Develop new AI infrastructure based on search engines and cloud services.

Since 2000, Baidu has emerged in the PC Internet era, expanding from search engines to content communities; in 2010-2015, Baidu missed opportunities for mobile Internet development; however, in the current AI era, Baidu's competitive advantage began to stand out, mainly due to the forward-looking layout of new AI infrastructure around 2016. Currently, Baidu's core business is driven by AI, contributing more than 70% of the company's revenue. Specifically, it is divided into three core business lines. One is a mature mobile ecosystem business, the second is a rapidly developing intelligent cloud business, and the third is the future smart driving business with hard technology.

AI drives Baidu's entire business transformation, and large model capabilities continue to iterate.

1) Mobile ecosystem: Online marketing business is the core, and AI restructuring unlocks commercial potential. Baidu started with a search engine and firmly grasped the search traffic portal. Baidu search led the domestic market. In 2023, Baidu released the 4.0 version of Wenxin's big model, and launched more than ten AI native applications such as Baidu search, such as streaming, maps, network disk, and library, based on Wenxin's restructured words. It also uses search as the core to create an X+Y strategic layout, forming a virtuous cycle from traffic platforms to content ecosystems. Implement AI native applications from 0 to 1, and use AI big models to empower the original business format.

2) Intelligent cloud: the second engine of continuous growth. With the release of Wenxin's big model and Wenxin's words, Baidu's leading position in the AI field has been further consolidated and expanded. Generative artificial intelligence and big language models have enhanced Baidu's competitive advantage in the cloud field. Baidu's smart cloud business grew steadily, benefiting from increased demand for large-scale model training and inference.

In 2019-2023, Baidu's smart cloud business revenue as a share of core revenue increased from 8.0% in 2019 to 18.1% in 2023.

3) Intelligent driving: Achieve 1 to N scenarios. Baidu was the first company to enter intelligent driving in China. It was the first to implement the intelligent driving business, and it has a first-mover advantage. At this stage, the company has clarified the three business models of Apollo's intelligent driving business. One is to provide automobile manufacturers with Apollo autonomous driving technology solutions; the second is Baidu builds cars and integrates Baidu's innovations in autonomous driving from end to end, such as cooperating with Geely; and third, sharing unmanned vehicles. The commercialization process of Baidu's unmanned vehicles has been accelerating, and Robotaxi has begun normalized commercial operations.

Investment advice:

2024 is Baidu's second year of generative artificial intelligence. The company is shifting from focusing on the Internet to prioritizing artificial intelligence, and is actively promoting the restructuring of ToB and ToC businesses using the Wenxin Big Model. Currently, Baidu's search and marketing business is stable, and the long-term opportunities brought by the AI model can not only reshape traditional businesses to explore new space, but are also a key driving force for cloud business and intelligent driving business. We expect the company to achieve revenue of 143.38/152.36/160.78 billion yuan in 2024-2026, and non-GAAP net profit of 29.09/30.94/32.71 billion yuan. According to the Segment Valuation Act, we finally gave the company a valuation of RMB 285.1 billion (HK$310.8 billion) in 2024, corresponding to a target price of HK$111. Covered for the first time, giving a “buy-A” rating.

Risk warning: Policy environments such as AI and intelligent driving have changed, competition in the intelligent driving industry has intensified, the construction of a mobile ecosystem falls short of expectations, AI technology iteration falls short of expectations, the progress of commercialization of the intelligent driving business falls short of expectations, and profit forecasts fall short of expectations.