KDDI From a Long-Term Owner's Perspective

KDDI's history is easy to understand. According to the FY2025 annual securities report, before FY2026 it was mainly divided into two reporting segments: Personal Services and Business Services. The former serves individual customers through multi-brand mobile communications, fixed broadband, finance, energy, and lifestyle services. The latter serves corporate customers through handsets and devices, networks and cloud, Telehouse data centers, IoT, DX, cybersecurity, and other solutions. The company has announced that from FY2027 it will reorganize its segments into Telecom Core, Personal Growth, Business Growth, Others. In substance, this separates the "telecom core" from the "growth domains extended from telecom" more clearly.

How this company makes money is not complicated. The first layer is typical telecom fees: mobile ARPU, fixed broadband subscriptions, device sales, and connectivity management. The second layer turns the communications relationship into a gateway for lifestyle services: finance, energy, Ponta Pass, Lawson, education, e-commerce, and more. The third layer uses communications infrastructure and customer relationships to provide enterprise services: corporate networks, cloud, data centers, AI integration, digital BPO, and Connected Solutions. In other words, KDDI's money-making model is essentially to "use highly recurring network subscription revenue to support relatively higher-growth value-added services and enterprise services."

Revenue recurrence is high overall. Mobile and fixed communications naturally have monthly and annual renewal characteristics. Enterprise networks, data centers, managed services, and solutions also have strong continuity. KDDI's FY2026 mobile ARPU rose from ¥4,340 in FY2025 to ¥4,440. Main-line churn edged up from 1.26% to 1.33%. Smartphone subscriptions increased from 32.877 million to 33.233 million. 5G penetration rose from 78.4% to 84.2%. Connected Solutions managed connections grew from 51.038 million to 60.632 million. These indicators show that the core business is not high-growth, but it is not declining either. It looks more like a mature-market business that keeps raising the value of existing customers.

The cost structure is also easy to grasp: network capex, spectrum-related investment, depreciation and amortization, stores and channels, device subsidies, customer service, and IT operating costs all account for meaningful shares. The FY2025 annual securities report states that group FY2025 capex was ¥678,907 million, mainly used to expand 5G coverage, handle data-traffic growth, and expand FTTH, CATV, and data-center facilities. The company planned to keep investing ¥720,000 million in FY2026. This means KDDI is not a "light-asset software company." It is a typical heavy-asset, highly regulated, cash-flow-stable but capital-intensive network infrastructure company.

Customer concentration is not the main issue. The company clearly discloses that Japan accounts for the bulk of revenue, but no single external customer accounts for more than 10% of consolidated revenue. The real dependencies are Japanese telecom regulation, spectrum and interconnection rules, industry competitive order, and the ability to keep investing in network quality. In other words, KDDI depends more on institutions and competitive structure than on any single major customer or key person.

If the stock market were shut for 5 years and the only question were "would I want to own this business," my answer would be: yes, provided the purchase price is not expensive and governance repair is observed properly. Communications demand will not disappear, customer switching costs are not low, and network assets have infrastructure characteristics. But this business does not have unlimited pricing power, nor is it a light-asset platform that can grow without reinvestment.

Business understandability score: 4/5.

Corresponding question: III. Industry and competitive structure

Japan's mobile communications industry is overall a mature industry, not a high-boom growth track. Demand is very stable, but incremental growth is limited, and the industry structure resembles an oligopoly game. Opensignal's 2026 Japan report directly defines au, NTT docomo, Rakuten Mobile, and SoftBank as the four major network operators. In regulatory materials, KDDI is also listed alongside NTT DOCOMO, SoftBank, Wireless City Planning, and others as an important regulated telecom operator.

Long-term demand stability is KDDI's most important industry advantage. Regardless of the macro environment, mobile communications, home broadband, enterprise leased lines, cloud connectivity, data centers, and cybersecurity are closer to "necessities" than discretionary consumption. The risk is not demand disappearing. The risk is price constraints, regulatory pressure on tariffs, and continuous capex from technology upgrades. Therefore, the difficulty in this industry is not "whether there is demand," but "whether shareholders can still receive sufficiently high free-cash-flow returns after maintaining network competitiveness."

