BUY at $75.03 per ADS, the November 16, 2026 12:20 p.m. Eastern reference quote. Base present fair value is $112.66, with a $65.07–$191.98 conditional scenario range. This is an investment judgment about price and prospective cash flows, not a prediction of the next trading session. The price is an intraday observation preserved in the source cache, not the closing price. Our research cut-off is November 16 and the quarter ended September 30.
| Metric | Value |
|---|---|
| Bear | 65.07 |
| Base | 112.66 |
| Bull | 191.98 |
| Scenario | Probability | Present value | vs $75.03 | What has to happen | The arithmetic |
|---|---|---|---|---|---|
| Bear | 25% | $65.07 | −13% | Weak cloud returns and faster ad erosion | 2036 cloud revenue RMB 60bn and 10% operating margin. |
| Base | 50% | $112.66 | +50% | Cloud scaling with continued ad decline | 2036 cloud revenue RMB 130bn and 18% operating margin. |
| Bull | 25% | $191.98 | +156% | Better utilization and monetization | 2036 cloud revenue RMB 210bn and 24% operating margin. |
| Subjective 25/50/25 weights; conditional range is not a confidence interval or loss floor. | |||||
| Signpost | Now (Q3 2026) | Green if | Red if | Next check |
|---|---|---|---|---|
| Cloud economics | RMB 8.2bn Q3 | Growth plus better cash conversion | Revenue without returns | Next results; date unannounced |
| Advertising | RMB 11.9bn Q3 | Decline moderates | Faster than model decline | Next results |
| M100 | Expected volume readiness | Commercial readiness | Delay or weak economics | End 2026 |
| Dividend | Conditional first declaration | Cash reaches shareholders | Unclear allocation or delay | Before end 2026 expected |
| Cloud revenue | Author 44bn annual base | Scale with profitability | Growth misses or excess capex | FY2027 |
Distinguish management milestones from author model checkpoints.
BUY at $75.03 per ADS, the November 16, 2026 12:20 p.m. Eastern reference quote. Base present fair value is $112.66, with a $65.07–$191.98 conditional scenario range. This is an investment judgment about price and prospective cash flows, not a prediction of the next trading session. The price is an intraday observation preserved in the source cache, not the closing price. Our research cut-off is November 16 and the quarter ended September 30.
Baidu's AI cloud operation is finally large enough to influence the group. General Business returned to growth even though advertising remains under considerable pressure. The key economic question has moved from whether AI can generate meaningful revenue to whether infrastructure spending can earn adequate returns for common shareholders. That distinction determines the model. We forecast separate cloud, marketing, application and residual-business paths, then charge the costs of equipment, working capital and compensation.
The current price appears attractive against our base assumptions, but confidence is limited by undisclosed cloud profitability and the difficulty of valuing cash and investments held across a consolidated group. The $65 bear value is not a loss limit. A severe impairment, poor allocation or constraints on cash access could produce a lower result. Investors should judge the thesis through operating evidence rather than the AI label or a large gross cash number.
Revenue was RMB 30.807bn in Q3 2026 versus RMB 31.174bn a year earlier and RMB 31.325bn in Q2. General Business contributed RMB 24.973bn versus RMB 24.659bn a year earlier and RMB 25.183bn sequentially. The change is meaningful but small: the release's rounded 1% growth should not be presented as rapid group expansion. iQIYI contributed RMB 6.007bn, down from RMB 6.682bn. Consolidated eliminations explain why business revenues do not simply add to the group figure.
Operating income was RMB 626m, a 2.0% margin. Adjusted operating income was RMB 1.712bn, down from RMB 3.785bn sequentially and RMB 2.205bn a year earlier. The GAAP year-over-year change is distorted by a RMB 16.190bn long-lived-asset impairment in Q3 2025. Removing that single prior charge does not establish that all remaining accounting adjustments are economically irrelevant.
Net income attributable to Baidu was RMB 1.395bn and adjusted net income RMB 2.395bn. Investment returns, currency movements and taxes affect these bottom-line numbers. Our enterprise valuation therefore starts with operating economics rather than capitalizing the reported earnings-per-ADS figure. All quarterly results are unaudited. The detailed AI revenue categories are drawn from internal management accounts, a disclosure qualification that also belongs beside any chart using them. Rounding of those categories prevents an exact reconstruction of every consolidated line.
| Metric | Value |
|---|---|
| Q3 2025 | 31,174.00 |
| Q2 2026 | 31,325.00 |
| Q3 2026 | 30,807.00 |
| Metric | Value |
|---|---|
| Q3 GAAP EBIT | 626.00 |
| Q3 adjusted EBIT | 1,712.00 |
| Q2 adjusted EBIT | 3,785.00 |
AI Cloud Infra revenue increased from a rounded RMB 4.2bn in Q3 2025 to RMB 7.3bn in Q2 2026 and RMB 8.2bn in Q3. GPU Cloud revenue rose 191% year over year and 43% sequentially. External customer token-usage revenue on Qianfan increased 1,159% year over year. The last figure can be large partly because of its starting base; the release does not disclose the absolute amount, and we do not convert it into an invented revenue contribution.
