Baidu says Chinese buyers want local AI chips due to ‘supply chain’ issues

Baidu, a major Chinese web company, recently informed investors about the promising outlook for its Kunlunxin chip business. Kunlunxin manufactures CUDA-compliant inferencing chips, which Baidu utilizes for its cloud services and has sold to various Chinese firms such as Huawei and ZTE. Executives from Baidu's AI Cloud Group expressed confidence in Kunlunxin's long-term growth and commercial potential, citing a few key reasons.

First, the demand for inferencing is steadily increasing, and Baidu anticipates this trend will persist. Second, the domestic market in China holds significant growth potential due to ongoing supply chain constraints for AI chips. This situation has led Chinese buyers to increasingly seek high-performance, reliable, and cost-efficient domestic AI chips, as highlighted by Baidu's executive vice president, Dou Shen.

The US government's recent decision to allow Nvidia to resume selling its products in China poses a challenge for local competitors, as Beijing maintains the power to veto any purchases by Chinese companies. As a result, Chinese buyers are turning to domestic chips due to supply chain issues. Nvidia CEO Jensen Huang has advocated for increased chip sales to China to solidify the US's dominance in AI, while Beijing has also encouraged the adoption of local technology to lessen reliance on US products.

Baidu's Q2 earnings revealed a robust performance in its AI business, with revenue from cloud infrastructure rental up 50 percent year-over-year to nearly $1.1 billion, and GPU cloud revenue surging 283 percent year-over-year. These figures indicate that Baidu believes owning its own stack of models, infrastructure, and chips will allow it to deliver AI services at competitive prices and give it a market advantage.