Baidu has announced integrating DeepSeekโs artificial intelligence (AI) models into its search engine. This follows closely behind Tencent, which recently incorporated DeepSeek technology into Weixin, Chinaโs largest social media platform.
In a statement released on Sunday, Baidu confirmed that it will fully connect DeepSeekโs AI models with its own Ernie large language models (LLMs) to provide users with a more diverse and advanced search experience. The company also revealed that DeepSeek will be added to its LLM platform for developers, expanding opportunities for AI-driven innovation.
This decision comes just days after Tencent beta-tested DeepSeekโs AI models within its Weixin app, the mainland China version of WeChat. Over the weekend, some users accessed the DeepSeek-R1 model through Weixinโs search bar, potentially introducing DeepSeekโs AI capabilities to the platformโs 1.3 billion active users. However, the AI model is currently only available for mainland China accounts.
Baiduโs stock takes a hit amid AI competition
Baiduโs share price dropped by around 8% in afternoon trading on the Hong Kong Stock Exchange, marking its sharpest decline since last November. The slump comes just ahead of the companyโs fourth-quarter earnings report, which is set to be released on Tuesday.
Meanwhile, DeepSeek has emerged as Chinaโs most talked-about AI start-up after making significant advancements in the tech industry. Last month, the company unveiled two powerful AI models, DeepSeek-V3 and DeepSeek-R1, which were developed at a fraction of the cost of other leading AI models. This breakthrough sent shockwaves through Wall Street, causing AI-related stocks to tumble.
DeepSeekโs popularity is largely due to its open-source approach, which allows anyone to use and modify its AI models. This strategy has prompted competitors, including Baidu, to rethink AI development methods. Historically, Baidu has focused on closed-source AI but has shifted towards a more open approach.
Last week, Baidu announced that the next generation of its Ernie LLM, scheduled for release on June 30, will be open source. Additionally, the company confirmed that its Ernie Bot service will be free from April 1, making AI more accessible to the public.
Local governments adopt DeepSeek AI
Chinese businesses and government agencies are rapidly deploying DeepSeekโs AI technology. Guangzhou and Shenzhen, located in southern Guangdong province, announced on Sunday that they have integrated the DeepSeek-R1 model into their civil service platforms.
Guangzhouโs municipal government has adopted DeepSeek-R1 and other LLMs to support domestic hardware. These AI models will improve public services, including policy interpretation and task management for government hotlines.
In Shenzhen, the Longgang district government implemented the DeepSeek-R1 model on February 8, making it the first district-level administration in China to adopt the start-upโs cost-effective, high-performance AI technology widely.
As competition in Chinaโs AI sector intensifies, Baidu and Tencentโs adoption of DeepSeek marks a significant shift in the industry. Open-source AI models are gaining traction among major tech companies and public institutions.
