BIDU · Forward model · Bear case
The Bear case, 20 quarters out
Each segment is projected from its own operating driver, rolled up into consolidated cash flow, and discounted back to a fair value per share. The assumptions below are editable — change them and every number on this page moves with them.
Two decisions decide how this model reads. First, the segmentation. Baidu publishes TWO overlapping cuts of Baidu General Business revenue - a Core AI-powered table (AI Cloud Infra + AI Applications + AI-native Marketing Services, plus Legacy and Others) and a reported split (Online Marketing Services plus Others). The two are the same revenue counted differently: AI-native Marketing Services sits INSIDE Online Marketing Services, and Legacy Business overlaps it too. Adding both cuts double-counts about RMB 2.6B a quarter. This model uses the Core AI-powered cut, which is additive, so 'Legacy Business' here is the traditional search-and-advertising business net of the AI-native slice already counted above it. Second, the currency. Baidu reports in renminbi and publishes a US dollar convenience translation at each quarter-end rate. Translating each quarter at its own rate puts a currency forecast inside an operating model - it is why Baidu's Q2 revenue reads +1% in dollars and -4% in renminbi. Every figure here is therefore converted at a CONSTANT RMB 6.7851 = US$1.00, the 30 June 2026 rate, so the projection contains no implicit view on the exchange rate. Prior quarters will not tie to the dollar figures Baidu published at the time. Apollo Go is not disclosed as a revenue line by Baidu and is carried here as an explicit estimate, marked as such. iQIYI is shown net of the inter-segment elimination so the verticals sum exactly to consolidated revenue. History runs six quarters, from 2025 Q1, rebuilt from three separate Baidu releases: the Q2 2026 print supplies 2025 Q2 / 2026 Q1 / 2026 Q2, the Q1 2026 print supplies 2025 Q1 and 2025 Q4, and the Q4 2025 print supplies 2025 Q3. One split is an outright estimate - the Q4 2025 release gives 2025 Q3's Core AI-powered components but not the Legacy/Others split beneath them, so Others is interpolated at RMB 2.6B and Legacy takes the residual. Finally, a definitional warning: like every model on this site, free cash flow here is EBITDA less capex less tax, NOT the operating-cash-flow-less-capex figure Baidu reports and this site stores in its series. Baidu's Q2 2026 EBITDA was about $906M against $506M of operating cash flow, so the model's free cash flow runs structurally ahead of the reported one by roughly the working-capital and interest gap. Do not read the model's Rule of 40 against the stock page's.
Q1's RMB 8.8B was the run-rate and Q2's RMB 7.3B is the trend: AI Cloud growth settles in the teens as Alibaba, Tencent and Huawei compete the price of Chinese GPU hours down. Legacy advertising keeps falling at double digits because the loss is share, not substitution. Capex stays high anyway, because stopping it concedes the market. Free cash flow never turns inside the horizon.
Latest: $5.87B (2031Q2E)
| Period | Value |
|---|---|
| 2025Q1 | $4.78B |
| 2025Q2 | $4.82B |
| 2025Q3 | $4.59B |
| 2025Q4 | $4.83B |
| 2026Q1 | $4.73B |
| 2026Q2 | $4.62B |
| 2026Q3E | $4.63B |
| 2026Q4E | $4.66B |
| 2027Q1E | $4.69B |
| 2027Q2E | $4.73B |
| 2027Q3E | $4.77B |
| 2027Q4E | $4.82B |
| 2028Q1E | $4.87B |
| 2028Q2E | $4.92B |
| 2028Q3E | $4.98B |
| 2028Q4E | $5.05B |
| 2029Q1E | $5.11B |
| 2029Q2E | $5.18B |
| 2029Q3E | $5.26B |
| 2029Q4E | $5.33B |
| 2030Q1E | $5.41B |
| 2030Q2E | $5.50B |
| 2030Q3E | $5.58B |
| 2030Q4E | $5.68B |
| 2031Q1E | $5.77B |
| 2031Q2E | $5.87B |
What drives each segment
AI Cloud Infrastructure
Growth pathThe growth engine and the reason the balance sheet looks the way it does. GPU Cloud - public-cloud AI compute sold by the hour and by contract - grew 283% year over year in Q2 2026, accelerating from 184% the quarter before, inside a segment that grew 50%. It is also the lowest-margin revenue Baidu has ever sold, and it is displacing the highest-margin revenue it ever sold. Every point of gross margin Baidu has lost since 2023 is this line arriving.