The industry is not immune to disruption. Technologically, 5G, Open RAN, AI network optimization, satellite communications, and 6G evolution may all change capex and cost curves. Commercially, Rakuten Mobile continues to act as a pricing disruptor. Regulationally, Japan's Ministry of Internal Affairs and Communications imposes long-term constraints through tariffs, interconnection, consumer protection, and restrictions on bundling devices with tariffs. The FY2025 annual securities report explicitly lists registration, interconnection, consumer-protection, and fair-competition requirements under the Telecommunications Business Act and Radio Act frameworks.

I would define KDDI's industry position as: one of Japan's three most important incumbent integrated telecom operators, and probably one of the highest-quality and most shareholder-friendly in capital allocation. But it is not an overwhelming monopolist, nor an unchallengeable sole winner. Its main competitors are the NTT DOCOMO system and SoftBank, while Rakuten is more like a "floor setter" for industry profitability. This structure means KDDI is unlikely to enjoy very high excess profits for long periods, but it is also unlikely to be easily displaced.

On pricing power, KDDI has some ability to raise price, but it is not strong. Mobile ARPU did rise from FY2025 to FY2026, showing that the company is not without elasticity in pricing and mix. But in Japan's mature mobile market, real pricing power is jointly constrained by competition and regulation, and costs cannot be passed through freely as they can at a strong consumer brand. For this type of company, the better metric is LTV improvement, rather than one-off increases in nominal tariffs.

So this looks more like an "excellent company in a moderately attractive industry" than "the absolute king in a superb industry." Stable industry demand, high entry barriers, and visible cash flow are the good side. Average growth, heavy regulation, capital intensity, and recurring price-war risk are the other side.

Industry attractiveness score: 3/5.

Moat and Management

Corresponding question: IV. Moat analysis

KDDI's moat is not singular. It is built from multiple overlapping layers.

First are licenses and regulatory barriers. Telecom is naturally governed by the Telecommunications Business Act and Radio Act. Network construction, spectrum, and interconnection are not things a new entrant can simply decide to do. Regulation is not always favorable to incumbents, but it does raise the barrier to replication. Even a well-funded new player would struggle to replicate a nationwide network, spectrum resources, and channel system in a short time.

Second are scale and network quality. KDDI's network experience has long been in the first tier in independent third-party assessments. Opensignal's April 2026 Japan report states that au again won the Consistent Quality and Reliability Experience national awards outright. Reliability Experience was won for the fourth consecutive time, Consistent Quality scored 88.3%, and Reliability Experience scored 967/1000. This matters for a telecom company because network quality is not a "nice to have." It is the foundation for churn, brand premium, and channel efficiency.

Third are switching costs and ecosystem stickiness. Mobile communications already involve multiple bindings: phone numbers, plans, family contracts, points, devices, card linking, broadband, finance, and energy. KDDI's multi-brand system, au, UQ mobile, and povo, layered with au HIKARI, finance, energy, Ponta/Lawson, and other services, means it is not really competing with a SIM card alone, but with an account relationship and a portfolio of daily-life use cases. FY2026 5G penetration has already reached 84.2%, and Connected Solutions managed connections continue to grow, which also shows customer relationships being further embedded into devices and scenarios.

Fourth are channels and execution. Telecom is not a purely online virtual product. Offline stores, customer service, billing, devices, after-sales service, enterprise sales, construction, and maintenance all matter. Through its group system, KDDI covers retail, engineering, BPO, network construction, and operations. Even if competitors could theoretically invest to replicate this, in practice it would require years and enormous capital.

But this must be stated clearly: KDDI's moat does not automatically widen. It is more like a stable infrastructure-type moat that requires constant maintenance. If capex weakens, the moat gradually narrows. If regulatory price pressure becomes too strong, profit is squeezed. If growth businesses earn poor returns on capital, they may dilute the advantage instead. My moat judgment is therefore: stable and fairly strong, but not continuously widening.

In an inflationary environment, KDDI is not the ideal price-raising asset, but it is not helpless either. It can more often indirectly increase unit customer value through plan upgrades, brand migration, cross-selling, and service bundling. During economic downturns, communications spending usually falls later than discretionary consumption, so downside earnings elasticity is relatively limited. This is also why it can work as a "defensive holding" within a portfolio.

