Tencent Holdings: A Long-Term Owner's Investment Analysis

Can this business be understood: [Opinion] I rate Tencent's "business understandability" at 4/5. It is not "simple," because there are many segments, many investment assets, complex accounting conventions, and strong regulatory factors; but if you grasp its economic essence -- using the strongest social gateway and digital infrastructure to connect attention, content, transactions, and payments, and continuously monetizing that system -- it is not hard to understand. Would I be willing to hold this business if the stock market closed for 5 years? [Opinion] Yes, but on the condition that the position cannot be too heavy, and I must accept that this is a composite business of "strong platform + heavy regulation + active capital allocation," not a single-category consumer stock you can see through at a glance.

Industry and competitive landscape: [Fact] As of the end of 2025, China's internet user base reached 1.125 billion, with internet penetration of 80.1%, and the generative-AI user base reached 602 million; this shows the Chinese internet has entered a stage of "mature user growth, penetration near its ceiling, with subsequent growth driven by ARPU/advertising efficiency/transaction penetration/AI value-add." Meanwhile, Newzoo estimated the global gaming market at roughly US$188.8 billion in 2025, and in a more optimistic update mentioned it could reach US$197 billion, indicating the gaming industry is not declining but looks more like "a maturing industry with structural growth."

[Inference] Tencent does not operate in a perfect industry. It is neither a lightly regulated industry nor a blue-sky industry where users are still expanding rapidly; but it is in an industry with stable long-term demand, a concentrated profit pool, and the strong getting stronger. Its main rivals include: NetEase in gaming; ByteDance/Kuaishou in traffic and advertising; Alipay/Ant in payments; and Alibaba Cloud, Huawei Cloud, and others in cloud and AI. [Opinion] More precisely, Tencent is "a top-tier company within a heavily regulated, mature industry," rather than "a good company within a naturally easy, good industry."

Industry attractiveness score: 3.5/5. Demand exists over the long term, but regulation, technological substitution, content review, and the AI arms race raise the level of uncertainty.

Moat and Management

Moat assessment:

The "factual basis" of the table above comes mainly from Tencent's 2025 annual report, its full-year 2025 / Q1 2026 earnings materials, and its history of capital returns; "my judgment" is an investment inference built on those facts.

Is the moat widening, stable, or narrowing: [Opinion] I lean toward judging it as "stable overall, widening locally, with new competition at the model layer." Moats such as social traffic, the payment gateway, the mini-game/Channels transaction loop, and the advertising-data loop remain solid; but in large AI models themselves, Tencent does not have an absolute, irreplaceable barrier like the WeChat relationship chain, and instead must keep investing capital and talent to catch up. In 2025, the company's CAPEX was RMB79.198 billion, and in the first quarter of 2026 CAPEX rose further to RMB31.9 billion, up 16% year over year; Reuters has also noted the company is increasing investment in AI infrastructure and models.

How long and how much capital would competitors need to replicate it: [Opinion] Replicating Tencent's most central moat is not as simple as spending money on GPUs and building data centers. What is truly hard to replicate is the social relationship chain of over 1.4 billion users, the integrated ecosystem of merchants/developers/advertisers/payment scenarios, and dozens of long-lived content assets. This requires at least many years, hundreds of billions in capital, and the policy and organizational conditions that only a very small number of companies possess. Replicating one of these businesses individually is possible; replicating the entire system is very hard.

Pricing power, inflation resistance, and counter-cyclical ability: [Inference] Tencent's pricing power is not the traditional consumer-goods kind of "I raise the price of a product," but rather raising unit economics through better advertising conversion, higher gaming ARPU, deeper embedding of payment/merchant services, and higher-value-added cloud/AI services. [Fact] Management explicitly attributed the 2025 growth in marketing services revenue to "increases in both price and ad impressions"; in the same year, overall and segment gross margins all rose. [Opinion] This indicates Tencent has a degree of "soft pricing power" in an environment of inflation and economic volatility, especially in advertising technology, gaming content, and merchant tools.

