An overview of the Baidu logo seen at the Shanghai New Expo Center during the World Artificial Intelligence Conference 2025 held in Shanghai, China on July 28, 2025.
Ying Tan | Null Photo | Getty Images
tech giant Baidu has emerged as one of China’s major players in artificial intelligence chips, positioning itself as a challenger to Huawei, with both looking to fill the void left by the industry leader. Nvidia are excluded from the country.
Baidu, best known as China’s largest search business, has in recent years refocused its business around self-driving cars and AI, including its subsidiary Kunlunxin, which designs chips.
Several analysts have raised their outlook for Baidu stock in the past few weeks, citing its semiconductor business and expecting further growth in domestic orders for the sector.
This month, Baidu announced a five-year roadmap for its Kunlun AI chips, starting with M100 in 2026 and M300 in 2027. The company is already using a combination of homegrown chips in its data centers to run ERNIE AI models and Nvidia products.
Baidu makes money by selling chips to third parties who build data centers and by renting computing power via the cloud. The company is trying to position itself as a so-called “full stack” AI product, with an infrastructure consisting of chips, servers, data centers, and AI models and applications.
And the chip business appears to be gaining momentum. Earlier this year, Kunlun New won an order from a supplier to China Mobile, one of China’s largest mobile carriers.
“Kunlunxin has emerged as the country’s leading AI chip developer with a focus on high-performance AI chips for large-scale language model (LLM) training and inference, cloud computing, communications and enterprise workloads,” Deutsche Bank analysts said in a note this month.
Nvidia’s graphics processing units (GPUs) are widely regarded as cutting-edge chips for training and running AI, but the company has been blocked by the U.S. government from selling its top-of-the-line products to China. The Chinese government is also reportedly trying to persuade local tech companies not to buy H20, an inferior Nvidia chip designed for the Chinese market and developed for export.
As Huawei, which has played a leading role in large-scale chips, fades into obscurity, analysts suggest Baidu will fill the gap and its chip business will see explosive growth.
“We believe domestic demand for AI computing in China remains strong and hyperscalers are increasingly sourcing from local solution providers,” JPMorgan said in a note on Sunday. “We believe the Kunlun AI chip is one of the most advantageously positioned chips.”
Analysts at the investment bank expect Baidu Chip’s sales to increase sixfold, reaching 8 billion Chinese yuan ($1.1 billion) by 2026.
Macquarie analysts estimate that Baidu’s Kunlun chip division could be worth about $28 billion.
Baidu is not alone among Chinese tech giants when it comes to self-developed semiconductors. CNBC reported in August. alibaba We are also progressing with the development of next-generation AI chips.
AI chip shortage hits China directly
Baidu’s chip push comes after the Chinese tech giant said this month it was experiencing supply shortages.
“The supply side is going to be a relatively big bottleneck” for the next two to three years, Alibab CEO Eddie Wu said, referring to the components and chips needed to build data centers.
tencent said this month that capital spending in 2025 will be lower than originally expected. However, Tencent President Martin Lau said this was not due to a lack of demand, but rather a lack of chips to fund.
“This does not reflect a change in our AI strategy… it is certainly a change in terms of the availability of AI chips,” Lau said.
This shortage is driven in part by global demand and resulting bottlenecks in the semiconductor supply chain. But China’s de facto blockade of NVIDIA chips also reduced supply.
Chinese tech companies are making their AI models more efficient to make more use of the semiconductors they have, and are trying to alleviate shortages by tapping into stockpiled chips.
Meanwhile, China has its own manufacturing challenges as it is the largest chip maker. SMICcannot compete with leaders such as Taiwan Semiconductor Manufacturing Co. in terms of scale and technology. That makes it difficult for China to manufacture enough domestic chips to fill the gap.
Like their U.S. peers, Chinese tech companies continue to report strong demand for AI.
“We see customer demand for AI as being and continuing to be very strong. In fact, we haven’t even been able to keep up with the increase in customer demand in terms of the pace at which we can deploy new servers,” Alibaba’s Wu said this week.
This gives Baidu an opportunity in China.
“Baidu’s chip push is both a necessity and an opportunity. It’s necessary because Chinese platforms can no longer steadily eat up U.S. GPUs, but it’s an opportunity because the multi-billion dollar domestic market for AI hardware that complies with both U.S. export rules and Beijing’s self-reliance policies is semi-captive,” Nick Patience, AI practice lead at Futurum Group, told CNBC.
“If Baidu can ship a competitive Kunlun generation on time, it will not only solve its own supply problems, but also become a strategic supplier to other AI industries in China.”