I am more inclined to view past margins as the result of structural advantages combined with a mature oligopolistic industry, rather than a simple cyclical tailwind. From FY2016 to FY2022, the operating margin broadly stayed in the 18.6%-20.0% range. FY2024, after restatement, temporarily fell to about 16%. FY2025 recovered to 18.6%, and FY2026 was 18.1%. This looks more like modest fluctuation at a stable mature enterprise than the boom-bust pattern of a highly cyclical company.

Moat strength score: 3/5.

Corresponding question: V. Management and capital allocation

Management and capital allocation are the areas where KDDI deserves the most credit, but also the ones that cannot be taken lightly.

Start with the positives. KDDI has long maintained stable and rising shareholder returns. The company officially discloses that as of FY2026 it had achieved 24 consecutive fiscal years of dividend increases. The FY2026 dividend per share was ¥80, and the FY2027 plan is ¥84. Over the longer history, the company conducted buybacks almost every year from FY2016 to FY2025, with annual buyback amounts gradually rising from ¥50bn to ¥400bn. The latest medium-term strategy also states that the company will maintain stable dividend growth, conduct buybacks flexibly after comparing returns on growth investments, and control CAPEX/Sales at ≤12% and net debt/EBITDA at around 2.0x. This capital-allocation language itself is mature and rational.

Second, KDDI's recent management language puts more emphasis on capital efficiency than traditional telecom companies often do. The new medium-term strategy directly incorporates ROE/ROIC spread into executive performance indicators, and links medium- to long-term equity incentives to TSR and ROIC spread. This shows that the company recognizes "scale growth" is not a substitute for "per-share intrinsic value growth." For long-term shareholders, this is a clearly positive signal.

Now consider capital-allocation actions themselves. Buybacks reached ¥300bn in FY2024 and ¥400bn in FY2025. In May 2026, the company approved another buyback of up to ¥300bn and approved the cancellation of treasury shares exceeding 5% of total outstanding shares. Securities reports and announcements also show that some buybacks were conducted through tender offers absorbing shares from long-term strategic shareholders such as Kyocera and Toyota. On one hand, this helps reduce cross-shareholdings and simplify the capital structure. On the other hand, it also means the buybacks are not driven entirely by "market undervaluation," but meaningfully by governance and shareholder-structure rebalancing. This is not bad, but it should not be mythologized as a "textbook bargain buyback."

But the negative must be stated plainly: the 2026 internal-control issue significantly lowered the management credibility score. On 2026-03-31, KDDI disclosed improper transactions in the advertising agency business at BIGLOBE and its subsidiary G-PLAN, and retrospectively corrected FY2023-FY2025 and parts of FY2026 interim financial statements. After correction, FY2025 full-year revenue decreased by ¥82.4bn, operating profit decreased by ¥31.2bn, and net profit attributable to owners of the parent decreased by ¥30.3bn. The Tokyo Stock Exchange subsequently required KDDI to submit an improvement report, which the company submitted on 2026-06-02. The special investigation committee disclosed that it found no prior knowledge by management and no other similar cases. For a long-term owner, however, "not found" is not the same as "governance is risk-free."

On equity incentives, KDDI is not a typical owner-operator with a founder holding a large stake and fully aligned with minority shareholders. In its shareholder structure, Master Trust, Kyocera, and Toyota still occupy important positions, and treasury shares are also substantial. Management-shareholder alignment depends more on institutionalized incentives and governance arrangements than on large personal holdings by managers. The company recently continued and revised its BIP trust-style performance share plan, and emphasized encouraging executives to actually hold KDDI shares during their tenure. The direction is right, but it is not enough to equate with a "strong ownership culture."

Overall, my assessment of KDDI's management is: its capital-allocation philosophy is better than that of most mature telecom companies, but its governance score was reduced by a real event in 2026. If the improvement report is properly implemented over the next 12-24 months and no further spread appears, this crack can be repaired. If similar issues appear again, the investment thesis must be substantially reassessed.

Management and capital allocation score: 3/5.