Management reports a more diversified customer base, increased customer prepayments and a shift toward longer-term contracts. These observations can reduce some demand and collection uncertainty. They do not reveal enforceable backlog, cancellation terms, concentration percentages, utilization or profit by contract. Consequently, no purported backlog coverage ratio appears in our valuation.
The economic chain is demand, capacity utilization, price per unit, hardware and energy cost, overhead, replacement spending and finally owner cash. Strong demand can coexist with disappointing shareholder returns if pricing falls as quickly as costs, or if capacity becomes obsolete before it earns back its cost. Baidu's full-stack approach could improve integration and supply options, but its value must show up in this chain. We model explicit margin and equipment paths rather than assign a premium merely for controlling more layers of technology.
| Metric | Value |
|---|---|
| Q3 2025 | 4,200.00 |
| Q2 2026 | 7,300.00 |
| Q3 2026 | 8,200.00 |
Online marketing revenue fell from RMB 15.3bn to RMB 11.9bn year over year, a 22% decline on reported rounded figures. AI-native marketing is a subset of that exposure rather than an extra independent revenue stream to add to all advertising. Our model uses online marketing as the forecast bucket and does not double-count AI-native marketing. Legacy-business revenue, a separate management presentation, fell 24% to RMB 9.4bn. These alternative taxonomies are informative but should not be summed indiscriminately.
Applications generated RMB 2.6bn, approximately flat year over year. Baidu reported higher AI feature penetration across Wenku and Drive, and lower digital-human production costs, but neither metric supplies an undisclosed cash margin or paid-customer conversion rate. The model allows applications to improve over time while recognizing that usage and monetization differ.
The residual forecast contains remaining General Business activity, iQIYI and consolidation eliminations. Apollo Go has an expanding geographic footprint, but the release does not disclose sufficient unit economics for a credible separate robotaxi valuation. We therefore do not add a speculative stand-alone robotaxi sale value on top of the operating model. Positive optionality may be captured in the bull residual-business path; execution costs and weak commercialization remain in lower scenarios. This avoids counting an exciting project twice while ignoring the investment required to develop it.
| Metric | Value |
|---|---|
| Q3 2025 | 15,300.00 |
| Q2 2026 | 13,100.00 |
| Q3 2026 | 11,900.00 |
| Metric | Value |
|---|---|
| Cloud infrastructure | 8,200.00 |
| Applications | 2,600.00 |
| AI-native marketing | 2,400.00 |
Cost of revenue increased to RMB 19.892bn from RMB 18.315bn, while consolidated sales declined. The calculated gross margin was 35.4%. Management attributes higher costs partly to cloud activity and traffic acquisition, offset in part by lower content costs. This mix shift is central to the analysis: an infrastructure revenue dollar can carry a different cost and reinvestment profile from an advertising dollar.
Selling, general and administrative expenses were RMB 5.460bn and research and development RMB 4.829bn. Expected credit losses helped explain the sequential SG&A increase. That line is a reminder that booked revenue and collectible cash need separate scrutiny. We do not assume the quarter's credit expense repeats at exactly the same amount forever, but the risk belongs in sustainable margins and collection assumptions.
Share-based compensation totaled RMB 1.030bn, including RMB 601m in research and development. It is noncash in the period but transfers economic value. Our modeled operating margins retain the economic cost of compensation. The model does not add back SBC and then present the resulting figure as free cash to current owners. It also avoids imposing another generic future dilution haircut for the same compensation. Existing diluted ownership is considered separately, with explicit limitations about the quarterly share proxy and the absence of an instrument-level option and convertible schedule.
| Metric | Value |
|---|---|
| Cost of revenue | 19,892.00 |
| SG&A | 5,460.00 |
| R&D | 4,829.00 |
| SBC within costs | 1,030.00 |
Q3 operating cash was RMB 2.778bn and capital expenditures RMB 7.349bn. Reported free cash flow was negative RMB 4.571bn. The comparable figures were negative RMB 2.145bn in Q3 2025 and negative RMB 7.954bn in Q2 2026. Sequential improvement is not the same as positive owner cash, and the year-over-year comparison shows a larger outflow despite cloud revenue growth.