Moat strength score: 4/5. If I disregarded institutional and geopolitical risk, I would rate it higher; but Buffett-style analysis must also factor in "shareholder rights and institutional variables."

Is management trustworthy, and is capital allocation rational: [Fact] As of the end of 2025, Pony Ma held roughly 8.82% of the company through his stake; the board had 8 directors in total, of whom 5 were independent non-executive directors. This means the founder still has a strong interest alignment, and the board is formally independent as well.

[Fact] On capital returns, in 2025 Tencent repurchased about 153.4 million shares for roughly HK$80 billion; it proposed a 2025 final dividend of HK$5.30 per share, totaling about HK$48 billion. In 2024 it repurchased about 307 million shares for roughly HK$112 billion; in 2023 it repurchased about 152 million shares for roughly HK$49 billion. Earlier still, the company had distributed JD.com and Meituan shares in kind to shareholders rather than locking its portfolio on the balance sheet forever. [Opinion] This shows Tencent's capital allocation is at least not the type that "only hoards cash and only loves building a bigger empire."

But there are areas that must be discounted: [Fact] In 2025, Tencent's average buyback price was about HK$521.70 per share, clearly higher than the market price of about HK$456.40 per share in mid-May 2026; management also indicated that, given the AI investment opportunities it sees, buybacks in 2026 would most likely be lower than in 2025 while dividends would rise. [Opinion] This means buybacks did not always occur in an "obviously undervalued range"; at least in hindsight, the 2025 buybacks were not particularly cheap.

Is the equity compensation reasonable: [Fact] In 2025, the share-based compensation expense within the non-IFRS adjustments was about RMB34.711 billion, of which the executive-compensation note mentions about RMB4 billion related to a one-off factor from the restructuring of existing commercial arrangements at an overseas subsidiary; on the other hand, the number of shares outstanding fell from 9.225 billion at the end of 2024 to 9.120 billion at the end of 2025, showing that buybacks overall offset the dilution from incentives. [Opinion] I therefore would not treat equity compensation as an out-of-control problem, but it does raise the "true cost of employees," and valuation cannot pretend it does not exist.

Management and capital-allocation score: 4/5. The strengths are long-termism, gradually strengthening buybacks and dividends, and a decent eye in historical investments; the deductions are that buyback prices have not always carried a sufficient margin of safety, and the inherent ceiling on governance transparency for Chinese platform companies.

Financial Quality

The table below prioritizes the most critical metrics that best answer "is this a good business." Unless otherwise noted, all figures are on a renminbi basis; some ratios are calculated by the author from the raw statement data.

Note: The 2021-2025 revenue, profit, and balance-sheet summaries come from Tencent's 2025 annual report, 2020 annual report, and 2021 full-year earnings release; the 2022-2025 free cash flow and CAPEX come from the company's annual earnings presentations/releases. Gross margins and operating margins in the table are calculated from the raw data. For 2021 the company did not disclose full-year FCF on the same consistent public basis that I was able to verify, so it is marked "data to be supplemented."

How to read these numbers: [Fact] The 2020-2025 revenue CAGR was about 9.3%, and the 2016-2025 revenue CAGR was about 19.4%; at the same time, the 2021-2025 overall gross margin rose from 43.9% to 56.2%, and the operating margin rose from 22.3% to 32.1%. [Opinion] This is not a model of "buying growth by spending money like crazy," but rather one of "revenue growth that is not exaggerated, but quality that keeps improving." For a long-term owner, this matters more than simply pursuing a higher revenue-growth rate.

Are profits real cash profits, or accounting profits: [Opinion] They are largely real, but IFRS net profit fluctuates significantly and is easily dragged around by changes in the fair value of investments, share of profits/losses of associates, and impairments. A very typical example: in 2021 IFRS profit was inflated by investment disposals/distributions, while in 2023 IFRS profit was depressed by investment-related losses and impairments, so looking at net profit alone easily leads to misjudgment. A better lens is: non-IFRS profit + operating cash flow + free cash flow. In 2025, Tencent's operating cash flow was RMB303.052 billion, up 17% year over year; free cash flow was RMB182.6 billion, up 18% year over year; and in the first quarter of 2026 free cash flow rose further to RMB56.7 billion, up 20% year over year, showing the "cash cow" character has not disappeared.

