Baidu's Q1 revenue tops estimates as AI cloud services offset ad weakness

Baidu's Q1 revenue tops estimates as AI cloud services offset ad weakness

FILE PHOTO: Baidu sign is seen at the World Artificial Intelligence Conference (WAIC) in Shanghai, China July 6, 2023. REUTERS/Aly Song/File Photo

Chinese search engine giant Baidu on Wednesday reported better-than-expected first-quarter revenue, fueled by increasing demand for its AI cloud services.

The developer of the Ernie AI chatbot offset losses in its advertising market by growing its AI cloud segment, which provides cloud computing, AI-powered solutions, and services to enterprises.

US-listed shares of Baidu rose more than 1.4% in premarket trading.

Total revenue in the first quarter rose 3% to 32.45 billion yuan ($4.50 billion), beating analysts’ average estimate of 30.9 billion yuan, according to data compiled by LSEG.

Baidu‘s online marketing business, which contributes the majority of its revenue, fell 6% to 17.31 billion yuan. Analysts had estimated 17.39 billion yuan.

Its non-online marketing revenue reached 9.4 billion yuan, up 40% year over year, mainly driven by its AI cloud business.

The company reported profit of 21.59 yuan per American Depositary Share, compared with profit of 14.91 yuan per share a year earlier.

Baidu has intensified its focus on AI in recent years to reduce its dependence on advertising revenue from its core search engine business.

In early 2023, the company was among the first to launch a chatbot following Microsoft-backed OpenAI’s release of ChatGPT in late 2022.

Despite this early advantage, Baidu‘s Ernie large-language model faces fierce competition from Chinese firms such as startup DeepSeek, which shook up the AI landscape earlier this year.

In response, Baidu eliminated fees for its premium chatbot services in April and introduced two new AI models in March – the reasoning-focused Ernie X1 and Ernie 4.5. The company upgraded these offerings to “Turbo” versions the following month.

Baidu‘s CEO Robin Li said last month that its cluster comprising 30,000 of its self-developed, third-generation P800 Kunlun chips can support the training of DeepSeek-like models.

Reuters

Bring stories like this into your inbox every day.

Sign up for our newsletter - The Daily Brief
Subscribe to Newsletter


This is your last free story for the month. Register to continue reading our content