Baidu, a prominent Chinese tech company, is integrating artificial intelligence (AI) features and voice functionality into its search platform to maintain its competitive edge against rivals such as DeepSeek. The company aims to transform its mobile search platform into an AI chatbot, assisting users with creative and organizational tasks. This shift prioritizes natural language searches over traditional keyword searches and supports voice search in multiple Chinese dialects.
Dan Ives, global head of tech research at Wedbush Securities, highlighted that while there has been some pressure on Baidu’s search business, the focus on AI and Ernie Bot is a strategic move forward. Baidu’s search lead, Zhao Shiqi, emphasized that the goal is not to outpace competitors but to surpass their former capabilities.
Baidu’s search platform now includes new features such as a voice search function, allowing users to input over a thousand characters, far exceeding the previous 28-character limit. Users can ask questions in a more natural, conversational style, similar to interacting with chatbots, and even prompt the search engine with images and files. The platform can generate photos and videos based on user requests, enhancing its functionality.
On Wednesday, Baidu launched its image-to-video model, aiming to enable marketers to create compelling videos on Baidu’s social feeds. The company is also leveraging AI to boost its cloud division, hoping to compete with major players like Alibaba Holdings and Tencent. Despite a 2.5% increase in shares in 2025, Baidu lags behind Alibaba’s 30.5% spike and Tencent’s 20% surge.
Baidu’s first-quarter 2025 results exceeded expectations, with adjusted earnings and total revenues surpassing estimates. The Baidu Core segment, particularly its AI Cloud business, saw a 42% surge compared to the previous year, reflecting increasing market recognition of Baidu’s full-stack AI solutions and competitive price-performance edge. However, increased costs and elevated selling, general, and administrative expenses impacted the company’s bottom line during the quarter.
Under the AI Cloud business, Baidu enhanced its MaaS platform, Qianfan, with an expanded model library and more comprehensive toolkits. This extension supports training and fine-tuning multimodal and reasoning models to facilitate AI-native application development. In April 2025, Baidu released the 3.0 version of its deep learning framework, PaddlePaddle.
Ives argued that Baidu needs to be aggressive and perception-wise to show it is not the little brother to Tencent on the AI front, encouraging more AI infusion. The company is confident that its Ernie foundation models will stand up to challengers like DeepSeek, fostering a broader ecosystem of AI-native applications. Baidu introduced the bot in 2023 and has been focused on updating the models ever since. This April, the tech giant released the Ernie 4.5 Turbo and Ernie X1 Turbo, advertised as faster and cheaper than previous versions. The X1 Turbo, in particular, is designed to compete with open-sourced models such as Alibaba’s Qwen and DeepSeek.
Source: https://www.ainvest.com/
