The new US rules define supercomputers broadly in terms of computing power in a defined space
Deciding who will be hurt by the new US curbs to sell technology to China will come down in part to what constitutes a “supercomputer”, experts told Reuters.
Around the world, the semiconductor industry began grappling on Friday with sweeping U.S. restrictions on the sale of chips and chip-making equipment to China.
Shares of chip makers fell, but industry experts said the new US definition of a supercomputer could be key to the impact of the new rules on China.
Supercomputers can be used in the development of nuclear weapons and other military technology, and experts say how to define them has long eluded regulators trying to determine the ever-advancing technology target.
The new US rules define supercomputers broadly in terms of computing power in a defined space: a machine with 100 petaflops – the ability to perform 100 trillion operations per second – in 41,600 cubic feet, with some other caveats.
Senior government officials said at a media press conference that their intention was to target only the most advanced Chinese systems that could pose a national security threat to the United States rather than commercial activity.
But experts wondered whether the densely packed data centers of Chinese tech giants owned by the likes of Alibaba Group Holding or TikTok owner ByteDance could soon reach supercomputer status based on the new definition, even if that’s not what U.S. regulators intended.
“Building data centers like Alibaba or ByteDance would have the potential to reach petaflop builds,” CCS Insight chip analyst Wayne Lam said.
The new definition is unlikely to change as industrial technology improves. China’s current supercomputers may one day become the corporate standard, but they still face limits that were put in place on Friday to prevent any chip made with American equipment or technology from entering China. Companies “may very well run into the limitations of supercomputers in the next few years,” Lam said.
Jack Dongarra, a computer science professor who helps lead a group called TOP500, which ranks the world’s fastest supercomputers, said he disagrees with a static definition.
“The problem is that the definition of a supercomputer changes over time,” he said by email.
Major Chinese companies with large data centers such as Baidu, Alibaba and ByteDance did not immediately respond to requests for comment. Tencent declined to comment.
Defining computing power per cubic foot can also offer room for creative solutions. For example, one expert said, use fiber optic cables to connect massive computing power over a larger area.
“They could spread their supercomputers over a larger area,” said one chip and data center expert, who spoke on condition of anonymity because of the politically charged nature of the new rules.
“The average supercomputer architect would say: This is not the way things are done! But not being able to do it any other way breeds a lot of creativity and a willingness to do things differently.”