Does growth require heavy capital investment: [Fact] In 2023 Tencent's operating CAPEX was only RMB17.2 billion, jumping to RMB63.4 billion in 2024 and rising further to RMB73.1 billion in 2025; total CAPEX rose from RMB23.9 billion in 2023 to RMB76.8 billion in 2024 and RMB79.2 billion in 2025. [Opinion] This is critical: Tencent used to be a platform company where "growth was highly profitable and capital intensity was low"; now, because of AI infrastructure investment, it is temporarily becoming a company where "growth is still profitable, but capital intensity has risen significantly." This has not broken the business model, but it has indeed lowered the certainty of the valuation.

Returns on capital and solvency: [Fact] Estimating from 2025 data, Tencent's ROE was about 21% and ROA about 11.8%; total liabilities/total assets at the end of 2025 were about 39.1%, the company-disclosed year-end net cash was RMB107.145 billion, and on 2025 adjusted EBITDA, net debt/EBITDA was about -0.32x; using IFRS operating profit/finance costs as a rough calculation, interest coverage was about 16x. [Opinion] This is a clearly strong balance sheet. Tencent's biggest risk is not financial leverage, but regulation and the return on capital allocation.

Working capital and accounting quality: [Fact] At the end of 2025, inventory was only RMB530 million, almost negligible; accounts receivable were RMB49.930 billion, up slightly from RMB48.203 billion at the end of 2024; accounts payable were RMB121.127 billion, up from RMB118.712 billion at the prior year-end; deferred revenue rose from RMB100.097 billion to RMB110.309 billion. In the cash flow statement, the main effects of working-capital changes on 2025 operating cash flow were: accounts receivable reduced operating cash flow by RMB1.242 billion, accounts payable added RMB13.741 billion, and deferred revenue added RMB6.143 billion. [Inference] This shows Tencent has not relied on aggressive credit sales to "pad the statements," and instead exhibits the healthy cash-conversion characteristics common to platform-type businesses.

Are there signs of financial fraud, aggressive accounting, or profit manipulation: [Opinion] From the structure of the public statements, I see no obvious red flags: receivables are not out of control, deferred revenue is rising, cash flow keeps up with profit, the debt structure is manageable, and the share count is falling rather than continuously expanding. But I would not say "there is no accounting complexity at all" either -- Tencent's investment assets, fair-value changes, and share of associates' results significantly distort IFRS profit, so one must cross-check using cash flow and the non-IFRS basis.

Owner Earnings and Valuation

The core idea of Owner Earnings: In Buffett-style analysis, I care more about a business's ability to generate cash truly distributable to owners than about the book net profit of any single year. The public metric closest to this concept for Tencent is the company's own disclosed Free Cash Flow, whose definition already deducts CAPEX, media-content payments, and lease-liability payments. In 2025 this value was RMB182.6 billion. But the 2024-2025 CAPEX includes a clear component of AI infrastructure expansion, so treating all CAPEX as "maintenance expenditure" would understate the long-term true earning power.

Conservative Owner Earnings estimate:

Why assume maintenance CAPEX of RMB45 billion: [Assumption] This number is clearly higher than the 2023 operating CAPEX of RMB17.2 billion and clearly lower than the 2025 operating CAPEX of RMB73.1 billion. In other words, in the valuation I have already assumed Tencent will need to bear a normalized level of infrastructure spending far above its historical baseline, but I do not treat all of the CAPEX in this 2024-2026 AI arms race as "necessary to maintain the status quo." This is a deliberately conservative -- but not excessively pessimistic -- way of treating Owner Earnings. On the current market capitalization, Tencent trades at roughly 16x conservative Owner Earnings.