Our model does not use quarterly free cash flow multiplied by four as a forward forecast. The quarterly run rate is useful context for scale, but prepayments, taxes, equipment deliveries and collection timing can be uneven. We explicitly forecast annual capex and depreciation, with working capital equal to 3% of incremental revenue. A declining sales scenario can release working capital; that is a modeled consequence, not a reported promise.
Base annual capex begins at RMB 21bn in 2027, rises to RMB 22bn, and remains about RMB 21bn in 2036. Depreciation rises as the equipment base matures. These assumptions allow cash conversion to improve without declaring that capital spending disappears after the initial build. The capex schedule is an author estimate, because the release does not provide numerical annual capital-spending guidance. Early forecast cash burn is deducted in the explicit valuation. The separate operating cash floor funds ongoing liquidity needs and is not a second reserve for those same forecast losses.
| Metric | Value |
|---|---|
| Operating cash | 2,778.00 |
| Capex outlay | −7,349.00 |
| Free cash flow | −4,571.00 |
| Metric | Value |
|---|---|
| Q3 2025 | −2,145.00 |
| Q2 2026 | −7,954.00 |
| Q3 2026 | −4,571.00 |
Because Baidu holds substantial liquid and investment assets and operates capital-intensive infrastructure, the independent cross-check examines common tangible book value. Start with reported Baidu shareholders' equity of RMB 273.487bn, subtract RMB 36.966bn goodwill and RMB 6.267bn intangible assets, and divide by 342.875m ADS and RMB 6.7045 per dollar. The result is $100.16 per ADS.
This method uses accounting asset and liability evidence rather than our future revenue, margin or terminal assumptions. It therefore provides a genuinely different perspective on what the share price represents. However, it is not a liquidation appraisal, a transaction price or evidence that all tangible assets earn their cost of capital. Equipment can become obsolete; receivables can fail to collect; investments can be illiquid; and restrictions can prevent immediate distribution. Tangible book also embeds historical accounting judgments.
The asset result sits below our base DCF because the cash model assigns value to future profitable operations beyond book capital. Its existence challenges a claim that the price requires extraordinary future success, but it does not independently confirm our 18% cloud margin. We do not average $100.16 and $112.66, nor choose whichever gives the desired rating. If assets are impaired or the operating returns fail, both measures can deteriorate. The cross-check is expressly limited by the quality, accessibility and productivity of the underlying assets.
| Metric | Value |
|---|---|
| Common book equity | 273,487.00 |
| Less goodwill | −36,966.00 |
| Less intangibles | −6,267.00 |
| Tangible common book | 230,254.00 |
At September 30, cash and equivalents were RMB 31.013bn, short-term investments RMB 155.331bn and long-term time deposits and held-to-maturity investments RMB 47.357bn. Together they total RMB 233.701bn. Restricted cash of RMB 199m is excluded. We separately include RMB 43.127bn of long-term investments at reported carrying value. This is not an assertion that every investment can immediately be sold for that amount.
Interest-bearing debt totals RMB 105.149bn from the disclosed loans, notes and convertible balances. Redeemable noncontrolling interests are RMB 12.699bn and ordinary noncontrolling interests RMB 9.223bn. Because the forecast covers consolidated operations, the valuation subtracts these minority claims once. Their carrying amounts are proxies for common-equity allocation, not independently appraised fair values.
The bridge also retains RMB 20bn as required operating liquidity and deducts RMB 3bn of estimated cash burn between September 30 and the November 16 valuation date. The result is RMB 126.757bn added to operating enterprise value. The estimate is transparent: no November 16 company balance sheet has been reported. A further 45-day forecast stub covers the period after the valuation date through year-end. The two periods do not overlap. We exclude investment interest income from enterprise cash flow because the underlying assets have already been added separately. That prevents adding both the investment principal and a perpetuity of its returns.
| Metric | Value |
|---|---|
| Liquid assets | 233,701.00 |
| Strategic investments | 43,127.00 |
| Debt | −105,149.00 |
| Minority claims | −21,922.00 |
| Operating floor | −20,000.00 |
| Interim burn estimate | −3,000.00 |
| Net bridge | 126,757.00 |
Current interest-bearing debt is RMB 31.418bn, comprising short-term loans, current long-term loans, current notes and current convertible debt. Noncurrent debt is RMB 73.731bn. The quarterly release supplies these classifications but not a full instrument-by-instrument face-value and maturity schedule. We do not invent dates or silently call carrying values principal. The comparison with reported liquid assets indicates substantial consolidated liquidity, although subsidiary restrictions and distribution mechanics can matter to common shareholders.
