China’s advancement in artificial intelligence is being hindered by U.S. trade restrictions on advanced chips, potentially widening the gap between the two nations, experts have indicated.
Despite high demand for Nvidia’s H20 chip, which was designed to circumvent export control licenses, trade restrictions are slowing China’s ability to advance its large language models (LLMs) that drive AI technologies like ChatGPT. Analysts from Jeffries highlighted this issue in a recent call, noting that China’s technological gap with the U.S. could either remain the same or widen.
The H20 chip, while not as powerful as Nvidia’s other chips sold to U.S. tech firms, has led Chinese companies to consider Huawei’s Ascend chip as a long-term alternative. However, experts express concerns over whether China has the advanced 7-nanometer process capacity required to meet future demand.
Reportedly, Huawei is struggling to ramp up production of its Ascend 910B chip, considered the best domestic alternative to Nvidia’s restricted chips, due to breakdowns in repurposed chip fabrication machinery.
U.S. officials have pointed out that another Chinese chip, the Kirin 9000s, which utilizes 7-nanometer technology and powers Huawei’s Mate 60 Pro smartphone, is still not as advanced as current U.S. chips. Secretary of Commerce Gina Raimondo remarked that the inferior quality of this chip shows that U.S. export controls are effective, emphasizing that the U.S. leads in semiconductor sophistication.
In contrast, Arati Prabhakar, director of the White House Office of Science and Technology Policy, noted that while the U.S. has decreased federal research and development funding, China has increased its R&D spending by 10%. During a recent AI boom, Prabhakar stressed the need for the U.S. to boost its efforts to remain competitive.
Experts also predict demand for computing power to continue growing as companies build larger and more advanced models. They warn that China’s AI industry may need to consolidate due to limited capital and computing resources.
The shortage of computing power could affect China’s global competitiveness in AI. While some experts suggest optimizing models to perform well with less computing power, others believe that the initial development stages of AI models, which require extensive computing resources, are crucial before optimization can be effectively implemented.