The relationship to net profit and free cash flow: [Fact] In 2023 Tencent's FCF exceeded IFRS net profit; in 2024-2025 FCF was about 0.8x IFRS net profit and about 0.7x non-IFRS net profit. [Opinion] This is not a typical sign of earnings fraud; it looks more like "AI CAPEX brought forward, depressing current free cash flow." Therefore, Tencent is now better valued using "conservative Owner Earnings + segment moat + value of investment assets" rather than mechanically looking at P/E alone.

Method 1: Owner Earnings discounting I value Tencent's "operating business" by discounting it, and add the portfolio and net cash back separately in Method 3 to avoid confusion. All of the following are [Assumptions], not company guidance.

Note: The exchange-rate conversion uses the HKD/CNY central parity of 0.87343 disclosed by China Foreign Exchange Trade System on 2026-05-15, i.e., 1 renminbi is roughly HK$1.145; value per share is estimated on about 9.120 billion shares outstanding at the end of 2025. The results in the table are calculations by the author based on public data.

Method 2: Relative valuation Here I prioritize comparing the metrics least affected by accounting conventions, avoiding forcing pseudo-precise numbers across Chinese and US internet platforms.

How to read this: [Opinion] Tencent is cheap relative to global advertising/platform giants, and only slightly more expensive than or close to NetEase. This valuation difference is broadly reasonable: Tencent's ecosystem breadth, cash-flow resilience, and investment assets are clearly superior to NetEase's; but its Chinese regulatory/VIE/geopolitical risk is also significantly higher than Meta's or Alphabet's. In other words, Tencent is not "the most undervalued among peers," but it belongs to the category whose "quality exceeds what its valuation reflects." For Tencent itself, on audited 2025 data and the current market capitalization, a rough calculation gives P/B of about 3.15x, P/FCF of about 19.9x, and EV/Adj EBITDA of about 10.5x.

Method 3: Asset or liquidation value [Fact] As of the end of 2025, the book value of Tencent's investment portfolio was about RMB957.219 billion; within this, the fair value of listed investees was about RMB672.7 billion, the book value of unlisted investees was about RMB363.1 billion, and year-end net cash was about RMB107.1 billion. This means Tencent is not a "hollow advertising/gaming shell"; it also owns a sizable pool of non-operating or semi-operating assets.

To avoid treating these assets as 1:1 cash, I present a discounted asset-value range:

Final combined intrinsic-value range After summing the "operating-business DCF" with the "discounted investments/net cash," I present the following total valuation range:

• Conservative intrinsic-value range: HK$360 - HK$430/share

• Fair intrinsic-value range: HK$430 - HK$560/share

• Optimistic intrinsic-value range: HK$560 - HK$640/share

[Opinion] At the current HK$456.4, Tencent sits roughly at "near the lower bound of fair value to the midpoint": relative to conservative value, there is no margin of safety; relative to fair value, there is only a small-to-moderate discount; relative to optimistic value, the discount is obvious. So I cannot define today's Tencent as "significantly undervalued"; a more accurate description is "a good company at a decent price, but not yet juicy enough."

Ideal buy/hold/overvalued ranges: [Opinion]

• Ideal buy price: HK$340 - HK$400

• Acceptable holding price: HK$400 - HK$560

• Clearly overvalued price: above HK$600

This is not fortune-telling the share price, but rather working "conservative and fair value" backward into buying-and-selling discipline: I want to step up buying when there is a further 15%-25% discount below the lower bound of fair value.

Margin of Safety, Risks, and Comparisons

Is the current price cheap enough: [Opinion] Not really. It is enough for me to keep holding, but not enough for me to add aggressively at a "Buffett-style high-conviction position." The most fragile valuation assumption is that I treat a significant portion of the 2024-2026 heavy AI investment as "growth CAPEX" rather than a permanently elevated maintenance CAPEX; if this judgment is wrong, Tencent's true free-cash-flow capacity will be lower than I estimate.

The most important risks and counterpoints:

The risk judgments in the table above correspond mainly to Tencent's disclosed AI investment, VIE structure, segment revenue mix, and industry maturity, combined with Reuters' reporting on its Q1 2026 results and the AI competitive environment.

