Tencent and Baidu, two leading Chinese tech giants, have developed plans to adapt to new conditions driven by US export restrictions on semiconductors, aimed at maintaining their competitiveness in artificial intelligence.
Tencent's Strategy for GPU Stockpiling
Tencent's president, Martin Lau, reported the company has built up a "pretty strong stockpile" of graphics processing units (GPUs) essential for training large AI models. By purchasing in advance, the company has secured enough chips to fuel its research for upcoming generations of models. Lau emphasized that rather than simply adding more GPUs, the focus has been on optimizing the performance of existing hardware. "We just need to sort of keep exploring these venues and probably spend more time on the software side rather than just brute-force buying GPUs," he said. Tencent is also exploring custom-designed chips produced within China.
Baidu's Full-Stack AI Ecosystem
Baidu, which runs the largest search engine in the country, emphasizes its 'full-stack' setup. This means it controls everything from the cloud servers to AI models. Dou Shen, president of Baidu’s AI cloud division, pointed out that even without access to the most advanced chips, their unique capabilities allow them to build strong applications and deliver meaningful value. Baidu is also implementing software tricks to reduce the cost of running AI tasks, making this a key competitive advantage as the demand for computing power grows.
Overall Progress in China's Chip Industry
China is pushing hard to build its own chip industry and while domestic GPUs and AI chips still lag behind Western offerings, analyst Gaurav Gupta of Gartner noted that Chinese firms have made steady progress in materials, equipment, chip design, and packaging. Executives in Washington and Silicon Valley have urged a reevaluation of export limits, claiming they do more harm to American firms than to Chinese buyers.
Tencent and Baidu's strategies illustrate how Chinese tech companies are adapting to changing market conditions and restrictions while continuing to advance their AI technologies.