Latest: $2.81B (2031Q2E)
| Period | Value |
|---|---|
| 2025Q1 | $722M |
| 2025Q2 | $722M |
| 2025Q3 | $619M |
| 2025Q4 | $855M |
| 2026Q1 | $1.30B |
| 2026Q2 | $1.08B |
| 2026Q3E | $1.16B |
| 2026Q4E | $1.24B |
| 2027Q1E | $1.33B |
| 2027Q2E | $1.41B |
| 2027Q3E | $1.49B |
| 2027Q4E | $1.58B |
| 2028Q1E | $1.66B |
| 2028Q2E | $1.75B |
| 2028Q3E | $1.83B |
| 2028Q4E | $1.92B |
| 2029Q1E | $2.01B |
| 2029Q2E | $2.09B |
| 2029Q3E | $2.18B |
| 2029Q4E | $2.27B |
| 2030Q1E | $2.35B |
| 2030Q2E | $2.44B |
| 2030Q3E | $2.53B |
| 2030Q4E | $2.62B |
| 2031Q1E | $2.72B |
| 2031Q2E | $2.81B |
Assumptions & reasoning
- The actuals are the management-account Core AI-powered table, published to one decimal place in RMB billions and unaudited. They are marked estimated because of that rounding, not because the line is invented.
- The six-quarter path in constant dollars is the argument for treating Q1 as the outlier: 722, 722, 619, 855, 1,297, 1,076. It is not a smooth ramp. Anyone extrapolating from the 2026 Q1 peak alone is extrapolating from the highest point in a jagged series.
- The awkward fact this line has to carry: revenue fell 17% SEQUENTIALLY, from RMB 8.8B to RMB 7.3B, in the same quarter GPU Cloud inside it grew 283% year over year. Baidu offers no reconciliation. The likeliest reading is that Q1 contained lumpy, non-recurring infrastructure contracts and Q2 is the cleaner run-rate. If instead Q1 is the run-rate and Q2 is the miss, this whole model is too conservative.
- 9% a quarter compounds to roughly 41% in year one and decays toward 4%, landing the line near 17% annual growth in the terminal year. That is deliberately below the 50% just printed: a line growing 283% off a small base cannot hold that rate as the base becomes the segment.
- Capex intensity starts at 125% of this line's revenue - Baidu spent RMB 11.4B of capex against RMB 7.3B of AI Cloud Infra revenue in Q2 - and glides to 30%. That single assumption drives more of the fair value than the growth rate does. Hold it at 125% throughout and free cash flow never turns.
- EBITDA margin starts at 12% and glides to 30% on utilisation and on Kunlun silicon displacing bought accelerators. If Baidu cannot get its own chips into the fleet at scale, the terminal margin belongs nearer 20%.
AI Applications
Growth pathConsumer and productivity AI that Baidu already charges for: Wenku, Baidu Drive, the DuMate agent, and the AI features layered onto them. It grew 3% year over year in Q2 2026, which is the quietest number in the AI table and the one that decides whether Baidu owns an AI platform or merely rents out the compute underneath other people's.
Latest: $495M (2031Q2E)
| Period | Value |
|---|---|
| 2025Q1 | $368M |
| 2025Q2 | $368M |
| 2025Q3 | $383M |
| 2025Q4 | $398M |
| 2026Q1 | $368M |
| 2026Q2 | $368M |
| 2026Q3E | $375M |
| 2026Q4E | $381M |
| 2027Q1E | $388M |
| 2027Q2E | $394M |
| 2027Q3E | $400M |
| 2027Q4E | $407M |
| 2028Q1E | $413M |
| 2028Q2E | $419M |
| 2028Q3E | $425M |
| 2028Q4E | $431M |
| 2029Q1E | $438M |
| 2029Q2E | $444M |
| 2029Q3E | $450M |
| 2029Q4E | $456M |
| 2030Q1E | $463M |
| 2030Q2E | $469M |
| 2030Q3E | $476M |
| 2030Q4E | $482M |
| 2031Q1E | $488M |
| 2031Q2E | $495M |
Assumptions & reasoning
- RMB 2.5B for three straight quarters. Rounded to one decimal place, this line has not visibly moved in a year, and Baidu's own YoY figure of +3% confirms it is close to flat.
- The one disclosed leading indicator is engagement, not money: AI DAU penetration across Wenku and Drive rose 27.4% year over year in June 2026. Usage is growing roughly nine times faster than revenue, which is either a monetisation opportunity or a monetisation problem.
