Baidu, Inc. (BIDU): what the price assumes

boothcheck covers Baidu, Inc. (BIDU) but does not put one priced-in number on it: here the defensible answer is the evidence rather than a point estimate. boothcheck publishes no house fair value, target price, or buy/sell rating. Narrative composed 2026-07-25.

Generated: 2026-08-30 · Exported: 2026-08-31 · Source: https://boothcheck.com/report/BIDU

Headline

FieldValue
TickerBIDU
CompanyBaidu, Inc.
Sector / IndustryTechnology
Current price$97.22/sh
CompositionOnline marketing services 53% / Others 47%

What The Price Assumes (Inversion)

The assumption today's price embeds, recovered by inverting the valuation.

FieldValue
Inversion basiswhole-company
Operating margin (value-band context)18.2%
Operating margin (mid-cycle)35.0%
Margin compression (value-band)-16.8pp
Trailing margin (depressed year)-4.5%
Multiple paid5x mid-cycle operating income

The operating-margin figure is value-band context at year 12: derived from the framework's value band, a separate calculation — not part of the priced-in solve.

The price sits below what even a 5%/yr operating-profit decline would warrant; the inversion reports a bound, not a solved growth path.

Solve inputs: computed at a 8.7% cost of capital with 4% terminal growth over a 5-year stage.

How unusual the bet is: n/a

ReferenceValue
vs own history-0.56σ

Valuation X-Ray

Every valuation family lands below the price. The price therefore sits beyond what those standard frames encode.

How the valuation models price the stock relative to the market price. Price/FV above 1.0 means the market pays more than that lens defends (expensive); at or below 1.0 the lens can defend the price.

FamilyMedian price/FVModelsReads
Asset5.31x4expensive
Earnings1.60x3expensive
Relative1.72x5expensive
Growth1.63x1expensive

Families that call it expensive: Asset, Earnings, Relative, Growth

The models below discount at their own flat-beta convention rates (cost of equity 9.3%, WACC 8.2%); the inversion above states its own rate.

Per-Model Detail (n=13)

ModelFamilyFVPrice/FVApplicableMethodology
DCF Perpetual GrowthGrowth$0.00noNegative/zero FCF — equity value floored at $0
DCF Exit MultipleGrowth$0.00noNegative/zero FCF or EBITDA — equity value floored at $0
Relative ValuationRelative$56.621.72xyesP/E 35x (static sector reference · 2026-04), scenarios: 29.4x / 35.0x / 40.6x (bear / base = reference held flat / bull), EV/EBITDA 55x
Simple DDMGrowthno
Two-Stage DDMGrowthno
Simple Excess ReturnAsset$25.393.83xyesBV/sh $111.93, ROE (TTM) 2.1%, ke 9.3%
Two-Stage Excess ReturnAsset$14.326.79xyes5yr excess ROE then converge to ke=9.3%
Discounted Future Market CapGrowth$59.601.63xyesRev $18.5B, growth -1% (input: historical growth; tapered), Terminal P/S: 1.5x / 1.8x / 2.1x (bear / base = today's held flat / bull, cap 8x)
Peter Lynch Fair ValueRelative$39.712.45xyesEPS $2.35, growth 17% (input: historical EPS growth), PEG=2.45 (Overvalued)
Margin TrajectoryGrowthno
Earnings Power ValueEarnings$60.891.60xyesNormalized EBIT (5y avg op income, one-time charges added back) $2.56B × (1−19%) / WACC 8.2% → EPV (no growth)
Residual IncomeAsset$10.429.33xyesBV $111.93 + 5yr PV of (ROE (TTM) 2.1% − Kₑ 9.3%) × BV; BV grows 1.4%/yr
Graham NumberAsset$76.901.26xyes√(22.5 × EPS $2.35 × BVPS $111.93) — Graham's conservative floor
EV/EBITDA RelativeRelative$6.8014.30xyesEBITDA $0.28B × sector EV/EBITDA 25.0x
FCF YieldEarningsno
SBC-Adj FCF YieldEarningsno
Ben Graham FormulaEarnings$75.771.28xyesEPS $2.35 × (8.5 + 2×15.0%) × (4.4 / 5.3%)
ROIC-Justified P/BAssetno
P/Sales SectorRelative$433.980.22xyesRevenue $18.46B × sector P/S 8.0x
PEG Fair ValueRelative$59.571.63xyesEPS $2.35 × (PEG 1.5 × growth 16.9% (input: historical EPS growth)) → PE 25.4x
Earnings YieldEarnings$25.393.83xyesEPS $2.35 / required return 9.3% (Rf 4.3% + ERP 5.0%)
Funds From Operations MultipleRelativeno
Clinical Phase NPVGrowthno
MertonAssetno
V5 Mechanicalno

Solvency

FieldValue
Net cash$8.5b
Net debt / NOPAT (after-tax)-1.61x (net cash)
Net debt / operating income (pre-tax)-1.31x (net cash)
Interest coverage16.2x
Burning cashyes

Leverage and coverage are computed on normalized mid-cycle operating income (mid-cycle margin 35.0%); the trailing year was depressed.