Lease liabilities total RMB 9.454bn. Rental and equipment-use costs remain within our economic operating assumptions; we do not subtract the lease liability again while also charging all future lease economics in operating cash. The model is not an EBITDA multiple that ignores replacement capex. Its lease treatment is a normalization assumption, and the detailed finance-lease funding schedule remains a source limitation.
The release reports 2,724m basic and 2,743m diluted weighted-average ordinary shares. One ADS represents eight ordinary shares, giving a 342.875m diluted ADS proxy. This is not a latest spot fully diluted cap table. Basic weighted shares were 2,713m a year earlier despite US$373m of repurchases since the start of 2026. Therefore we assume no automatic future share-count benefit. Existing options, conversion rights, exercise proceeds and vesting timing require more granular disclosures than this exhibit provides. A 5% higher denominator reduces base value to about $107.29, illustrating the consequence without pretending the uncertainty is resolved.
October 31 ownership reconciliation. The November 4 monthly share return reports 2,209,541,178 Class A and 524,020,320 Class B ordinary shares: 341.695187m ADS-equivalent at eight ordinary shares per ADS. It separately identifies 10,172,850 repurchased ordinary shares pending cancellation. Subtracting those gives 340.423581m ADS-equivalent; this adjustment is a reconciliation, not a newly adopted model denominator. The return excludes reserved depositary shares, earlier repurchases and subsidiary-held shares specified in its notes. Our 342.875m quarterly diluted weighted-average proxy reflects a different date and basis.
Outstanding plan options total 8,295,928 ordinary shares and restricted units 93,085,256. These are gross potential awards, not all currently vested or in the money. October had no new issuance under the listed plans. Exercise prices, service conditions and compensation already charged matter; adding every award to the denominator while retaining all future economic compensation would risk duplication. The return marks parent-share convertibles not applicable.
The June 30 interim report, page 10 identifies US$873m convertible-note principal issued by iQIYI, maturing 2026–2030. These are subsidiary obligations, not automatic Baidu ADS dilution. It also identifies parent unsecured notes maturing 2027–2035 and US$2bn exchangeable bonds due 2032. Those June figures contextualize the structure; September carrying balances remain the model inputs, rather than presenting old principal as current debt.
| Metric | Value |
|---|---|
| Current debt | 31,418.00 |
| Noncurrent debt | 73,731.00 |
| Lease liabilities separately | 9,454.00 |
| Metric | Value |
|---|---|
| Q3 2025 basic | 2,713.00 |
| Q3 2026 basic | 2,724.00 |
| Q3 2026 diluted | 2,743.00 |
The main model is a component revenue-to-economic-FCFF transition DCF. A current P/E is poorly suited to the combination of a shrinking advertising business, cloud infrastructure expansion, financial investments and impairment-related comparisons. A single cloud revenue multiple would miss both the cost of equipment and the legacy cash engine. An asset model alone would miss future returns on Baidu's technology and distribution. The selected DCF directly connects those economics to owner value.
In the base, cloud revenue grows from RMB 44bn in 2027 to RMB 130bn in 2036. Advertising declines from RMB 43bn to RMB 25bn. Applications grow from RMB 11bn to RMB 20bn; the consolidated residual grows from RMB 30bn to RMB 33bn. Total revenue reaches RMB 208bn. None of these values is management guidance. The cloud path deliberately slows as the revenue base becomes larger rather than extending the latest 94% growth indefinitely.
Cloud operating margin rises from 4% to 18%, advertising from 12% to 18%, applications from 4% to 18%, and the residual from negative 1% to 3%. These component margins include allocated overhead and economic compensation and are author attributions rather than reported segment margins. Economic EBIT is the sum of revenue times margin for each component. Cash tax is 20% of positive EBIT, with no immediate tax benefit assumed on losses. FCFF equals EBIT less cash tax plus depreciation less capex less working-capital investment. The annual scenario tables show the revenue, profit, tax, reinvestment and cash-flow assumptions used in each case.