- 3% a quarter is about 13% a year - ahead of the 3% just printed, on the view that agent products launched in mid-2026 start converting. If they do not, this line is worth roughly what it is worth today and nothing more.
- High margin and almost no capex: this is software running on infrastructure already paid for by the cloud line above.
AI-native Marketing Services
Growth pathAdvertising sold through AI formats rather than the classic search-results page - the part of the marketing business Baidu counts as reborn rather than declining. It was flat year over year and up 11% sequentially in Q2 2026. It is the bridge line: every dollar it gains is a dollar Legacy does not have to lose.
Latest: $447M (2031Q2E)
| Period | Value |
|---|---|
| 2025Q1 | $251M |
| 2025Q2 | $383M |
| 2025Q3 | $413M |
| 2025Q4 | $398M |
| 2026Q1 | $339M |
| 2026Q2 | $383M |
| 2026Q3E | $386M |
| 2026Q4E | $389M |
| 2027Q1E | $392M |
| 2027Q2E | $395M |
| 2027Q3E | $398M |
| 2027Q4E | $401M |
| 2028Q1E | $404M |
| 2028Q2E | $408M |
| 2028Q3E | $411M |
| 2028Q4E | $414M |
| 2029Q1E | $417M |
| 2029Q2E | $420M |
| 2029Q3E | $424M |
| 2029Q4E | $427M |
| 2030Q1E | $430M |
| 2030Q2E | $434M |
| 2030Q3E | $437M |
| 2030Q4E | $440M |
| 2031Q1E | $444M |
| 2031Q2E | $447M |
Assumptions & reasoning
- This line sits INSIDE Baidu's reported Online Marketing Services revenue. It is separated here because the Core AI-powered cut separates it; do not add it to a figure taken from the reported revenue split or you will count it twice.
- Flat year over year but +11% quarter over quarter, and the reported Online Marketing Services total also rose 4% sequentially - the first sequential increase in the tracked window. Two quarters of that and the advertising bottom is real.
- 2% a quarter, about 8% a year, is a deliberately unheroic number for a line that has not grown year over year. The Bull case is that AI ad formats price better than the blue links they replace; nothing in the disclosures proves that yet.
- Margin is set near the old search business because that is economically what it is - inventory Baidu already owns, sold through a new interface.
Legacy Business
Growth pathThe old Baidu: classic search advertising, feed monetisation and the non-AI mobile ecosystem. It fell 23% year over year in Q2 2026 to RMB 10.4B and is now smaller than the AI half for the first time. This is the melting ice cube, and how fast it melts decides whether the company's total revenue ever grows again.
Latest: $872M (2031Q2E)
| Period | Value |
|---|---|
| 2025Q1 | $2.11B |
| 2025Q2 | $2.00B |
| 2025Q3 | $1.84B |
| 2025Q4 | $1.81B |
| 2026Q1 | $1.50B |
| 2026Q2 | $1.53B |
| 2026Q3E | $1.47B |
| 2026Q4E | $1.41B |
| 2027Q1E | $1.36B |
| 2027Q2E | $1.31B |
| 2027Q3E | $1.27B |
| 2027Q4E | $1.23B |
| 2028Q1E | $1.19B |
| 2028Q2E | $1.16B |
| 2028Q3E | $1.13B |
| 2028Q4E | $1.10B |
| 2029Q1E | $1.07B |
| 2029Q2E | $1.04B |
| 2029Q3E | $1.02B |
| 2029Q4E | $993M |
| 2030Q1E | $970M |
| 2030Q2E | $949M |
| 2030Q3E | $928M |
| 2030Q4E | $909M |
| 2031Q1E | $890M |
| 2031Q2E | $872M |
Assumptions & reasoning
- Down 23% year over year - but UP 3% sequentially, off Q1's low. One quarter is not a bottom; it is the first thing a bottom looks like.
- 2025 Q3 is the one estimated split in this model: that quarter's release published the Core AI-powered components but not the Legacy and Others lines beneath them, so Others is interpolated at RMB 2.6B and Legacy takes the residual, RMB 12.46B. Every other quarter's Legacy figure is Baidu's own.
- Modelled at -3% a quarter decaying toward -0.5%: roughly -11% in year one, easing to about -2% a year in the terminal year. That is materially kinder than the -23% just printed, on the view that the AI-native line is cannibalising this one internally rather than the whole advertising pool shrinking at 23% forever.
- The single most important interaction in this model is between this line and AI-native Marketing. If the decline here is genuine substitution, the two roughly net; if it is share loss to Tencent, ByteDance and Alibaba, the AI line does not catch it and total revenue keeps falling.