Bullet Takeaways

Bull Case

The advertising decline at the center of the bearish read here is not something that happened to Baidu. It is something Baidu did. The annual report is unusually direct about it: revenue from online marketing services "declined in 2024 and 2025, primarily due to our ongoing AI transformation, which impacted the monetization" of the search experience. Rebuilding a results page around generated answers removes the blue links that carried the ads. The revenue goes first. Whatever replaces it arrives later, if it arrives.

That is an expensive way to run a search engine, and it is the kind of choice a business makes when it thinks standing still costs more. The substitution is already visible in the segment disclosure, where the core held roughly flat while its composition changed underneath: "Baidu General Business revenue was RMB104.7 billion in 2024, increasing by 1% compared to RMB103.5 billion in 2023. The increase was primarily driven by growth in cloud services, partially offset by a decline in online marketing revenue." One line went down, another came up, and the total barely moved. A shrinking advertising business inside a flat total is a different situation from a shrinking business.

What is coming up is a different kind of revenue. Auction-priced search advertising reprices every quarter with the macro mood of Chinese advertisers. Cloud sells compute, model access and deployment work under contracts, and the filing describes a stack that runs the length of a customer's project rather than a single transaction: "This enables us to support enterprises and public sector customers at every stage of their AI journey, from model training and fine-tuning to AI-native application and agent development." Selling training capacity and then the applications that run on it is stickier than selling clicks, and it grows with the customer rather than with the advertising cycle.

Sitting on top of that is the driverless business, which the market is largely not paying for. The annual report describes agreements "to deploy fully autonomous vehicles across key European markets through the Lyft platform, starting with Germany and the United Kingdom", and notes that "In October 2025, Apollo Go established a strategic partnership with PostBus, Switzerland's leading public transport operator, to launch an on-demand autonomous mobility service." Those are distribution deals for a service that has been operating commercially in China for years. Whether it becomes a real profit line is unknowable from here; what matters to the bull is that nothing in the current price appears to require it.

Then there is the price itself. Against a market value near $35.9B, the annual report states the company had "RMB115.5 billion (US$16.5 billion) in cash, cash equivalents, restricted cash" and related holdings at the end of 2025. Roughly 8 billion dollars of funded borrowings sits against that. The buyer of the stock is therefore paying a good deal less than the headline market value for the operating business, and paying it for a company whose revenue mix is moving toward the part that is growing.

The honest objection is that this transition is capital hungry and unfinished. The filing concedes the company "made significant prepayments to acquire relevant resources from third-party suppliers" to build out the AI stack, and export controls constrain what hardware it can get. Both are real. Neither is evidence the transition failed; they are the bill for attempting it, and the bill is being paid out of a cash position most companies attempting the same thing do not have.

Bear Case

Start with where the money actually sits. The ADSs are a claim on a Cayman holding company that reaches its Chinese operating businesses through contracts rather than ownership, and the annual report puts a figure on how much of the balance sheet that structure walls off. Restricted net assets, which include "the net assets of the VIEs in which the Company has no legal ownership" alongside paid-in capital and statutory reserves of the Chinese subsidiaries, came to US$5.8 billion at the end of 2025. That is money on the balance sheet that is not freely available to the entity whose shares trade in New York.

The structural risk behind that number is stated plainly in the filing. If regulators "were to find our legal structure and contractual arrangements to be in violation of any laws, administrative regulations or provisions of Chinese mainland", the consequences run to license revocation and forced restructuring, and the company notes it "would not be able to continue to consolidate the financial results of these entities in our financial statements." Deconsolidation is not a stock that trades down. It is a stock that stops describing the same business.

The valuation argument has a softer joint than it looks. The case that the shares are cheap rests on a through-cycle profit figure, not on what the business has been earning. On the trailing annual basis the operating margin is about negative 4.5%, and the filing explains why in its own words: "Our operating margin and net income attributable to us as a percentage of revenue fluctuated notably from 2021 to 2025 due to various factors such as the macroeconomic environment, the impact of the COVID-19 pandemic and impairment of long-lived assets." The normalized read assumes a through-cycle operating margin near 35%. The earnings-power method, which averages five years of operating profit and adds one-time charges back, reaches a normalized figure closer to 2.3 billion dollars. Those two views of the same "normal" are not close to each other, and everything about the cheapness claim depends on which one is right.

Meanwhile the peer set is doing what Baidu is trying to do, from a stronger position. GOOGL is running a 32.7% operating margin on revenue of $422.5B, growing 17.5%. META is at a 41.2% operating margin on revenue of $215.0B, growing 26.2%. Both are rebuilding search and feed experiences around the same generative technology, and both are funding it out of profits Baidu is not currently earning. The filing also names a slower erosion underneath: as vertical service providers expand, their "reliance on search engines may decline, especially if they can consolidate their industry verticals." Search intent leaking to destination apps is the same pressure that has been reshaping Western search, and it does not require a competitor to win a model benchmark.

