“It will give other vendors’ products, such as Intel’s CPUs, GPUs, FPGA, ASICs from major CSP, and AI chips from Chinese fabless companies, an opportunity to catch up with Nvidia or AMD’s current level of single-chip performance,” Zeng said. “But, in addition to single-chip performance, Nvidia and AMD still have advantages when it comes to new products, interconnect solutions, and AI model technology in the future.”
Another point to note here is that AMD already has a weak presence in China. Manish Rawat, a semiconductor analyst at Techinsights, pointed out that the latest hurdle could prompt rivals to take advantage of this to solidify their position in the market and increase their market share.
“Meanwhile, Chinese tech firms may accelerate efforts to develop indigenous AI chip technologies, intensifying competition in the long run,” Rawat said. “Increased R&D spending by companies is probably going to result from this heightened competitiveness, which will also spur greater innovation overall. The decision may, strategically speaking, heighten tensions between the US and China and trigger retaliatory actions that might harm US tech companies operating in China.”
Effectiveness of the restrictions
Perhaps a more significant question is the success of US restrictions in denying Chinese companies access to the most advanced technology.
The FT report pointed out that the Chinese government’s initiative to assist AI startups was prompted by internet companies and cloud service providers terminating contracts. This shift occurred as tighter US regulations led these companies to reserve GPUs for their own use.
Chinese tech behemoths Alibaba, Tencent, and ByteDance are restricting the rental of Nvidia’s GPUs, keeping the bulk of their hoarded AI processors for in-house applications and key customers.