Financial Quality and Owner Earnings

Corresponding question: VI. Financial quality analysis

Start with the financial skeleton. In FY2026, KDDI generated revenue of ¥6,071.9bn, up 4.1% year over year; operating profit of ¥1,099.1bn, up 1.1%; net profit attributable to owners of the parent of ¥707.1bn, up 7.9%; free cash flow of ¥708.4bn; depreciation and amortization of ¥686.3bn; cash-flow-basis capex of ¥679.0bn; net debt of ¥4,296.5bn; ROE of 14.0%; and net debt/EBITDA of 2.33x. FY2027 guidance gives revenue of ¥6,410.0bn, adjusted operating profit of ¥1,210.0bn, adjusted net profit attributable to owners of the parent of ¥731.0bn, and a dividend of ¥84.

Over a longer period, KDDI's growth is not spectacular, but it is steady enough. According to the official historical financial highlights, reported revenue rose from ¥4,466.1bn in FY2016 to ¥5,918.0bn in FY2025. Including the latest FY2026 result of ¥6,071.9bn, the ten-year revenue CAGR is roughly 3%. Net profit rose from ¥494.9bn in FY2016 to ¥707.1bn in FY2026, also at around 3%-4% CAGR. This is the typical curve of a mature, high-quality public-network company: slow, but not poor; heavy, but not fragile.

On margins, the long-term operating margin basically stayed in the high teens to around 20%. The FY2016-FY2022 historical range was about 18.6%-20.0%. FY2024, after restatement, fell to about 16.0%. FY2025 recovered to 18.6%, and FY2026 was 18.1%. For a mature telecom business, this is very good margin stability, but it also shows that the company does not have the explosive operating leverage of an internet platform.

Cash-flow quality should be viewed in two layers. The first layer is whether accounting profit is a "false boom." I think overall it is not. Operating cash flow has been strong over the past decade. FY2016-FY2025 reported operating cash flow was in the ¥884.5bn-¥1,706.5bn range, showing that this business does indeed keep producing cash. The second layer is whether free cash flow is stable. The answer here is: stable through the cycle, but volatile year by year. FY2025 free cash flow was only ¥68.9bn, while FY2026 recovered to ¥708.4bn. But if we look at the company's preferred "Core FCF," FY2026 was ¥990.2bn, clearly above headline FCF. This reflects that KDDI's headline FCF is easily disturbed by strategic investments, finance businesses, and investment-cash-flow classification, while core cash earnings are closer to the long-term cash generation that owners truly care about.

The balance sheet is not fragile, but leverage is no longer low. At the end of FY2026, the equity ratio declined to 26.6%, the D/E ratio was 1.06, and net debt/EBITDA was 2.33x, up materially from 1.95x in FY2025. The good side is that the company still has an R&I AA / a-1+ credit rating, and the direct-bond maturity structure is relatively long, indicating strong financing capacity. The bad side is that if the company keeps conducting large buybacks and growth investments over the next few years while the telecom core slows, leverage headroom will be thinner than before.

For receivables, payables, and inventory, the current English materials do not provide a full FY2026 working-capital breakdown in a unified, restated format, so I will not fabricate numbers. What can be said clearly is that KDDI has no single major customer dependency, and most revenue comes from Japan through recurring subscriptions and services. This model itself does not look like a business where receivables easily spin out of control. The real item to monitor is not traditional bad debt, but transaction authenticity at subsidiaries and the ability of internal controls to see through the group.

Below is a key financial table prioritized for "decision usefulness" and based as much as possible on restated figures. For FY2023-FY2025, I prioritize the 2026-03-31 correction announcement and restated FY2026 Data Book data. For earlier years, I use historical highlights from the Integrated Sustainability and Financial Report 2025. Items that cannot be confirmed are marked unknown directly.

Table note: FY2016/FY2020/FY2022 are from the historical highlights in KDDI's 2025 Integrated Report. FY2023/FY2024/FY2025 revenue, operating profit, and net profit use the 2026-03-31 correction announcement. FY2025/FY2026 FCF, capex, net debt, and ROE are from the FY2026 Data Book and FY2026 financial summary. Because the official English historical highlights page states that it has not yet fully retrospectively reflected the 2026 financial-statement corrections, and because the English cash-flow details for FY2023-FY2024 were not decision-grade in the current retrieval chain, those items are left as unknown.

Judgment: Profit is broadly real cash profit, not "paper prosperity." But growth is not effortless, and the business still requires continuous capital investment. KDDI is not a company that "needs more and more money the more it grows," but it is also absolutely not a company that can grow with almost no reinvestment. My conclusion: financial quality is good and cash flow is real, but growth quality must be evaluated together with capex and returns.