The hardware constraint compounds it. The company acknowledges that "the introduction of a series of export control measures in recent years could potentially impose limitations on our access to advanced semiconductor technologies, which may hinder the process of the research and development of our AI technology and capabilities." A cloud business whose growth depends on compute it may not be able to buy has a ceiling that is not set by demand.

Even the driverless option carries policy exposure rather than pure upside. The filing flags that on "January 30, 2026, eight departments including the MIIT issued the Guidelines on Automotive Data Outbound Transfer Security (2026 Edition)", effective immediately, restricting how algorithm files and source code cross borders, with the company noting a scenario in which revenue from driverless ride-hailing declines. The European and Middle Eastern expansion is exactly the activity those rules govern.

The downside is not open-ended. Around 6 billion dollars of equity stakes sit outside the operating business, close to a sixth of the current market value, and those holdings survive an impairment of the operating thesis. That is a boundary, not a cushion. It is also a reminder that a meaningful part of what the buyer owns is not the search engine at all.

Valuation

Two readings of the same $105.37 price point in opposite directions, and the disagreement between them is the most informative thing in the file.

Run the standard methods against what the company has actually reported lately and the price looks stretched against every one of them. The peer-multiple methods land with the price roughly 86% above their central estimate. The earnings-power methods land at about half the price, so the price sits near double that family. The asset-value methods, which ask what return the company earns on its own book, land furthest away, with the price close to six times their central read. Even the forward-growth methods, the most generous family here, leave the price about 63% above where they land. No family reaches the price.

All four share one denominator problem. The trailing annual period carries an operating loss, and a loss makes every earnings-based method either uninformative or punitive. That is not a subtlety the reader should skip: it is the difference between a business that cannot earn and a business that spent a year rebuilding the thing that earns.

Invert the price against the company's own through-the-cycle margins instead and it reads the other way entirely. On that basis the market is paying roughly five times normalized operating profit. A multiple that low is not pricing growth at all. It is consistent with a profit stream that shrinks by a mid single-digit percentage every year and never recovers. That is a bound rather than a solved answer: it says the price embeds decline, and it does not say how much.

So the concrete question is not what growth rate to believe. It is which normal to believe. The normalized read assumes a through-the-cycle operating margin near 35% applied to current revenue. The five-year average of reported operating profit, with one-off charges added back, sits closer to 2.3 billion dollars. A buyer at today's price is underwriting the first figure while the recent record supports something nearer the second, and the gap between them is wider than any of the model disagreements above.

The peer group offers context rather than a verdict, because the comparable set here is whole-company rather than segment-matched. GOOGL and META are both large advertising franchises being retooled around generative models, and both currently convert that scale into operating margins in the thirties and forties while growing revenue at double digits. Baidu is attempting the same retooling from a position where the advertising line is contracting first.

The balance sheet is what makes the wait survivable. The annual report states the company had "RMB115.5 billion (US$16.5 billion) in cash, cash equivalents, restricted cash" and related holdings at the end of 2025, against funded borrowings of roughly 8 billion dollars. Set against that, a portion of those net assets is restricted inside the Chinese operating entities and the consolidated affiliates, which is where the liquidity argument and the structure argument meet.

Catalysts

The March quarter was the first period in which the transition showed up as a majority of the business. Core AI-powered revenue rose 49% year over year to RMB13.6 billion and crossed half of the core mix for the first time, while total revenue came in at RMB32.1 billion. AI cloud revenue reached RMB11.3 billion, with the infrastructure component growing 79% year over year. The advertising line remained the drag; the cloud line remained the offset. That is the same trade the annual report describes, now visible at quarterly frequency.

The driverless business scaled in the same period, with 3.2 million fully driverless rides in the quarter and operations extended to 27 cities globally, including a standalone autonomous ride-hailing app launched in Dubai. Vehicles have been positioned in London ahead of trials with Uber and Lyft, and road testing in Switzerland follows the PostBus arrangement disclosed in the annual report. Ride counts are not revenue, and the filing is explicit that the regulatory framework around autonomous vehicle data tightened in January 2026, so this line is best watched for whether unit economics follow the volume.

Second-quarter results are scheduled for August 26, 2026. The specific thing to watch is whether cloud growth stays fast enough to hold the total revenue line while advertising continues to reset, since that combination is what separates a deliberate transition from a decline with a story attached.

Peer Cohorts (Per Segment, With Filing Citations)

Core business (reported)

Methodology Note

Fundamentals sourced from SEC EDGAR filings. Current price from Databento. The priced-in inversion and valuation x-ray are computed by the boothcheck engine; narrative composed by AI from the structured data.

Sources

company earnings calendar, July 2026 · Baidu Q1 2026 results announcement, May 2026 · Baidu Q1 2026 earnings call, May 2026

View the full interactive BIDU report on boothcheck