- Margin starts at 36% and DECLINES to 30%. A shrinking ad business loses operating leverage; it does not gain it. This is the only vertical here whose margin gets worse.
- Little capex. The infrastructure this business runs on was built years ago and is being repurposed for the cloud line.
Apollo Go (Robotaxi)
Units × priceThe long-duration option. 28 cities, over 350 million cumulative autonomous kilometres and 240 million of them fully driverless, live commercial driverless service in Dubai, open-road testing in London with Uber and Lyft, and the first fully driverless permits in a right-hand-drive market. It is also, on any honest estimate, well under half a percent of revenue today.
Latest: $183M (2031Q2E)
| Period | Value |
|---|---|
| 2025Q1 | $4M |
| 2025Q2 | $7M |
| 2025Q3 | $10M |
| 2025Q4 | $12M |
| 2026Q1 | $12M |
| 2026Q2 | $14M |
| 2026Q3E | $16M |
| 2026Q4E | $18M |
| 2027Q1E | $21M |
| 2027Q2E | $24M |
| 2027Q3E | $27M |
| 2027Q4E | $31M |
| 2028Q1E | $35M |
| 2028Q2E | $40M |
| 2028Q3E | $46M |
| 2028Q4E | $52M |
| 2029Q1E | $59M |
| 2029Q2E | $67M |
| 2029Q3E | $76M |
| 2029Q4E | $86M |
| 2030Q1E | $98M |
| 2030Q2E | $111M |
| 2030Q3E | $126M |
| 2030Q4E | $142M |
| 2031Q1E | $161M |
| 2031Q2E | $183M |
Assumptions & reasoning
- Baidu does NOT disclose Apollo Go revenue. Every figure in this vertical is an estimate and should be treated as one. It is carved out of the Others line of Baidu General Business, where it actually sits, so carving it out does not add revenue to the model.
- The estimate is rides multiplied by an assumed fare. Disclosed ride counts: 3.4M fully driverless rides in Q4 2025 and 3.2M in Q1 2026, with weekly rides peaking above 350,000 in March 2026. Q2 2026 is estimated at 3.8M because Baidu stopped publishing the ride count this quarter and gave only cumulative kilometres - which rose about 20 million in the quarter.
- That Baidu dropped the quarterly ride count in the same quarter it added Dubai, Hong Kong and London is worth noticing. A company whose robotaxi rides were compounding usually keeps printing the number.
- Fare is assumed at US$3.70 a ride, blended. Chinese robotaxi fares are heavily discounted - materially below a human-driven ride - and the international cities price far higher. The fare drifts UP 1.5% a quarter as the international mix grows, which is the opposite of most ride-hailing models and is the whole reason the line is interesting.
- Volume compounds at 14% a quarter against a ceiling of 60 million rides a quarter. That reaches roughly 52 million rides and about $259M of quarterly revenue in the terminal year - around 4% of company revenue. Anyone modelling Apollo Go as a material part of Baidu's value inside five years is making an argument this model does not make.
- Capex intensity starts at 200% of the line's revenue: every vehicle is bought before it earns. It glides to 40%, which assumes the fleet reaches the point where replacement rather than expansion dominates.
Other Baidu General Business
Growth pathEverything else inside Baidu General Business once the AI table, Legacy and Apollo Go are taken out: intelligent-driving licensing to automakers, Xiaodu devices, maps, and the smaller mobile-ecosystem lines. RMB 2.3B a quarter, down 15% year over year, and mostly a rounding item next to the rest.
Latest: $357M (2031Q2E)
| Period | Value |
|---|---|
| 2025Q1 | $300M |
| 2025Q2 | $384M |
| 2025Q3 | $373M |
| 2025Q4 | $373M |
| 2026Q1 | $313M |
| 2026Q2 | $337M |
| 2026Q3E | $338M |
| 2026Q4E | $339M |
| 2027Q1E | $340M |
| 2027Q2E | $341M |
| 2027Q3E | $342M |
| 2027Q4E | $343M |
| 2028Q1E | $344M |
| 2028Q2E | $345M |
| 2028Q3E | $346M |
| 2028Q4E | $347M |
| 2029Q1E | $348M |
| 2029Q2E | $349M |
| 2029Q3E | $350M |
| 2029Q4E | $351M |
| 2030Q1E | $352M |
| 2030Q2E | $353M |
| 2030Q3E | $354M |
| 2030Q4E | $355M |
| 2031Q1E | $356M |
| 2031Q2E | $357M |
Assumptions & reasoning
- This is a residual: Baidu General Business less the four lines above it and less the Apollo Go estimate. It therefore absorbs the rounding in Baidu's one-decimal-place management table, roughly $10-15M a quarter.
