According to Nvidia, the Nvidia A800 graphic processing unit is “another alternative product to the Nvidia A100 GPU in China. “The A800 meets the clear test for reduced export control established by the US government. And cannot be programmed to exceed it.” Reuters was the first to report on the new chip on Monday.
Two months after the United States blocked China’s access to two of Nvidia’s high-end microchips, the American semiconductor design giant unveiled a replacement with a slower processing speed for its second-largest market.
The A100 processor has been used to power supercomputers, artificial intelligence, and high-performance data centers in industries ranging from biotech to finance to manufacturing.
One of its customers has been Alibaba’s cloud computing business. The A100 and Nvidia’s enterprise AI chip H100 have been placed on a US export control list. It listed to “address the risk that the covered products may be used in. Or diverted to, a military end use’ or military end user’ in China and Russia.”
Nvidia previously stated that the US ban cost the company up to $400 million in potential third-quarter sales to China. So the new chip appears to be an attempt to make up for the financial loss. According to an Nvidia spokesperson, the A800 GPU went into production in Q3.
Indeed, Chinese chip distributors like Omnisky are already marketing A800 in their product catalogs. And this chip has been designed to avoid US export regulations while still performing other core computing functions.
Nvidia isn’t the only company slowing down its chips to avoid US sanctions. According to the Financial Times, Alibaba and Chinese chip design startup Biren are modifying the performance of their latest semiconductors. This is because they outsource their manufacturing to Taiwan’s TSMC.