© Reuters. Semiconductor chips are seen on a computer circuit board in this illustrative photo taken February 25, 2022. REUTERS/Florence Lo/Illustration/Files
By Jane Lanhee Lee and Stephen Nellis
(Reuters) – Deciding who gets hurt by sweeping new U.S. restrictions on selling technology to China will partly depend on what constitutes a “supercomputer”, experts told Reuters.
The semiconductor industry around the world began to grapple Friday with sweeping U.S. restrictions on the sale of chips and chipmaking equipment to China.
Shares of chip-equipment makers fell, but industry experts said a new US definition of a supercomputer could be key to the new rules’ impact on China.
Supercomputers can be used to develop nuclear weapons and other military technologies, and experts say their definition has long perplexed regulators trying to pin down an ever-changing technological target.
The new US rules broadly define supercomputers in terms of computing power in a defined space: a 100 petaflops machine – the capacity to perform 100 trillion operations per second – in 41,600 cubic feet, with a few other keep.
Senior government officials told a news conference their intention was to target only China’s most advanced systems that could pose a threat to US national security rather than commercial activities.
But experts have questioned whether the densely populated data centers of Chinese tech giants owned by people like Ali Baba (NYSE:) Group Holding or ByteDance, which owns TikTok, could soon achieve supercomputer status based on the new definition, though that’s not what US regulators had expected.
“Data center builds like Alibaba or ByteDance would have the potential to hit petaflops,” said CCS Insight chip analyst Wayne Lam.
The new definition is unlikely to change as industry technology improves. China’s current supercomputers may one day become the standard for businesses, but they will still face limits imposed on Friday to prevent any chip made with American equipment or technology from entering China. Enterprises “could very well run into supercomputing limitations over the next couple of years,” Lam said.
Jack Dongarra, a computer science professor who helps lead a group called TOP500 that ranks the world’s fastest supercomputers, said he disagreed with the static definition.
“The problem is that the definition of a supercomputer will change over time,” he said by email.
Major Chinese companies with large data centers such as Baidu (NASDAQ:), Alibaba and ByteDance did not immediately respond to requests for comment. Tencent declined to comment.
Defining computing power per cubic foot can also provide room for creative workarounds. For example, said one expert, use fiber optic cables to link immense computing power across a larger space.
“They could spread their supercomputers over a larger space,” said a chip and data center expert who requested anonymity due to the politically charged nature of the new rules.
“The average supercomputer architect would say, ‘That’s not how things are done!’ But not being able to do it any other way breeds a lot of creativity and the will to do things differently.”