Baidu Targets Enterprise Cloud Market with a Raft of Industry-Specific AI Agents
By Liu Peilin and Denise Jia
Chinese tech giant Baidu is intensifying its push into the enterprise cloud market with a suite of industry-specific artificial intelligence (AI) agents designed to deeply integrate AI into business operations across sectors such as energy, healthcare, transportation and automotive..
The company recently unveiled its new lineup of AI agents tailored for vertical industries including transportation, automotive, medical and environmental services. The rollout underscores Baidu’s intent to use intelligent cloud services as a strategic entry point into China’s enterprise market amid growing demand for AI-driven infrastructure.
Shen Dou, Baidu’s intelligent cloud president, said at a launch event that 65% of China’s central state-owned enterprises are already using Baidu’s AI products. Major clients such as State Grid Corp., China Iron & Steel Research Institute Group and China Merchants Bank have begun deploying Baidu’s in-house Kunlun P800 AI chips at scale.
In February, Baidu lit up a 10,000-card cluster of the P800 chips, and by April had scaled it up to 30,000 cards. Sources familiar with the matter report that the cluster is located in Ningxia and is now actively powering the training of Baidu’s large language models.
In the energy sector, Baidu’s AI agent collaborates with State Grid to generate electricity supply plans. When a user submits a request through the State Grid app, the AI identifies intent, breaks down the task, engages expert models and tools, and automates the workflow to produce a tailored plan.
In transportation, the AI agent has been deployed on the Beijing-Xiong’an Expressway in Hebei province. It detects incidents in real time and rapidly formulates strategies for an emergency response . Baidu claims the system’s hybrid use of small and large models has raised early warning accuracy to over 95%.
In the automotive sector, Baidu Cloud has teamed up with Deepal, a brand under Changan Auto, to launch an intelligent in-car assistant. The AI agent entertains children with personalized storybooks, English lessons and interactive cartoon characters during rides.
The AI cloud boom has fueled strong financial growth. In the first quarter, Baidu’s smart cloud revenue jumped 42% year-on-year to 6.7 billion yuan ($930 million), now comprising 26% of its core revenue, up from 20% a year earlier. Operating profit margin for the unit surpassed 10%.
Shen said during the company’s first-quarter earnings call that enterprise cloud services now drive most of Baidu’s AI cloud growth, with subscription models accounting for the majority of revenue. He said the core battleground among Chinese cloud providers has shifted to building and optimizing massive GPU clusters for model training and inference.
On the foundational model front, Baidu released its new large-scale pretraining model Ernie 4.5 and inference model X1 in March. The company claims Ernie 4.5 outperforms OpenAI’s GPT-4.5 on several benchmarks and costs just 1% of GPT-4.5’s API fee. The X1 model reportedly rivals DeepSeek-R1 in performance at half the API price.
By late April, Baidu introduced upgraded versions — Ernie 4.5 Turbo and X1 Turbo — which offer faster inference speeds, stronger capabilities and significantly reduced costs. The company plans to open-source the Ernie 4.5 series on June 30.
Robin Li, Baidu’s CEO, previously stated that the company has no intention of leading in every possible AI direction. Instead, Baidu will focus on areas with real application value — such as multimodal AI, search transformation and digital humans — which it sees as key to the next phase of tech evolution.
Contact reporter Denise Jia (huijuanjia@caixin.com)
caixinglobal.com is the English-language online news portal of Chinese financial and business news media group Caixin. Global Neighbours is authorized to reprint this article.
Image: InfiniteFlow – stock.adobe.com