| Metric | Value |
|---|---|
| 2027 | 44,000.00 |
| 2028 | 55,000.00 |
| 2029 | 67,000.00 |
| 2030 | 80,000.00 |
| 2031 | 91,000.00 |
| 2032 | 101,000.00 |
| 2033 | 110,000.00 |
| 2034 | 118,000.00 |
| 2035 | 124,000.00 |
| 2036 | 130,000.00 |
| Metric | Value |
|---|---|
| 2027 | −2,995.16 |
| 2028 | −968.00 |
| 2029 | 2,471.00 |
| 2030 | 6,459.60 |
| 2031 | 10,830.40 |
| 2032 | 14,977.60 |
| 2033 | 18,751.00 |
| 2034 | 21,398.80 |
| 2035 | 23,344.20 |
| 2036 | 24,815.00 |
The bear case assumes weaker cloud scale and returns, persistent advertising decline and limited application monetization. In 2036, cloud revenue is RMB 60bn at 10% operating margin, advertising RMB 15bn at 6%, applications RMB 11.5bn at 6%, and the residual RMB 21.5bn around breakeven. This is a commercially viable but disappointing outcome, not liquidation or the worst conceivable loss.
The bull case reaches RMB 210bn cloud revenue at 24% operating margin, with a steadier advertising business and stronger applications. Capital spending remains substantial: annual capex reaches RMB 40bn by 2036. The model does not give higher sales a free pass on infrastructure investment. These assumptions are demanding, but they articulate what stronger utilization, commercialization and operating leverage could mean for cash rather than merely assigning a higher target price.
All scenarios use 11.5% WACC and 2.5% terminal growth. Terminal reinvestment is linked to sustainable returns on new capital of 10%, 16% and 22% for bear, base and bull. Terminal cash flow is next-year NOPAT multiplied by one minus growth divided by return on capital. The base terminal contribution is 62.8% of operating enterprise value, a meaningful source of model dependence. Scenario probabilities of 25%, 50% and 25% are subjective narrative weights, not statistical frequencies. Reported liquidity supports the modeled deficits without assumed new equity issuance; failed refinancing or inaccessible cash would require a separate financing-stress case rather than an unexplained additional discount.
Main-model present values are $65.07 bear, $112.66 base and $191.98 bull per ADS. The subjective weighted present value is $120.59. The published base fair value is the base DCF result, not the weighted value and not an average with another method. Rounding to $65, $113 and $192 on slides improves readability while the underlying record retains full precision.
The base valuation gap versus $75.03 is about 50.2%. That is not a dated investment return. For a separate December 31, 2029 horizon, we recompute the value of remaining cash flows and terminal value at that date, then add the initial claims bridge and retained cash from the intervening forecast. Net interest assumes 3% cash yield, 4.5% debt coupon and 20% cash tax, with principal refinanced unchanged. This produces approximately $71.51, $138.14 and $249.53 future prices across the scenarios.
Conditional annualized returns are approximately negative 1.5%, positive 21.6% and positive 46.9% over 3.123 years. We assume constant economic shares and exchange rate and no distributions. The company intends a first dividend declaration subject to approval, but no amount is supplied; inventing one would overstate precision. A future dividend should move value from retained cash to a separately timed distribution, not simply be added to an unchanged horizon price. The base return clears our 12% equity hurdle, while the downside case shows that patient ownership can still fail to earn an adequate return.
| Metric | Value |
|---|---|
| Bear | 65.07 |
| Base | 112.66 |
| Bull | 191.98 |
| Metric | Value |
|---|---|
| Bear | 71.51 |
| Base | 138.14 |
| Bull | 249.53 |
| Metric | Value |
|---|---|
| Bear | −1.53 |
| Base | 21.58 |
| Bull | 46.93 |
To interpret the $75.03 quote, solve for a constant shift in the entire base cloud operating-margin path, holding other revenue, costs, investment, claims and discount assumptions fixed. The solution lowers cloud margins by about 12.16 percentage points, leaving approximately 5.84% in 2036 rather than 18%. It also implies losses in earlier cloud years. This is a conditional model translation of price, not a survey of what investors actually expect.
A revenue-only reverse solution is less useful here. Even at zero modeled cloud scale, the linked reduction in equipment investment leaves a value around $77.81, above the quote. There is no nonnegative revenue-scale solution for $75.03 under those fixed assumptions. We disclose that bound rather than report a negative revenue forecast or silently force a result. Profitability is the more informative disputed driver.