The strongest bear case: [Opinion] Those bearish on Tencent would most likely argue as follows: Tencent used to be a "low-capital-intensity traffic + content + payment platform," but now it has no choice but to join the AI arms race; China's internet penetration is already very high, and regulation and geopolitics demand a higher discount; and Tencent's VIE structure means what you hold is not the most direct ownership of Chinese operating assets. If these concerns materialize at the same time, Tencent would turn from a "high-quality compounding machine" into a "low-growth, capital-intensive asset with decent profits but that can never command a high valuation." This bear logic is not absurd.

The largest scenario of permanent capital loss: [Opinion] I believe the worst permanent loss is not a short-term 30% share-price drop, but the following combination: regulation materially weakening economic rights + AI investment returns falling short over the long term + the three major cash-flow pillars of WeChat advertising/payments/gaming simultaneously entering low growth. In this case, Tencent could be priced by the market for the long term at "low-growth Owner Earnings of around 10x," plus an asset discount, with a permanent loss range potentially reaching 40%-60%; under an extreme institutional tail risk, the loss could be even higher.

Comparison with the index, bonds, and other opportunities: [Fact] On 2025 free cash flow, Tencent's current FCF yield is about 5.0%; on my conservatively estimated Owner Earnings, the yield is about 6.2%. Over the same period, China's 10-year government bond yield was about 1.75%, and the US 10-year government bond about 4.595%. So Tencent carries a clear premium over China's risk-free rate, while its premium over US long bonds is not especially wide.

[Opinion] Is buying Tencent clearly better than buying the index? Not obviously. For most global investors, buying a broad-based index carries lower institutional risk and a more stable portfolio; but if you specifically want to allocate to China's strongest platform asset and can tolerate regulatory and single-company risk, Tencent has more quality than "just buying any basket of Chinese internet stocks."

If I could hold only 5 assets, does it qualify for the portfolio? [Opinion] For investors who can tolerate Chinese policy and geopolitical volatility, it qualifies to compete for a seat in the portfolio; for investors who only want to hold businesses with "an extremely clear institutional environment and an extremely simple business model," not necessarily.

Investment Checklist:

The key conclusion of this checklist is: business quality mostly passes, price discipline does not.

Limitations of the data: I have tried to use official disclosures and verifiable public sources, but several limitations remain: first, different platforms have slightly different conventions for the current "TTM P/E / market cap / share count"; second, P/B, EV/EBITDA, and ROIC are not fully comparable between Tencent and overseas tech stocks, because the conventions for investment assets, fair value, and associates differ greatly; third, the consistent full-year FCF for 2021 was not presented as standardized in the official summaries I directly verified this time as it was for 2022-2025, so I did not force it into the table. For these items, I choose to state them clearly rather than feign precision.

Final Investment Conclusion

[Final Rating] Hold

[One-Sentence Investment Thesis] [Opinion] Tencent remains a high-quality, deep-moat, strong-cash-flow leading Chinese platform, but its current price is closer to "fair and holdable" than to "a heavy-position buy point with ample margin of safety."

[Core Bull Case]

• The relationship chain, payments, Mini Programs, Channels, and advertising loop of Weixin/WeChat form extremely deep network effects and channel advantages that are very hard to replicate.

• In 2025, revenue, gross profit, operating profit, and free cash flow all rose together, and in the first quarter of 2026, even with CAPEX continuing to rise, FCF still grew 20% year over year, showing the cash engine remains robust.

• The balance sheet is robust, with year-end net cash of about RMB107.1 billion, while it also holds large-scale listed/unlisted investment assets, providing an "implied asset cushion" for the valuation.

• Long-term capital allocation is more shareholder-friendly than many platform companies: substantial buybacks, higher dividends, and the in-kind distribution of JD/Meituan all show management does not pursue scale alone.

• Relative to global platform leaders such as Meta and Alphabet, Tencent's current P/E is clearly lower, reflecting part of the China discount but also leaving medium-to-long-term return upside.

