By Jonathan O’Callaghan
China’s AI sector could be five to ten years behind the United States, partly because it cannot access the most advanced computer chips.
In 2022, the United States imposed export controls on the fastest chips and on equipment that could be used to produce such chips. Even foreign suppliers that produce chips for the US market won’t sell to Chinese developers to avoid falling foul of US sanctions. This has “dramatically limited” China’s progress with training AI models, says computer engineer Yiran Chen.
The AI boom has sparked a global race to produce increasingly powerful computer chips that can cope with the large data sets needed to train and execute models. Nvidia, one of the leading developers of such chips, based in Santa Clara, California, saw its market value shoot past US$2 trillion for the first time last March.
Export controls imposed by the US Department of Commerce in October 2022, and subsequently tightened a year later, prohibited the sale of certain technologies to China, including chips that can operate above speeds of 300 teraflops, or 300 trillion operations per second. It also limited the sale of state-of-the art manufacturing equipment that could be used to produce such chips.
Generations of chip design are labelled according to nanometre ratings, with smaller numbers denoting more advanced chips. Nvidia’s latest chip — the GB200 Blackwell, intended to sell at $30,000–$40,000 per chip according to chief executive Jensen Huang — is rated at just 4 nanometres, approaching the width of a strand of human DNA.
Nvidia and other US companies, such as Intel, along with Samsung in South Korea, are now transitioning to 3 nm technology and even pushing down towards 2 nm.
China’s best efforts, from Huawei, are still at around 7 nm, meaning that Chinese companies have to use more chips to achieve the same computing power as one advanced chip.