Despite the restrictions imposed by Washington, the Asian giant continues to take advantage of US technology.
In the current global dispute between the United States and China, one of the areas of greatest concern to Washington is semiconductors. For this reason, in recent months it has imposed significant restrictions on Beijing.
However, US microchip export controls, to freeze the development of Chinese supercomputers to produce nuclear weapons and artificial intelligence systems, have minimal effect on the Asian giant’s technology sector.
The rules limit shipments from Nvidia Corp and Advanced Micro Devices. For example, to comply with the regulations, Nvidia has created variants of its chips for the Chinese market, such as the Nvidia H800, which will likely take 10-30% more time to perform some AI tasks and may double some costs compared to faster US. chips.
But even the slower Nvidia chips represent an improvement for the Chinese company: Tencent Holdings estimated in April that the system using the H800 will cut the time it takes to train its largest AI system by more than half, out in eleven days to four days.
The US government’s goal is to slow China’s progress in high technology without hurting domestic companies. So part of the strategy was to avoid such a shock that the Chinese would abandon American chips completely and redouble their own development efforts.
The export restrictions have two parts: One puts a limit on a chip’s ability to calculate extremely precise numbers, a measure designed to limit supercomputers that can be used in military research.
But this is less important in AI tasks, such as large language models, where the amount of data the chip can process is more important.
The second limitation is on chip-to-chip transfer speed, which affects the AI. The models behind the technologies, such as ChatGPT, are too large to fit on a single chip, so they must be spread over many chips that need to be used. communicate with each other