[Core Bear Case]

• Tencent has entered a phase of heavy AI capital expenditure; if CAPEX stays normalized at a high level in the future, both free cash flow and the valuation midpoint could be depressed.

• Chinese regulation, the VIE structure, and geopolitical risk are not market volatility but long-term sources of discount at the level of ownership and the valuation system.

• China's internet user base is approaching maturity, with subsequent growth depending more on ARPU and monetization efficiency than on a user dividend.

• The 2025 average buyback price was higher than the current share price, showing that although capital allocation is decent overall, it is not always executed in an "obviously undervalued range."

• The business is very good, but the structure is not simple: the investment portfolio, fair-value volatility, associates, regulation, and AI investment keep it from being among the "clearest, simplest" value-investing targets.

[Key Assumptions]

• The WeChat ecosystem can continue to sustain rising user time and commercialization efficiency.

• A significant portion of the 2024-2026 AI CAPEX is growth-oriented rather than permanent maintenance spending.

• Evergreen games and new products can keep overall gaming cash flow stable.

• Regulation will not undergo changes that materially damage economic rights.

• Capital-return policy will not deteriorate significantly, and the share count can keep falling or staying stable.

[Fair Buy Price] [Opinion] HK$340 - HK$400/share. The rationale is that this range gives you a more meaningful margin of safety on top of the lower bound of the fair intrinsic value I present, and can also hedge the uncertainty of AI and regulation.

[Target Holding Horizon] 10+ years. Tencent will be affected in the short term by regulation, the macro environment, AI, and market-style swings, but to capture the compounding of the relationship chain, advertising technology, gaming IP, the payment ecosystem, and cloud/AI services, one must hold it as a long-term business owner.

[Expected Annualized Return]

• Conservative scenario: 5% - 7%

• Neutral scenario: 8% - 11%

• Optimistic scenario: 12% - 15%

These are not share-price forecasts, but rough estimates based on the current Owner Earnings yield, long-term growth, capital returns, and the possibility of valuation convergence.

[Maximum Loss Risk] [Opinion] If it is merely deteriorating market sentiment, a periodic share-price drop of 30%-50% would not be surprising; if the permanent-loss scenario I described earlier of "deterioration in regulation/VIE + low returns on heavy AI investment + stalling of the core cash-flow businesses" occurs, then permanent capital loss could be in the range of 40%-60%, and under an extreme institutional tail even higher.

[Tracking Metrics]

• Changes in the combined MAU and user time of Weixin/WeChat.

• The revenue growth of Channels, WeChat Search, and closed-loop advertising.

• The revenue of domestic games, international games, and the number/grossing performance of evergreen games.

• The revenue growth and gross margin of fintech and business services.

• Annual and quarterly changes in CAPEX, FCF, and net cash.

• Changes in the share count, buyback amounts, average buyback price, and total dividends.

• The share of equity-compensation expense and whether it is offset by buybacks.

• The fair value and impairment of the listed/unlisted investment portfolio.

• Regulatory and VIE-related disclosures.

• The commercialization progress of AI-related products, not just user growth.

[Signals That Trigger Reassessment]

• Free cash flow stagnates or declines for 2-3 consecutive years while CAPEX stays high for the long term.

• Advertising growth slows significantly while gross margins no longer expand.

• The grossing of core games and the relay of new products clearly malfunction.

• Management stops large-scale buybacks or the share base resumes continuous expansion.

• Regulation materially tightens on payments, content, gaming, or the VIE structure.

• Large-scale impairments on the investment portfolio begin to eat into the ability to return capital and cash.

[Final Recommendation] [Opinion] If you already hold Tencent, I lean toward continuing to hold, and shifting attention from the short-term share price to "free cash flow, the return on AI investment, WeChat-ecosystem monetization, and capital-return discipline." If you do not yet hold it, I would suggest patiently waiting for a better price or building a position in tranches, rather than rushing to buy just because "it is a good company." Tencent meets most of the conditions for being "worth owning for the long term," but at today's price it looks more like a stock worth holding, worth watching, and worth waiting for better odds than one you must overweight immediately.

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