Positive Comments: The AI Strategy Begins to Yield Results, and Baidu’s New Growth Curve is Gradually Taking Shape
Since the establishment of the “All in AI” strategy in 2017, Baidu has spent nearly a decade going through the pains of transformation from a “search giant” to an “AI technology company.” The release of the financial report for the third quarter of 2025 marks that its AI strategy has entered the “results verification period.” The “progress” at this stage is not only reflected in the rapid growth of AI business but also in the multi – dimensional breakthroughs in technology, products, and commercialization, providing a reference sample for the strategic transformation of large Internet companies.
First of all, the large – scale growth of AI business verifies the effectiveness of strategic investment. In the third quarter, Baidu’s AI business revenue increased by more than 50% year – on – year, accounting for 32% of the total revenue. For the first time, it formed a situation of “equal competition” with the traditional online marketing business. Behind this data is the coordinated efforts of the three major sectors of AI business: the revenue of intelligent cloud infrastructure reached 4.2 billion yuan (a year – on – year increase of 33%), and the subscription revenue of AI high – performance computing doubled (+128%), indicating the strong demand for enterprise – level AI; the revenue of AI native marketing services soared by 262% and was defined as the “second growth curve outside the traditional business,” which reflects the in – depth reconstruction of the marketing scenario by AI; although the growth rate of AI applications such as Baidu Wenku and Baidu Netdisk is relatively gentle (+6%), the user scale has approached 300 million, becoming the largest personal application matrix, laying the groundwork for subsequent commercialization. This full – chain growth from the “infrastructure layer – application layer – scenario layer” shows that Baidu’s AI business has shifted from “single – point breakthrough” to “ecological expansion.”
Secondly, the AI transformation of core products has reshaped user value. The upgrade of Baidu Search’s “intelligent box” (supporting multi – modal input and AI collaborative generation), 70% of search result pages containing AI content, and the data from QuestMobile showing that the monthly active users of AI search reached 382 million (a quarterly increase of 18.63%) all indicate that AI technology has been deeply integrated into users’ information acquisition habits. More notably, 56% of the users who left DeepSeek switched to Baidu, reflecting that Baidu still maintains its irreplaceability as an “information gateway” in the AI era. This irreplaceability is not simply about traffic competition but an active choice after improving the user experience through AI capabilities.
Thirdly, continuous investment in cutting – edge technologies has consolidated long – term competitiveness. ERNIE Bot 5.0 (with 24 trillion parameters and unified multi – modal modeling), the planning of self – developed chips Kunlun Chip M100/M300 (to be launched in 2026 – 2027), and the large – scale implementation of Apollo Go (3.1 million global trips and a cumulative driverless mileage of 140 million kilometers) all reflect Baidu’s full – stack layout in the “model – chip – scenario.” This closed – loop ability of “technology – product – commercialization” is the core barrier that differentiates Baidu from other AI players. For example, the explosion of digital human live – streaming during the Double 11 shopping festival (83% of anchors used Huibo Xing, and GMV increased by 91%) and the full – process automation of the AI short – drama creation platform are essentially the results of the spill – over of underlying technical capabilities, proving that AI has transformed from a “cost center” to a “productivity tool.”
Negative Comments: Hidden Growth Concerns and Intense Competition Coexist, and Baidu’s Transformation Still Needs to Overcome Three Challenges
Although Baidu’s AI business shows strong growth momentum, its transformation is far from being in a “care – free” stage. From the financial report data and the industry competition situation, the challenges it faces are equally acute: the continuous decline of traditional business, the profit pressure of AI business, and the white – hot technological competition together form the “three hurdles” on the transformation path.
The first challenge is the “stalling” pressure of traditional business. The online marketing revenue has declined year – on – year for six consecutive quarters (-18% in the third quarter), directly dragging down the total revenue (-7.1%), exposing the vulnerability of Baidu’s “foundation.” The weakening of the search entry advantage is the core reason: the rise of large AI model platforms (such as DeepSeek) and short – video platforms has diverted users’ information needs; the monthly active users of Baidu App only increased by 1% (708 million), indicating that user growth has reached the ceiling. The contraction of traditional business not only affects short – term cash flow but may also weaken Baidu’s bargaining power among advertisers. When advertisers’ budgets shift to AI native marketing (such as digital human live – streaming), if Baidu cannot achieve a “smooth transition” between old and new businesses, it may face a passive situation of “bleeding in old business and new business not fully filling the gap.”
The second challenge is the “profit dilemma” of AI business. Although the proportion of AI business revenue has increased, the profit margin problem is prominent: the cost of sales in the third quarter was 18.3 billion yuan (+12%), mainly driven up by the cost of intelligent cloud and content; the R & D expense was 5.2 billion yuan (accounting for 16.7% of revenue), and the cumulative investment has exceeded 100 billion yuan since the release of ERNIE Bot in 2023; although the number of trips of Apollo Go has soared, the quarterly revenue is less than 100 million yuan (estimated by the average ticket price), and the commercialization efficiency still needs to be verified. This “high – input, low – return” model may be acceptable in the short term (since the AI business is still in the expansion stage), but in the long run, if Baidu cannot improve the gross profit margin of the intelligent cloud (current revenue of 4.2 billion vs. high cost) and optimize the average revenue per order of Apollo Go (such as reducing operating costs through autonomous driving technology), the “quality of growth” of the AI business will be questioned.