The primary sensitivity varies cloud revenue scale and cloud operating margin. Incremental cloud revenue also adds capex at 15% and depreciation at 10% of the revenue change, preventing a costless growth assumption. Rate and terminal-growth sensitivities are additional context. Separately, setting strategic investments to zero removes $18.78 per ADS from value; a 10% weaker RMB reduces dollar value proportionally through the conversion rate. These stresses are not added again as generic risk discounts in the base. If cash is less accessible or minority claims are worth more than carrying amounts, the quote can be explained with less pessimism about operating margins.
| Metric | Value |
|---|---|
| Base | 18.00 |
| Quote implied | 5.84 |
The release contains no numerical next-quarter revenue, profit or capital-spending guidance. Management expects Kunlunxin M100 to be ready for volume production by the end of 2026. A production milestone could improve supply availability and economics, but readiness does not establish commercial yield, customer adoption, unit cost or cash return. Those follow-on measures matter more than treating a product announcement as immediate intrinsic value.
Baidu also expects to declare its first dividend before year-end, subject to Board approval and market conditions. This is a useful capital-allocation signal. The cash amount, payment timing and sustainability were not supplied in the exhibit, so the forecast does not include an invented yield. The reported US$373m of buybacks should be judged alongside the actual ownership denominator, rather than treated as proof that every remaining ADS represents a larger claim.
The next earnings date is not asserted here. At the next report, track cloud revenue, advertising erosion, cash collections, capex, credit-loss expenses and consolidated free cash. For 2027, the base cloud checkpoint is RMB 44bn annual revenue with improving economics; it is our benchmark, not management's promise. Evidence of larger contracts without corresponding cash returns would weaken the thesis. Disclosure of utilization, margins and sustainable distributions would make the investment case more assessable and could raise conviction even before every forecast year has passed.
Business quality, financing strength and valuation should be judged separately. Baidu retains technology capability, distribution and meaningful liquid resources. Its advertising monetization is deteriorating, cloud returns are not directly disclosed, and capital spending is substantial. Thus we describe quality as mixed and financial/model uncertainty as high even though consolidated liquidity appears ample.
Our purchase discipline seeks a 20% margin of safety below $112.66 base fair value, roughly $90.13. This is an entry threshold, not a second haircut to the valuation. The $75.03 reference quote meets it, and the separately modeled base return exceeds 12%. That supports a BUY call within the disclosed assumptions. It does not establish suitability for every portfolio or justify ignoring concentration and drawdown risk.
The thesis would weaken if advertising falls faster than our declining base path, cloud pricing prevents margin improvement, capex stays elevated without utilization, or cash and investments prove less accessible than assumed. Credit deterioration and minority leakage matter because reported growth can fail to reach common owners. We locate these risks in operating paths, investment, claims and specific stresses rather than apply repeated generic country, holding-company and execution discounts. The analysis would improve with a detailed debt and convertible schedule, current diluted cap table, subsidiary liquidity reconciliation and cloud profitability disclosure. Source limitations remain visible; they are not converted into false certainty simply to meet a publishing deadline.
| Checkpoint | Date | Evidence sought | Model connection |
|---|---|---|---|
| M100 volume readiness | End 2026 expected | Commercial readiness and usable supply | Cloud costs and capacity |
| First dividend declaration | Before end 2026 expected; conditional | Amount, timing and Board decision | Distribution versus retained cash |
| Next quarterly results | Date not announced in exhibit | Cloud sales, cash collection and capex | Cash conversion and forecast returns |
| Annual cloud scale | FY2027 author checkpoint | RMB 44bn cloud revenue with improving margins | Base operating path |
| Advertising durability | Next report and FY 2027 | Slower erosion versus Q3 decline 22% | Base advertising decline |
The primary operating source is Baidu's exact Q3 2026 SEC results exhibit, an unaudited release dated November 16 covering the quarter ended September 30. Financial amounts are in renminbi unless explicitly stated otherwise. The source's convenience conversion of RMB 6.7045 per US dollar is held fixed in our model; it is not described as a November 16 foreign-exchange quote. Each ADS represents eight Class A ordinary shares. The release's detailed AI category data come from unaudited internal management accounts.
The dated price and consensus reference come from the preserved StockAnalysis forecast snapshot. Its 30-analyst average one-year target is $131.91, with observations from different dates. We do not call that a wholly refreshed post-results consensus or compare its horizon directly with our present intrinsic value. The model's December 2029 outcome has a different date again and must retain its own label.
