a live overhang.

4. Sentiment and the near-term ugliness

Q1 2026 (actual): Revenue KRW 6,778bn (−1.0%), operating profit KRW 482.7bn (−29.9%), net income KRW 388.3bn (−31.5%), EBITDA KRW 1,440bn (−13.1%). Wireless KRW 1,683bn (+0.4%, 29.16M subs, 82.7% 5G); broadband +1.8%; media +1.3%; B2B −2.2%; KT Estate +72.9%.

Q2 2026 is not yet reported — expected mid-August, i.e. within days. Consensus: revenue ~KRW 6.89tn (−7%), operating profit ~KRW 609bn (−40%). ⚠️ Consensus preview only.

Those declines are largely the 2025 property comp unwinding, not core deterioration — but the tape will not read it that way, and that is the setup that creates the entry.

The security breach — KT's own, and worse than SKT's. Attackers cloned femtocell certificates and moved through KT's network from 2024-10-08 to 2025-09-05 — eleven months undetected, discovered only via a customer complaint. 16,647 subscribers exposed; ~KRW 240m of fraudulent micropayments across 368 victims. PIPC fined KT KRW 53.979bn on 2026-07-30, with corrective orders and a police referral. Root causes were basic hygiene failures: ten-year certificate validity, no source-IP restriction, a route bypassing the management server.

Subscriber consequence. KT had gained ~196,700 SKT defectors in May 2025; after its own breach it lost ~230,000 subscribers, with SKT taking roughly three-quarters, accelerated by a penalty-fee waiver. Regulators voided carrier liability waivers in Aug 2026. The fine is immaterial (~0.4% of market cap); the churn and the reputational damage are not.

Governance. CEO Kim Young-shub declined reappointment citing responsibility for the breach. Ex-enterprise head Park Yoon-young takes over from March 2026. Nomura explicitly flagged that its thesis depended on Kim's cost discipline continuing. State-linked appointment risk is structural — the 2023 precedent saw an administration block a reappointment outright.

Analysts. Nomura cut Buy → Neutral (2025-12-05) while raising its target to KRW 56,000 — an explicit valuation call: EV/EBITDA 3.97×, P/B 0.76×, with the 49% foreign ownership cap named as a ceiling on incremental foreign buying. Goldman upgraded to Buy (KRW 59,000); UBS initiated Buy (KRW 55,000). Consensus skews Hold. Yahoo's mean target $22.44 implies ~21% upside.

The most underrated constraint: foreign ownership is at the 49% statutory cap under the Telecommunications Business Act. There is effectively zero headroom for incremental foreign buying, which structurally limits how far a foreign-flow-driven re-rating can run.

5. Valuation

Company type: mature dividend-paying utility-like telecom with real book value. Weight DYT, DDM and P/B heavily; Graham moderately (there is genuine asset backing); FCF yield as a cross-check. Growth models are irrelevant.

Multiples (per ADR, on normalised earnings): Reported P/E(ttm) 8.24× · Normalised P/E ~9.8× (normalised net ~KRW 1.25tn ≈ $906M on $8.83B cap) · P/B 0.73× · P/FCF 13.7× · FCF yield 7.4% · Dividend yield 4.48%

Graham IV — √(22.5 × 2.25 × 25.54) = $35.95, 94% above the price. The most bullish Graham reading in today's batch, and it is directionally meaningful because unlike the services names KT has real, hard, depreciating assets behind the book value. But Graham on a capital-intensive carrier systematically overstates: book value here is network equipment that must be continually replaced out of the same cash flow that funds the dividend. Weight it as a floor indication, not a target. Call the asset-based floor ~$16–18 (roughly 0.65–0.70× book).

Dividend Yield Theory — and it does not say what the yield implies. Current 4.48% against a 5-year average of 5.30%. The yield is BELOW its own historical band. Per Knowledge/Playbook/pitfall-dyt-inverts-when-price-caused-the-yield, decompose: the dividend has grown and the price has re-rated (P/B 0.56 → 0.73), with price doing most of the work. Forward, the KRW 2,400 minimum DPS = KRW 1,200/ADR ≈ $0.87 → 4.7% forward yield — still below the 5.3% average. DYT verdict: fairly valued to slightly expensive versus its own history. This is the single most important counterweight to the Value-Up bull case, and it directly contradicts the Graham reading.

Dividend coverage (mandatory high-yield test). Correct denominator is FCF. FY25 dividends KRW 578bn on FCF KRW 901bn = 64% FCF payout; including buybacks, 92%. The 2026 commitment (50% of adjusted standalone net income) is formulaic, not progressive — meaning the dividend falls with earnings. Given 2026 earnings will decline against the inflated 2025 base, the KRW 2,400 minimum is the floor doing real work here, and it is only guaranteed for 2026. Stress test: at −25% to normalised FCF (a plausible capex-heavy, churn-hit year), FCF ≈ KRW 675bn against KRW 578bn of dividends — coverage survives, buybacks do not. Verdict: the dividend is covered; the buyback is the variable that absorbs a bad year.