The third challenge is the “dissipation of first – mover advantage” in technological competition. Baidu once took the lead due to its early “All in AI” layout, but the “arms race” in the large – model track has become white – hot: ERNIE Bot 5.0 dropped from the “first in China” to the 24th place in the LMArena ranking, being surpassed by Alibaba and DeepSeek; in the overseas market, Google’s Gemini 3 topped the list with 1501 points, and its general reasoning ability exceeded that of GPT – 5 Pro. This change warns that the iteration speed of AI technology far exceeds expectations, and the first – mover advantage may quickly turn into “catch – up pressure.” If Baidu cannot continuously widen the gap in model performance (such as multi – modal understanding and general reasoning), chip adaptation (the collaborative optimization of Kunlun Chip and ERNIE Model), and scenario implementation (such as users’ willingness to pay for AI short – dramas and digital humans), the uniqueness of its “AI label” will be diluted.
Suggestions for Entrepreneurs: Lessons on “New and Old Growth Momentum Transformation” from Baidu’s Transformation
Baidu’s “progress and setbacks” provide a vivid transformation sample for entrepreneurs. Under the dual pressures of technological change and market competition, the following three suggestions are worth considering:
Strategic Focus: Clearly Define the Priority of the “Second Growth Curve” while Maintaining the Resilience of the “Foundation”
Baidu’s success lies in its decade – long investment in AI, but the decline of its traditional business also reminds entrepreneurs that the connection between old and new businesses needs to be “advance while defending.” Entrepreneurs need to clearly define the strategic priority of the “second growth curve” (such as Baidu’s AI business), and at the same time, extend the life cycle of the “foundation” through technological empowerment (such as the AI transformation of Baidu Search) to avoid a break in the transformation caused by the “collapse of old business.” For example, when traditional enterprises transform, they can optimize the efficiency of their original business through digital tools and use the increment of new business to support the stock of old business, forming a virtuous cycle of “using new to drive old and old to support new.”Technological Investment: Balance “Long – term R & D” and “Short – term Commercialization” to Avoid the Trap of “Technology for Technology’s Sake”
Baidu’s investment of tens of billions in ERNIE Bot and Kunlun Chip is the foundation of its competitiveness, but the low profit margin of its AI business also suggests that technological R & D needs to be deeply integrated with commercialization scenarios. Entrepreneurs should avoid blindly pursuing “technological leadership” and design technological paths around the real needs of target users (such as the multi – modal input of Baidu AI Search and the live – streaming e – commerce of digital humans). For example, B2B enterprises can prioritize solving customers’ “cost pain points” (such as high – performance computing subscriptions for intelligent cloud), while B2C enterprises need to enhance users’ “experience increment” (such as the intelligent agent function of Baidu Netdisk), so that technological investment can be directly converted into revenue or user stickiness.Competition Response: Build “Differentiated Barriers” through “Rapid Iteration” Instead of Relying on the “First – Mover Advantage”
The decline of Baidu’s large – model ranking and the low average revenue per order of Apollo Go are essentially the results of the rapid catch – up of its “first – mover advantage.” Entrepreneurs need to realize that the “window period” in the technology track is getting shorter and shorter, and they must build barriers through “rapid iteration + differentiated innovation.” For example, in the AI application layer, they can focus on vertical scenarios (such as Baidu’s AI short – drama creation platform) and form a unique user experience through “full – process automation + human – machine collaboration”; in the technology layer, they can strengthen the collaboration of “model – hardware – scenario” (such as the adaptation of Kunlun Chip and ERNIE Model) to increase the imitation cost of competitors.
In summary, Baidu’s “progress and setbacks” are a microcosm of the transformation of technology companies. It proves the value of long – term strategic investment and reveals the complexity of the transformation of new and old growth momentum. For entrepreneurs, the key is not “whether to transform” but “how to maintain strategic determination, technological patience, and business acumen during the transformation” – this may be the core inspiration of Baidu’s case.
- Startup Commentary”Why Do Leading Beverage Brands Flock to Discount Stores Despite Selling Only a Few Dollars Cheaper Than Convenience Stores?”
- Startup Commentary”Distributors to Face a Major Shake-up in the Next Five Years”
- Startup Commentary”Is There a “Mass Exodus” of Internet Celebrities in Hangzhou? It’s Not That Scary!”
- Startup Commentary”In the Anti – social Media Era, Young People Rebuilding Relationships at “Mystery Dinner Parties” | Early – stage Projects”
- Startup Commentary”Industry Research Report on Multi-scenario Karaoke Systems”