The annual scenario tables below show the operating assumptions and cash-flow calculations for each case. The balance-sheet bridge explains how operating value translates into common-equity value, while the sensitivity tables show how changes in cloud economics, discount rates and terminal growth affect the result. Reported facts, management expectations, consensus and author estimates are distinguished, with links to the underlying sources. At future checkpoints, compare actual cloud and advertising revenue, cash conversion, capex, share count and distributions with these forecast paths. Judge forecast errors and total returns separately: a favorable share move does not prove every operating assumption was correct, and a disappointing quarter does not alone invalidate a multiyear thesis.
| Year | Cloud | Ads | Apps | Other | Revenue | EBIT | Cash tax | D&A | Capex | WC use | FCFF | PV FCFF |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2027 | 38,000.0 | 39,000.0 | 10,000.0 | 28,000.0 | 115,000.0 | 3,060.0 | 612.0 | 12,000.0 | 19,000.0 | −246.8 | −4,305.2 | −3,809.7 |
| 2028 | 42,000.0 | 33,000.0 | 10,000.0 | 27,000.0 | 112,000.0 | 2,905.0 | 581.0 | 12,500.0 | 17,000.0 | −90.0 | −2,086.0 | −1,655.5 |
| 2029 | 45,000.0 | 29,000.0 | 10,100.0 | 26,000.0 | 110,100.0 | 3,512.0 | 702.4 | 12,000.0 | 14,500.0 | −57.0 | 366.6 | 260.9 |
| 2030 | 48,000.0 | 26,000.0 | 10,300.0 | 25,000.0 | 109,300.0 | 3,894.0 | 778.8 | 11,000.0 | 12,000.0 | −24.0 | 2,139.2 | 1,365.6 |
| 2031 | 50,000.0 | 23,000.0 | 10,500.0 | 24,000.0 | 107,500.0 | 4,560.0 | 912.0 | 10,000.0 | 10,500.0 | −54.0 | 3,202.0 | 1,833.3 |
| 2032 | 52,000.0 | 21,000.0 | 10,700.0 | 23,500.0 | 107,200.0 | 5,317.5 | 1,063.5 | 9,200.0 | 9,500.0 | −9.0 | 3,963.0 | 2,034.9 |
| 2033 | 54,000.0 | 19,000.0 | 10,900.0 | 23,000.0 | 106,900.0 | 6,114.0 | 1,222.8 | 8,500.0 | 9,000.0 | −9.0 | 4,400.2 | 2,026.4 |
| 2034 | 56,000.0 | 17,500.0 | 11,100.0 | 22,500.0 | 107,100.0 | 6,476.0 | 1,295.2 | 8,000.0 | 8,500.0 | 6.0 | 4,674.8 | 1,930.8 |
| 2035 | 58,000.0 | 16,000.0 | 11,300.0 | 22,000.0 | 107,300.0 | 6,858.0 | 1,371.6 | 7,600.0 | 8,200.0 | 6.0 | 4,880.4 | 1,807.8 |
| 2036 | 60,000.0 | 15,000.0 | 11,500.0 | 21,500.0 | 108,000.0 | 7,590.0 | 1,518.0 | 7,300.0 | 8,000.0 | 21.0 | 5,351.0 | 1,777.7 |
| Year | Cloud | Ads | Apps | Other | Revenue | EBIT | Cash tax | D&A | Capex | WC use | FCFF | PV FCFF |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2027 | 44,000.0 | 43,000.0 | 11,000.0 | 30,000.0 | 128,000.0 | 7,060.0 | 1,412.0 | 12,500.0 | 21,000.0 | 143.2 | −2,995.2 | −2,650.5 |
| 2028 | 55,000.0 | 39,000.0 | 12,000.0 | 30,000.0 | 136,000.0 | 9,090.0 | 1,818.0 | 14,000.0 | 22,000.0 | 240.0 | −968.0 | −768.2 |
| 2029 | 67,000.0 | 36,000.0 | 13,200.0 | 30,300.0 | 146,500.0 | 11,607.5 | 2,321.5 | 15,500.0 | 22,000.0 | 315.0 | 2,471.0 | 1,758.8 |
