The revenue growth rate of Baidu's AI business exceeded 50%, indicating that Baidu's internalized AI has reached the moment of realization.

On November 18th, Baidu released a "special" third - quarter earnings report.

Data shows that in the third quarter of 2025, Baidu achieved an operating income of 31.2 billion yuan, among which the revenue of Baidu Core was 24.7 billion yuan.

However, the performance figures themselves are not the whole story. The real "special" thing about this earnings report lies in the comprehensive upgrade of data disclosure dimensions.

Since Baidu announced the AI - driven reshaping of its applications in 2023, the market has been waiting for a quantifiable AI report card. In the third - quarter earnings report of 2025, Baidu for the first time put the detailed data of its AI business on the table.

According to the earnings report, in the third quarter of 2025, Baidu's AI business revenue increased by more than 50% year - on - year. Among them, the revenue of AI Cloud increased by 33% year - on - year; the revenue of AI applications reached 2.6 billion yuan; the revenue of AI - native online marketing services increased by 262% year - on - year, reaching 2.8 billion yuan.

This move sent two key signals: First, Baidu's strategy of reshaping its business with AI is on the right track and has now entered a breakthrough stage of reaping results; Second, the continuous high - intensity strategic investment ensures the sustainability of Baidu's future AI development, and this itself is a confident display of strength.

01

Qualitative Change in Revenue Structure

If a core indicator is used to measure the commercialization process of Baidu's AI, the proportion of AI revenue in Baidu Core's revenue is undoubtedly the best yardstick. In the third quarter of 2025, this key figure was close to 40%. This marks a qualitative change in Baidu's business structure, enabling it to join the ranks of a very small number of domestic Internet giants whose revenue growth is truly driven by AI.

Behind this figure is not simply the superposition of technologies, but Baidu's strategic decision to "internalize" AI from an external tool into the core of its products and business. While the market is still discussing how to "use" AI, Baidu is committed to making the company "become" AI. Its approach is to comprehensively reconstruct all - line businesses with AI - native thinking, rather than simply grafting functions.

The implementation of this strategy first started with the C - end products with the largest user base. As the first core product to be reconstructed, Baidu Search underwent its biggest overhaul in a decade and is also the most aggressively AI - transformed search engine globally. Its intention is to transform the search from an Internet application mainly based on text content and links into an AI application mainly based on rich media content such as pictures and videos.

The earnings report revealed that as of October, about 70% of mobile search result pages contained AI - generated content. The same AI reconstruction logic was also applied to Baidu Wenku and Baidu Netdisk. In April 2025, they jointly launched the GenFlow general AI agent tool platform, which has now been iterated to the third generation, with over 20 million active users, becoming the "world's largest general agent".

On the basis of AI reshaping user products, there is a full - scale blossoming of commercialization capabilities.

In the most direct advertising field, AI is no longer just an optimization tool, but the revenue - generating model itself. The "Huibo Star" digital human designed specifically for live - streaming e - commerce drove a year - on - year surge of 91% in GMV during this year's "Double 11"; the "Merchant Agent" created for merchants can, as shown in some cases, increase the order conversion rate in the tourism industry by 30% and reduce the lead cost in the education industry by 30%. The digital human and agent businesses are expected to become the second growth curve of the company's online marketing business.

This "internalization - commercialization" logic also runs through the B - end business.

The recently released "Baidu Famo" agent can independently evolve and find the optimal solution in complex R & D scenarios. Within a few hours of its launch, more than 1,000 enterprises applied for testing. In addition, at the beginning of August, Baidu Smart Cloud launched the world's first batch of out - of - the - box enterprise - level agents - AI digital employees, which have a high - degree of anthropomorphic interaction ability and have covered seven major scenarios such as marketing and recruitment, serving thousands of enterprises in finance, automotive, and cultural tourism industries.

Luobo Kuaipao is the "other pole" of Baidu's AI layout. In the third quarter of 2025, the number of global travel services provided by Luobo Kuaipao reached 3.1 million, a year - on - year increase of 212%. In October, the number of fully driverless orders per week by Luobo Kuaipao exceeded 250,000. As of November, its cumulative global travel service times exceeded 17 million, ranking first in the world.

Currently, Luobo Kuaipao has launched its globalization process. On October 22nd, Luobo Kuaipao officially announced a strategic cooperation with PostBus under Swiss Post to launch the autonomous driving travel service "AmiGo" in Switzerland. On November 10th, Luobo Kuaipao announced that it had obtained the first batch of fully driverless commercial operation licenses issued by the Integrated Transport Centre (ITC) of Abu Dhabi, marking the first large - scale fully driverless operation for the public outside China by Luobo Kuaipao.

Not long ago, Baidu CEO Robin Li proposed "internalizing AI capabilities" at the Baidu World Conference. The high - speed growth of the AI business in this earnings report also indirectly verified Baidu's progress in internalizing AI capabilities. Behind the high - speed growth of the AI business is the inevitable "external manifestation of effects" after Baidu successfully "internalized" AI from an external tool into the native ability within the organization, which is gradually pushing the development process of Baidu's AI business from the investment stage to the return stage.