- Down 15% year over year, up 2% sequentially. Modelled at 1.5% a quarter - flat to slightly up, on the view that intelligent-driving licensing grows while devices shrink.
- Low margin. Hardware and licensing at this scale do not carry the software economics of the lines above.
iQIYI
Growth pathThe consolidated streaming business, RMB 6.3B a quarter and falling 5% year over year, with an operating loss that widened to RMB 105M. Baidu does not own all of it and does not appear to want more of it, but it consolidates fully, so it sits in every headline number the market reacts to.
Latest: $710M (2031Q2E)
| Period | Value |
|---|---|
| 2025Q1 | $1.03B |
| 2025Q2 | $952M |
| 2025Q3 | $960M |
| 2025Q4 | $977M |
| 2026Q1 | $895M |
| 2026Q2 | $905M |
| 2026Q3E | $890M |
| 2026Q4E | $876M |
| 2027Q1E | $862M |
| 2027Q2E | $849M |
| 2027Q3E | $837M |
| 2027Q4E | $826M |
| 2028Q1E | $815M |
| 2028Q2E | $805M |
| 2028Q3E | $795M |
| 2028Q4E | $786M |
| 2029Q1E | $777M |
| 2029Q2E | $768M |
| 2029Q3E | $760M |
| 2029Q4E | $752M |
| 2030Q1E | $745M |
| 2030Q2E | $737M |
| 2030Q3E | $730M |
| 2030Q4E | $723M |
| 2031Q1E | $717M |
| 2031Q2E | $710M |
Assumptions & reasoning
- Shown NET of the inter-segment elimination (RMB 145M in Q2 2026) so the seven verticals sum exactly to consolidated revenue. That is why these figures are marked estimated - the elimination is a group-level adjustment, not an iQIYI disclosure.
- Revenue -5% year over year, +1% sequentially; segment operating loss widened from RMB 46M to RMB 105M and adjusted EBITDA was RMB 11M, essentially break-even.
- Modelled at -0.5% a quarter easing to +0.5%: a business that stops shrinking without ever really growing. Nothing in the disclosures argues for better.
- The honest caveat this model does not price: Baidu consolidates 100% of iQIYI's revenue and cash flow but does not own 100% of it. The net cash figure below carries a haircut for that; the revenue here does not.
From cash flow to fair value
| Present value of free cash flow, 20 quarters | -$6.32B |
| Terminal-year revenue | $22.90B |
| Terminal-year EBITDA | $4.52B |
| Exit multiple, on revenue | 0.8x |
| Terminal value | $18.32B |
| Discounted at 15.0% a year, terminal value becomes | $9.11B |
| Enterprise value | $2.79B |
| Net cash | $24.00B |
| Equity value | $26.79B |
| Diluted shares | 0.34B |
| Fair value per share | $77.78 |
| Against the current price of $91.97 | -15% |
1.6x terminal revenue, on a terminal year with roughly a 27% EBITDA margin - about 6x terminal EBITDA. That is a mature-large-cap-in-China multiple, and it is deliberately unexciting. The reference point that matters: Baidu's enterprise value TODAY is roughly 0.3x trailing revenue, because the market is netting out a cash pile nearly as large as the market capitalisation. So 1.6x is already a large re-rating - it assumes the market eventually pays for the operating business at all. The discount rate is 12%, above Tesla's 10% and Micron's 11%, for the obvious reasons: a Chinese ADS, a variable-interest structure, and a capital programme currently funded with short-term borrowing. Move the slider: this input and the AI Cloud capex glide together decide most of the answer. At 1x the fair value falls to roughly the current share price; at 3x it roughly doubles it.
Read the other way round: at $91.97 the market is paying 1.2x terminal-year revenue, holding every other assumption on this page fixed. That is the number to argue about.