DDM — KRW 2,400 DPS/common growing 3% at a 9% required return → KRW 41,000/common ≈ $29.7/ADR. Highly sensitive to the growth assumption; at 1% growth it gives $22.3/ADR. Treat the $22–30 band as the optimistic bookend.

Bogle expected return: 4.7% forward dividend + ~2% buyback + ~2% earnings growth = ~8.7% before multiple change, plus optionality on P/B moving 0.73 → 0.85–0.90 if the 2028 ROE target lands. That is a respectable, low-beta (β 0.09) total return — it is simply not a bargain.

Fair value range: $18.00 – $24.00, centre ~$21. - Bear $18 — 0.70× book; 2026 earnings disappoint against the property comp, churn continues, buyback absorbed. - Base $21 — 0.80× book / ~10× normalised earnings; Value-Up delivers as committed, AICT stays immaterial. - Bull $24 — 0.90× book; ROE moves toward the 9–10% target, Korean reform re-rates the whole cohort, AICT begins contributing.

At $18.55 the stock sits at the bottom of the range — the most attractive relative position of the seven names analysed today. The catch is that the range itself is narrow: this is an 8–9% total-return asset, not a double.

6. Tensions and the debate round

Valuation internally contradicts itself, and that is the whole finding. Graham says $35.95 (94% upside). DYT says the yield is below its own five-year average, i.e. slightly expensive. Both are computed correctly from the same data.

Resolution. Graham is measuring the asset base; DYT is measuring the income stream against its own history. For a capital-intensive carrier whose book value is depreciating network equipment that must be replaced from operating cash flow, the asset base is not distributable and Graham overstates. The income stream is what the shareholder actually receives. Weight DYT and the FCF yield above Graham, which is why the fair value lands at $18–24 and not near $36. Graham's contribution is the floor: at 0.73× book with a covered dividend, permanent capital loss is unlikely.

Fundamentals vs. Sentiment. Fundamentals sees 10.9% net income CAGR and a 7.4% FCF yield. Sentiment replies that the CAGR is an artefact of two distorted base years and that Q2 is about to print operating profit down ~40%. Resolution: Sentiment wins on the near term, Fundamentals on the medium term. The comps are ugly and the underlying business is not deteriorating as fast as they will suggest. That gap is the opportunity — but it argues for waiting for the print, not buying into it.

7. Risks

  1. The 2026 comp wall — Q1 operating profit −29.9%, Q2 consensus −40%, driven by the non-repeating 2025 property income. Headlines will be worse than the business.
  2. Formulaic payout policy — 50% of net income means the dividend falls with earnings; only the 2026 KRW 2,400 minimum is committed.
  3. Regulatory price pressure from the Korean government on retail tariffs.
  4. Breach fallout — ~230,000 subscribers lost, KRW 54bn fine, ongoing reputational damage and a police referral.
  5. Governance / CEO risk — new CEO from March 2026; state-linked appointment risk is structural, with a 2023 precedent of political intervention.
  6. Foreign ownership at the 49% legal cap — near-zero headroom for incremental foreign buying.
  7. KRW/USD translation on a KRW-reporting ADR, plus Korean dividend withholding tax.
  8. Capex cycle — KRW 4tn/yr and non-discretionary; any AI-datacentre buildout adds to it.
  9. ADR liquidity and the 0.5:1 ratio confusing screens and price alerts.
  10. AICT decelerating while being marketed as the growth story.

A portfolio-specific passage was removed from the public build.

9. Verdict

WATCH — conviction 5.5 / 10. Fair value $18–24, centre $21. Entry $16.50–18.00. Trim at 9× forward earnings.

The highest-quality catalyst in today's batch attached to the lowest-quality business. The Value-Up commitment is board-approved, dated, tax-advantaged and legally reinforced — that is as concrete as a re-rating catalyst gets. Against it: sub-cost-of-capital returns, no growth, a formulaic dividend that falls with earnings, a foreign ownership cap with zero headroom, and a yield that is below its own five-year average, which is the market telling you the re-rating has partly happened already.

The specific opportunity is timing. Q2 prints within days and will show operating profit down roughly 40% on a comp distortion, not a business collapse. If that print pushes the ADR toward $17, the forward yield goes above 5% — back inside its historical band — with a board-committed floor under it. That is the version of this trade worth taking.

What would change the rating: the ADR below $17 post-print · FY26 DPS confirmed above the KRW 2,400 minimum · evidence AICT revenue is inflecting rather than declining · subscriber share stabilising · normalised ROE tracking toward the 9–10% target.

Next check: Q2 2026 print, expected within a week of this report.