Corresponding question: VII. Owner Earnings analysis

The Buffett-style key is not accounting profit, but estimating how much the company can distribute to owners without harming its competitive position.

My conservative estimate for KDDI starts from FY2026: net profit attributable to owners of the parent of ¥707.1bn; add back depreciation and amortization of ¥686.3bn; then subtract maintenance capex. The problem is that the company does not publicly disclose a precise split between maintenance capex and growth capex, so this step requires assumptions. Given that KDDI is both a mature network-asset operator and still investing in 5G, data centers, and enterprise growth, I believe it is reasonable to treat about 72%-96% of FY2026 cash-flow-basis total capex of ¥679.0bn as maintenance capex. The midpoint is about 80%, corresponding to maintenance capex of about ¥543bn. Assuming working capital is roughly neutral around the long-term average, this produces neutral Owner Earnings of about ¥850bn. Under more conservative maintenance-capex assumptions, normalized Owner Earnings could fall into the ¥700bn-¥800bn range.

This estimate is reconcilable with the company's disclosed cash framework. FY2026 reported free cash flow was ¥708.4bn, while the company-defined Core FCF was ¥990.2bn. The former is more disturbed by investment cash flows, while the latter is closer to operating substance. I therefore would not treat ¥708.4bn as the only "true earnings," nor would I directly treat ¥990.2bn as fully distributable owner earnings. The middle region between the two is more likely to reflect KDDI's long-term distributable capacity.

At the current share price of ¥2,706.5, using conservative Owner Earnings of ¥700bn-¥800bn, the equity value corresponds to a P/OE of about 13-15x. Using neutral Owner Earnings of ¥850bn, it is a bit above 12x. This is not expensive, but it is not an obvious bargain either. It looks more like the market assigning high-quality Japanese telecom cash flows a "reasonable but slightly rich" price.

My conclusion is: KDDI's true earning power is probably higher than single-year net profit and higher than the worst-year headline FCF, but not high enough to make the current price easy to call "materially undervalued." This is exactly why I assign "Watch," not a direct "Buy."

Intrinsic Value and Margin of Safety

Corresponding question: VIII. Intrinsic value estimate

Owner Earnings Discount Method

I use three scenarios, all based on the normalized Owner Earnings assumptions above, while explicitly acknowledging that maintenance capex cannot be directly verified, which is the most fragile part of this valuation step.

These figures are not company-disclosed values. They are estimates based on FY2026 net profit, depreciation and amortization, capex, capital structure, and business maturity. They are better suited for judging whether there is a margin of safety than for placing a bet with decimal-point precision.

Relative Valuation Method

Based on a roughly contemporaneous market snapshot, KDDI's current valuation is about: P/E 14.75x, P/B 2.01x, P/S 1.70x, EV/EBITDA 7.63x. Compared with large Japanese peers, NTT is about P/E 11.72x, P/B 1.24x, EV/EBITDA 7.09x; SoftBank Corp is about P/E 19.31x, P/B 3.49x, EV/EBITDA 8.16x. This means the market positions KDDI as: more expensive than NTT because the market considers it higher quality and steadier in execution; cheaper than SoftBank Corp because its growth story and financial-leverage narrative are less aggressive.

In other words, relative valuation does not signal "obvious undervaluation." It only says KDDI's price is broadly within a reasonable range for a quality Japanese telecom stock. If you value certainty over an exciting story, this valuation is not absurd. But if you insist on buying only when the price is "significantly below intrinsic value," it is not cheap enough.

Asset Value or Liquidation Value Method

For a company like KDDI, liquidation value is of limited use because the value of networks, machine rooms, and telecom equipment mainly lies in continued operation, not in selling them off piece by piece. At the end of FY2026, total equity was ¥5,592.7bn, and equity attributable to owners of the parent per share was about ¥1,333. The current share price corresponds to about 2x PB. At the parent-company level, the fair value of external equity investments at the end of 2025 was ¥209.9bn, above book value of ¥98.4bn, showing that some hidden asset value does exist. But the scale is not enough to change the overall valuation conclusion. My judgment is: the asset method can only provide a weak floor for KDDI and should not be the main valuation anchor.