02

Full - Stack AI Capability is the Core Foundation

So, what is the underlying logic behind the rapid growth of Baidu's new AI business? What supports the continuous revenue generation of the AI business is the full - stack technology foundation that Baidu has built over the years.

This vertically integrated system from chips to applications is the physical foundation and prerequisite for realizing "internalizing AI capabilities". Only when full - stack autonomy and controllability are achieved can an enterprise truly integrate the pulse of AI with its own business development, rather than being controlled by others. The efficient feedback loop created by this in - depth integration makes AI no longer an external cost, but truly becomes the endogenous productivity of the enterprise.

Currently, Baidu has a full - stack AI ecosystem from chips - deep learning frameworks - basic models/Qianfan Maas platform - upper - layer applications, and is one of the few Internet companies globally that can provide a complete AI technology stack.

The core value of this vertical integration lies in creating an efficient feedback loop: the optimization of underlying chips and computing power infrastructure directly serves the training and inference of upper - layer large models, and the improvement of large - model capabilities in turn feeds back user needs through upper - layer applications and helps accumulate more data, thus achieving a positive cycle and ensuring the continuous and rapid iteration of AI capabilities.

More importantly, the full - scale self - research from chips to models means that Baidu firmly controls the cost and iteration rhythm of AI in its own hands, which builds a more resilient and sustainable profit model for it.

At the infrastructure layer, Baidu takes its self - developed Kunlun Chip and Smart Cloud as the core to build the foundation for its AI strategy, ensuring the autonomy and controllability of computing power infrastructure. At the same time, self - developed chips are also an important way to get rid of the problem of inverted AI value chain.

On November 13th, Baidu released the new - generation Kunlun Chip and Tianchi Super Node and announced that it will launch new products every year in the next five years. The new Kunlun Chip includes two models: the M100, which focuses on large - scale inference and has high cost - effectiveness, will be launched in 2026; the M300, which is for ultra - large - scale multi - modal model training and inference, will be launched in 2027. The Tianchi 256 and Tianchi 512 super nodes will be launched next year. Currently, tens of thousands of Kunlun Chips have been deployed, and together with the Baige platform, they provide high - performance and scalable computing power for enterprises, becoming a highly competitive domestic chip alternative. This has also accelerated the commercialization process of the Kunlun Chip. In August this year, the Kunlun Chip won a billion - level procurement contract from China Mobile, achieving a major breakthrough for domestic AI chips in large - scale projects.

In the model layer, as the "brain" of the technology system, the Wenxin Big Model is the core engine of Baidu's AI capabilities. The newly released Wenxin Big Model 5.0 at the Baidu World Conference is a native full - modal model that can seamlessly process text, images, audio, and video, and has achieved generational leaps in understanding, logic, and creativity.

According to the latest ranking of the LMArena Big Model Arena on November 8th, the Wenxin model ERNIE - 5.0 - Preview - 1022 ranked second globally and first in China in the text task evaluation, especially excelling in creative writing and complex problem understanding.

In terms of verifying commercialization results, benefiting from the leading AI capabilities and rich product system under full - stack self - research, as well as the steady growth of domestic AI training and inference demand, Baidu has gained more market recognition in the AI field. In this quarter, Baidu's AI Cloud revenue increased by 33% year - on - year, and the subscription revenue of AI high - performance computing infrastructure increased by as much as 128% year - on - year.

03

Valuation Reshaping Continues

Baidu's latest earnings report is not only a phased realization of AI results. Its updated business disclosure caliber is more like a declaration, demonstrating the management's full confidence in future growth.

This confidence stems from its rare long - term strategic determination in the AI field. At the latest earnings conference, Baidu clearly stated that it will continue to maintain high - intensity long - term investment. This continuous and firm investment has continuously strengthened its full - stack layout from chips to applications and formed a strong closed - loop for AI in its business ecosystem, continuously deepening its moat.

Looking ahead, as the "internalization of AI capabilities" continues to deepen, Baidu's growth model is showing a stronger compound interest effect. When AI is no longer just a department or a technology, but becomes the "muscle memory" of the company's thinking and operation, new AI - native businesses and growth points will continue to emerge.

This is exactly the core logic for the market to start re - evaluating Baidu's value: It is no longer a company that "owns" AI technology, but a company driven by AI as the core. The growth resilience and imagination space brought about by this essential change are becoming the continuous engine for its value re - evaluation.

The capital market has smelled this change. Since this year, especially after September, more than 20 top domestic and overseas institutions, including Goldman Sachs and Morgan Stanley, have successively raised their investment ratings and target prices, which is the best endorsement of this core logic change.

*Disclaimer:

The content of this article only represents the author's views.

The market is risky, and investment should be made with caution. In any case, the information in this article or the opinions expressed do not constitute investment advice for anyone. Before making an investment decision, if necessary, investors must consult professionals and make decisions carefully. We have no intention to provide underwriting services or any services that require specific qualifications or licenses for the trading parties.