| 2030 | 80,000.0 | 33,500.0 | 14,500.0 | 30,700.0 | 158,700.0 | 14,782.0 | 2,956.4 | 16,500.0 | 21,500.0 | 366.0 | 6,459.6 | 4,123.6 |
| 2031 | 91,000.0 | 31,500.0 | 15,700.0 | 31,100.0 | 169,300.0 | 18,310.5 | 3,662.1 | 17,500.0 | 21,000.0 | 318.0 | 10,830.4 | 6,200.8 |
| 2032 | 101,000.0 | 30,000.0 | 16,800.0 | 31,500.0 | 179,300.0 | 22,222.0 | 4,444.4 | 18,000.0 | 20,500.0 | 300.0 | 14,977.6 | 7,690.7 |
| 2033 | 110,000.0 | 28,500.0 | 17,800.0 | 31,900.0 | 188,200.0 | 25,647.5 | 5,129.5 | 18,500.0 | 20,000.0 | 267.0 | 18,751.0 | 8,635.3 |
| 2034 | 118,000.0 | 27,000.0 | 18,700.0 | 32,300.0 | 196,000.0 | 28,291.0 | 5,658.2 | 19,000.0 | 20,000.0 | 234.0 | 21,398.8 | 8,838.2 |
| 2035 | 124,000.0 | 26,000.0 | 19,400.0 | 32,700.0 | 202,100.0 | 30,659.0 | 6,131.8 | 19,500.0 | 20,500.0 | 183.0 | 23,344.2 | 8,647.3 |
| 2036 | 130,000.0 | 25,000.0 | 20,000.0 | 33,000.0 | 208,000.0 | 32,490.0 | 6,498.0 | 20,000.0 | 21,000.0 | 177.0 | 24,815.0 | 8,244.1 |
| Year | Cloud | Ads | Apps | Other | Revenue | EBIT | Cash tax | D&A | Capex | WC use | FCFF | PV FCFF |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2027 | 50,000.0 | 46,000.0 | 12,000.0 | 32,000.0 | 140,000.0 | 11,320.0 | 2,264.0 | 13,000.0 | 25,000.0 | 503.2 | −3,447.2 | −3,050.4 |
| 2028 | 68,000.0 | 44,500.0 | 14,500.0 | 34,000.0 | 161,000.0 | 16,210.0 | 3,242.0 | 15,500.0 | 28,000.0 | 630.0 | −162.0 | −128.6 |
| 2029 | 88,000.0 | 44,000.0 | 17,500.0 | 36,500.0 | 186,000.0 | 22,890.0 | 4,578.0 | 18,500.0 | 31,000.0 | 750.0 | 5,062.0 | 3,603.1 |
| 2030 | 110,000.0 | 44,000.0 | 21,000.0 | 39,000.0 | 214,000.0 | 31,110.0 | 6,222.0 | 22,000.0 | 33,000.0 | 840.0 | 13,048.0 | 8,329.5 |
| 2031 | 130,000.0 | 44,500.0 | 24,500.0 | 42,000.0 | 241,000.0 | 40,505.0 | 8,101.0 | 25,000.0 | 35,000.0 | 810.0 | 21,594.0 | 12,363.3 |
| 2032 | 149,000.0 | 45,000.0 | 28,000.0 | 45,000.0 | 267,000.0 | 48,950.0 | 9,790.0 | 28,000.0 | 36,000.0 | 780.0 | 30,380.0 | 15,599.6 |
| 2033 | 166,000.0 | 45,500.0 | 31,500.0 | 48,000.0 | 291,000.0 | 55,940.0 | 11,188.0 | 30,000.0 | 37,000.0 | 720.0 | 37,032.0 | 17,054.1 |
| 2034 | 182,000.0 | 46,000.0 | 35,000.0 | 51,000.0 | 314,000.0 | 63,180.0 | 12,636.0 | 32,000.0 | 38,000.0 | 690.0 | 43,854.0 | 18,112.8 |
| 2035 | 197,000.0 | 46,500.0 | 38,000.0 | 54,000.0 | 335,500.0 | 70,500.0 | 14,100.0 | 34,000.0 | 39,000.0 | 645.0 | 50,755.0 | 18,801.0 |
| 2036 | 210,000.0 | 47,000.0 | 41,000.0 | 57,000.0 | 355,000.0 | 76,780.0 | 15,356.0 | 36,000.0 | 40,000.0 | 585.0 | 56,839.0 | 18,883.1 |
| Cloud revenue scale | Margin −3 percentage points | Base margin | Margin +3 percentage points |
|---|---|---|---|
| 80% of base | $98.26 | $105.69 | $113.11 |
| 100% of base | $103.38 | $112.66 | $121.94 |
| 120% of base | $108.49 | $119.62 | $130.76 |
| WACC | Terminal growth 1.5% | Terminal growth 2.5% | Terminal growth 3.5% |
|---|---|---|---|
| 10.5% | $120.21 | $122.64 | $125.69 |
| 11.5% | $111.11 | $112.66 | $114.53 |
| 12.5% | $103.79 | $104.77 | $105.92 |