Final intrinsic value range:

• Conservative intrinsic value range: ¥2,300-2,700/share

• Reasonable intrinsic value range: ¥2,900-3,500/share

• Optimistic intrinsic value range: ¥4,000-5,000+/share

Current price relative to intrinsic value: At about ¥2,706.5, the current price is roughly near the upper end of the conservative range and below the base range, but it does not provide a very generous discount for conservative investors.

Ideal buy price range: ¥2,200-2,500 This range means buying near or below conservative value, better matching a "margin of safety first" principle.

Acceptable holding price range: ¥2,500-3,200 Within this range, the stock is more suitable for "holding rather than adding."

Clearly overvalued price range: above ¥3,600 If there is no unexpectedly strong, high-quality growth and the valuation continues to climb into this range, I would view it more as a bond substitute being chased than a true value-investing opportunity.

Corresponding question: IX. Margin of safety

Is the current price cheap enough? I do not think so. Not because the company is poor, but because your risk preference is conservative and your horizon is long. What you need most is a purchase price that will not cause serious damage even if you are wrong. KDDI's current payoff looks more like a medium-return, low-to-medium-risk investment than a situation where "you lose little if wrong and make a lot if right."

The most fragile valuation assumption is the share of maintenance capex. If true maintenance capex is far higher than I assume, Owner Earnings will fall materially, and the base intrinsic value range will be revised downward. The second fragile assumption is that non-telecom growth domains can continue to contribute growth above the cost of capital. If these businesses merely "look like growth" but require higher investment or produce lower returns, the base case will also be distorted.

If growth falls short of expectations, does the investment still work? It can still work, but returns become less attractive. Suppose KDDI grows only 1%-2% in real terms going forward, while valuation converges toward NTT. Your total return may be reduced to the dividend plus a little earnings growth, potentially falling to low single digits annualized. For existing holders, this is not necessarily a disaster. For new buyers, it means opportunity cost is not low.

If margins decline, does the investment still work? In the short term, yes; in the long term, it depends on the magnitude. Telecom-core demand is stable, so a mild margin decline usually does not break the business model. But if regulatory price pressure, price wars, and weaker returns from growth businesses combine, and the operating margin trends down toward the low-to-mid teens while capex cannot be reduced, this is no longer a case of "a good company just growing slowly." It becomes "the return-on-capital structure has deteriorated."

My final judgment: good company, and the price is not bad; but good company at an ordinary price is the accurate description. Conservative capital should wait for a better price.

Risks, Checklist, and Final View

Corresponding question: X. Risks and opposing view

The most important permanent-capital-loss risks are not about share-price volatility, but about the following issues:

First is competition and regulatory risk. Japan's mobile industry is mature and crowded, and price wars plus MNP competition appear cyclically. NTT Docomo has already acknowledged in 2026 investor materials that prioritizing customer-base reinforcement, increased sales-promotion expenses, and intense MNP competition hurt profit. This shows that industry competition has not disappeared. KDDI's quality is steadier, but it cannot stand entirely apart.

Second is technology and capex risk. Network quality is not free. 5G, data centers, AI networks, and future 6G evolution will continue to consume capital. If capex and lease investment stay high over the next several years while revenue growth remains in the low single digits, the real cash return to shareholders will be pressured.

Third is governance and accounting risk. The 2026 false circular transactions at subsidiaries are the negative factor this report must treat most seriously. Even if the special investigation committee said it found no prior knowledge by management and no other similar cases, and even if the company has submitted an improvement report, the incident still proves that KDDI is not an "absolutely flawless top student." If the scope later expands, remediation proves incomplete, or similar issues appear again, the assessment of management credibility and valuation premium must be reduced directly.

Also note valuation and opportunity-cost risk. KDDI's current dividend yield is about 2.96%, below NTT's 3.59% and SoftBank Corp's 3.93%. If you buy it not because it is clearly cheap, but simply because "it is stable and many people like it," you can easily fall into the trap of a good company with average returns.

There is also ADR and currency risk. KDDIY reflects the same equity, but U.S.-dollar investors are additionally exposed to yen exchange-rate fluctuations. Even if the ordinary shares perform steadily in operating terms, the ADR's U.S.-dollar return can be materially eroded by currency moves.