The projected path
| Quarter | AI Cloud Infrastructure | AI Applications | AI-native Marketing Services | Legacy Business | Apollo Go (Robotaxi) | Other Baidu General Business | iQIYI | Revenue | YoY | EBITDA | Capex | FCF | R40 | PV of FCF |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2026 Q3E | $1.16B | $375M | $386M | $1.47B | $16M | $338M | $890M | $4.63B | +1% | $680M | $1.57B | -$890M | -18 | -$860M |
| 2026 Q4E | $1.24B | $381M | $389M | $1.41B | $18M | $339M | $876M | $4.66B | -3% | $698M | $1.52B | -$824M | -21 | -$769M |
| 2027 Q1E | $1.33B | $388M | $392M | $1.36B | $21M | $340M | $862M | $4.69B | -1% | $718M | $1.48B | -$759M | -17 | -$683M |
| 2027 Q2E | $1.41B | $394M | $395M | $1.31B | $24M | $341M | $849M | $4.73B | +2% | $740M | $1.43B | -$694M | -12 | -$603M |
| 2027 Q3E | $1.49B | $400M | $398M | $1.27B | $27M | $342M | $837M | $4.77B | +3% | $764M | $1.39B | -$631M | -10 | -$530M |
| 2027 Q4E | $1.58B | $407M | $401M | $1.23B | $31M | $343M | $826M | $4.82B | +3% | $788M | $1.36B | -$571M | -8 | -$463M |
| 2028 Q1E | $1.66B | $413M | $404M | $1.19B | $35M | $344M | $815M | $4.87B | +4% | $813M | $1.33B | -$513M | -7 | -$402M |
| 2028 Q2E | $1.75B | $419M | $408M | $1.16B | $40M | $345M | $805M | $4.92B | +4% | $839M | $1.30B | -$459M | -5 | -$347M |
| 2028 Q3E | $1.83B | $425M | $411M | $1.13B | $46M | $346M | $795M | $4.98B | +5% | $865M | $1.27B | -$409M | -4 | -$299M |
| 2028 Q4E | $1.92B | $431M | $414M | $1.10B | $52M | $347M | $786M | $5.05B | +5% | $892M | $1.25B | -$362M | -2 | -$255M |
| 2029 Q1E | $2.01B | $438M | $417M | $1.07B | $59M | $348M | $777M | $5.11B | +5% | $919M | $1.24B | -$319M | -1 | -$217M |
| 2029 Q2E | $2.09B | $444M | $420M | $1.04B | $67M | $349M | $768M | $5.18B | +5% | $946M | $1.23B | -$280M | +0 | -$184M |
| 2029 Q3E | $2.18B | $450M | $424M | $1.02B | $76M | $350M | $760M | $5.26B | +5% | $973M | $1.22B | -$244M | +1 | -$155M |
| 2029 Q4E | $2.27B | $456M | $427M | $993M | $86M | $351M | $752M | $5.33B | +6% | $1.00B | $1.21B | -$211M | +2 | -$130M |
| 2030 Q1E | $2.35B | $463M | $430M | $970M | $98M | $352M | $745M | $5.41B | +6% | $1.03B | $1.21B | -$183M | +2 | -$108M |
| 2030 Q2E | $2.44B | $469M | $434M | $949M | $111M | $353M | $737M | $5.50B | +6% | $1.06B | $1.21B | -$157M | +3 | -$90M |
| 2030 Q3E | $2.53B | $476M | $437M | $928M | $126M | $354M | $730M | $5.58B | +6% | $1.09B | $1.22B | -$134M | +4 | -$74M |
| 2030 Q4E | $2.62B | $482M | $440M | $909M | $142M | $355M | $723M | $5.68B | +6% | $1.11B | $1.23B | -$115M | +4 | -$61M |
| 2031 Q1E | $2.72B | $488M | $444M | $890M | $161M | $356M | $717M | $5.77B | +7% | $1.14B | $1.24B | -$98M | +5 | -$51M |
| 2031 Q2E | $2.81B | $495M | $447M | $872M | $183M | $357M | $710M | $5.87B | +7% | $1.17B | $1.26B | -$84M | +5 | -$42M |
Every row is projected. A year-over-year change is shown only where the quarter it compares against exists — an em dash means there is no comparable quarter, not a flat year.
Model revisions
Assumptions are marked to reality as each quarter prints. Every change is appended here, with the fair value the model produced at the time, so the model's own history stays visible.
| Date | Changed | Fair value then | Note |
|---|---|---|---|
| 2026-08-18 | all | $136.65 | Initial model, built off the Q2 2026 print: RMB 31.33B ($4.62B) consolidated revenue, Core AI-powered business at exactly half of Baidu General Business, RMB 11.39B of capex and free cash flow of -$1.17B. Seven verticals on the Core AI-powered cut rather than the reported revenue split, because the two cuts overlap. History rebuilt to six quarters from the Q2 2026, Q1 2026 and Q4 2025 releases so the first projected quarter carries a real year-over-year comparison. |