Tencent Prioritises Domestic AI Chips Amid China's Self-Sufficiency Drive
- tech360.tv
- Sep 17
- 2 min read
Tencent Holdings, which operates the world’s largest video gaming business by revenue and China’s biggest social media platform WeChat, is now supporting domestically designed artificial intelligence (AI) chips. This move aligns with China’s efforts to reduce reliance on imported semiconductors, including those from Nvidia.
Tencent Cloud, the company’s cloud computing unit based in Shenzhen, announced its support for "mainstream domestic chips" within its AI computing infrastructure. No specific Chinese integrated circuit brands were identified.
Tencent has "fully adapted to mainstream domestic chips" and "participates in the open-source community," said Qiu Yuepeng, Tencent Cloud president. He spoke at the company’s annual Global Digital Ecosystem Summit on Tuesday.
This commitment reflects a broader push within China’s semiconductor and AI sectors to advance Beijing’s tech self-sufficiency agenda. It comes amid US export restrictions on China and increasing geopolitical tensions.
Weeks prior, AI start-up DeepSeek announced its updated V3.1 model was engineered for the next generation of "home-grown chips to be released soon."
In July, the Model-Chips Ecosystem Innovation Alliance was formed by a group of Chinese semiconductor and AI companies. Its goal is to promote the adoption of locally developed processors for AI projects.
Tencent Cloud is collaborating with several domestic chip companies to utilise "the most suitable hardware" for various scenarios. Dowson Tong Tao-sang, Tencent senior executive vice-president and CEO of the firm’s Cloud and Smart Industries Group, conveyed this to Chinese media at the summit.
Tencent Cloud is also pursuing "long-term strategic investment" to optimise hardware and software collaboration. This strategy involves creating cost-effective AI computing power through the use of diverse chips.
This follows an assessment by Martin Lau Chi-ping, Tencent president, during the company’s recent quarterly earnings call. He stated Tencent had sufficient processors for training and "many options for inference chips," despite US export policy uncertainties.
Lau also mentioned that Tencent is developing software improvements to allow existing chips to handle increased workloads.
Tencent Cloud’s latest initiative coincides with a Monday announcement from the State Administration for Market Regulation. The regulator stated Nvidia violated antitrust laws and Chinese government approval terms for its 2019 acquisition of Mellanox Technologies, an Israeli interconnect solutions provider.
The regulator, which began its antitrust investigation into the US$6.9 billion deal in December, indicated it would continue its inquiry. Further details were not provided.
At the summit, Tencent Cloud also unveiled an updated version of its Hunyuan 3D-generation system. Its Yuanbao AI assistant now ranks among China’s top three native AI applications by daily active users.
The company also plans foreign market expansion, with data centres slated for Saudi Arabia and Japan.
During the first half of this year, Tencent Cloud held a 7% market share, ranking as the No. 4 AI cloud services provider on the mainland, Omdia research firm reported. It trailed behind cloud services units from Alibaba Group Holding, ByteDance-owned Volcano Engine, and Huawei Technologies.
Tencent’s Hong Kong-listed shares closed flattish at HK$645 on Tuesday. This marked the highest stock level since early 2021.
Tencent Holdings is actively integrating domestically designed AI chips into its infrastructure.
This strategy supports China's goal of tech self-sufficiency amidst US export restrictions.
Tencent Cloud is collaborating with various Chinese chip companies and investing in hardware and software optimisation.
Source: SCMP