Strongest opposing view: Bears would say that KDDI is still essentially a mature utility-style telecom operator with slow growth, high capex, and limited real pricing power. The so-called growth stories in finance, energy, Lawson, enterprise AI/DX, and similar areas may simply put a "growth coat" on a mature core business, without being enough to support a sustained high valuation. Combined with the 2026 financial restatement incident, the governance premium the market is willing to give it should decline. If price wars return, KDDI's "quality premium" will be compressed from both sides: slower earnings growth and a lower valuation multiple. I think this opposing view is not weak.

Facts that would overturn the investment view: If any two or three of the following happen at the same time, I would admit the view was wrong: first, internal-control problems continue to spill over, with a second round of corrections or further penalties; second, net debt/EBITDA stays above 2.5x for a long time with no clear path downward; third, core-business ARPU declines, churn rises, and 5G/connection growth also slows; fourth, non-telecom growth businesses earn margins and returns below the cost of capital; fifth, management keeps conducting buybacks merely to absorb cross-shareholdings or polish EPS, rather than to create per-share intrinsic value.

Largest permanent-capital-loss scenario: The danger is not a 20% share-price decline. It is that "the market finally realizes this is just a low-growth, high-capex, governance-discounted utility stock," meaning you bought at a price that was not cheap and over the next ten years receive only low-single-digit annualized returns, or even negative dollar returns if the exchange rate turns unfavorable. That is more dangerous than short-term volatility.

Corresponding question: XI. Comparison with other opportunities

Compared with its strongest competitors, I would place KDDI between NTT and SoftBank Corp:

• Relative to NTT, KDDI is more expensive, but perceived quality and execution are also better;

• Relative to SoftBank Corp, KDDI is more conservative and cleaner, but growth and valuation elasticity are weaker. If you seek a higher dividend and lower valuation, NTT may be more attractive. If you seek a stronger growth narrative, SoftBank Corp may be more exciting. If you seek "a more balanced mix of quality and certainty," KDDI is indeed good.

Compared with a broad-market index, KDDI is probably not an opportunity that is "significantly better than the index." It is more likely to provide a return curve with lower volatility but also lower upside elasticity than the index. If your portfolio needs a defensive Japanese equity position with clear cash flow and lower correlation to high-growth assets, KDDI has a place. If you are seeking significant excess return, it does not naturally stand out. This judgment is mainly my inference based on its growth rate, capital intensity, and current valuation.

Compared with the risk-free rate or high-grade bonds, KDDI's advantage is not the current dividend itself, but dividend growth + buybacks + slow-growth total return. But this only works if the purchase price is reasonable. If you buy too expensively, you take equity risk while earning only a little extra return.

If I could hold only 5 assets, I would not put KDDI at the current price directly into the core 5. It qualifies for the candidate list, but it is better suited to enter after a better price or clearer governance repair. At present, it is more like a "second-line core asset" than a "must-own heavy position."

Corresponding question: XII. Investment Checklist

The above judgments are based on KDDI's latest financial summary, annual securities report, historical highlights, capital-allocation policy, credit ratings, third-party network-quality reports, and this year's financial-statement correction incident.

Corresponding question: XIII. Final investment conclusion

【Final Rating】 Watch

【One-sentence investment thesis】 KDDI is a high-quality, stable Japanese communications cash-flow business that can be held for the long term, but the current price does not provide enough margin of safety for a relatively conservative long-term investor to buy comfortably.

【Core bull case】 First, telecom-core demand is stable, customer relationships are recurring, and the business is easy to understand. Second, licenses, network scale, and network quality form a stable moat, and au remains in the first tier in independent third-party network-experience assessments in Japan. Third, the long-term shareholder-return culture is good, with 24 consecutive years of dividend increases and long-term buybacks. Fourth, FY2026 financial quality remained sound, with revenue, profit, FCF, and ROE all staying at good levels. Fifth, the new capital-efficiency framework incorporates ROE/ROIC spread into incentives, helping restrain the impulse to "chase scale instead of value."

【Core bear case】 First, the growth rate is fundamentally not fast, and the business remains a mature telecom utility at its core. Second, the business is capital-intensive, and maintaining network competitiveness requires continuous capex. Third, the 2026 subsidiary improper-transaction and financial-restatement incident shows governance is not flawless. Fourth, the current valuation is not cheap enough to cover estimation error. Fifth, non-telecom growth businesses may not all sustain high returns over the long term, creating the risk of "apparent diversification that actually dilutes returns."

【Key assumptions】 First, the 2026 internal-control event does not continue to expand. Second, telecom-core ARPU and churn remain broadly stable. Third, CAPEX/Sales can stay near the company's target, and leverage does not continue to deteriorate materially. Fourth, growth domains such as finance, energy, Lawson, enterprise AI/DX, can keep contributing profits above the cost of capital.

【Fair buy price】 ¥2,200-2,500/share. The basis is that this range is closer to my conservative value range and leaves a more adequate buffer for risks such as misjudging maintenance capex, governance repair falling short of expectations, and valuation multiple compression.

【Target holding period】 More than 10 years. The real advantage of this type of asset lies in dividend growth, buybacks, customer stickiness, and capital-allocation compounding, not in 12-month valuation swings.

【Expected annualized return】 At the current price, my estimate is:

• Conservative scenario: 3%-5%

• Base scenario: 6%-8%

• Bull scenario: 9%-11%

If a position is built in my ideal buy range, the above returns can broadly move up by about 2 percentage points. These are inferences based on Owner Earnings, dividend growth, and valuation reversion, not company guidance.

【Maximum loss risk】 If regulation and competition suppress tariffs, growth-business returns fall short, a second round of governance issues emerges, and valuation converges toward cheaper NTT levels, medium- to long-term permanent capital loss could reach 30%-45%. For ADR investors, if yen depreciation is also added, U.S.-dollar returns would be worse.

【Tracking indicators】 I recommend continuously tracking the following 10 items: first, mobile ARPU; second, main-line churn; third, 5G penetration; fourth, smartphone subscriptions and Connected Solutions managed connections; fifth, operating-profit growth in Telecom Core and growth segments; sixth, the gap between Core FCF and reported FCF; seventh, CAPEX/Sales; eighth, net debt/EBITDA; ninth, execution of dividends and buybacks; tenth, progress on implementing the Tokyo Stock Exchange improvement report and whether new accounting/internal-control issues appear.

【Signals that trigger reassessment】 New financial-statement corrections; clear underperformance in executing the improvement report; ARPU weakens consecutively and churn rises; net debt/EBITDA exceeds 2.5x and keeps moving upward; growth-business profit growth is far below revenue growth; management continues high-intensity buybacks while per-share intrinsic value does not improve.

【Final recommendation】 If you already hold KDDI, it is still worth holding and tracking, because its business quality has not been completely overturned by the 2026 incident. If you do not yet hold it, given your "balanced but conservative" risk preference, I would be more inclined to wait. Act when either of two opportunities appears: the price returns to a more attractive range, or governance repair has more time to prove itself. Calm long-term investing is not buying immediately whenever you see a good company. It is waiting until "good company + good price + acceptable uncertainty" appear together. For KDDI today, "good company" is true, but "good price" is not yet solid enough.

Open Questions and Limitations

There are three points you should especially know about this report. First, as of 2026-06-03, the latest complete annual-report-style statutory filing was still the FY2025 annual securities report. FY2026 full-year results summary and Data Book were available, but the FY2026 annual securities report had not been obtained in this retrieval chain. Therefore, the most detailed governance, segmented asset, and working-capital disclosures still rely mainly on FY2025 statutory disclosures, with FY2026 supplemented by the results summary.

Second, some English historical financial materials for FY2023-FY2025 had not fully retrospectively reflected the impact of the 2026 financial-statement corrections. Therefore, for FY2023-FY2025 revenue, operating profit, and net profit, I prioritize the correction announcement. Some annual cash-flow/working-capital details were not fully reconstructed in the current English materials, and I did not force-fill them.

Third, Owner Earnings valuation is highly sensitive to the maintenance-capex assumption. This is not unique to KDDI. It is an unavoidable difficulty in valuing all mature telecom stocks; I have handled it with conservative ranges rather than a single-point valuation, but it remains the part of the entire report where humility is most needed.

This report is based on public information and does not constitute investment advice. Markets carry risk; invest with caution.