Beijing
In an effort to fortify existing regulations on the export of AI chips to China, the United States government has outlined plans to curb the potential circumvention of restrictions by American chipmakers. According to a U.S. official, new guidelines, set to be announced this week, will target specific AI chips, bolstering the comprehensive measures implemented last October. The revised rules will prevent the export of certain advanced datacenter AI chips that were previously not subject to restrictions. While the official did not specify the exact chips to be affected, industry sources have suggested that the prominent semiconductor, Nvidia’s H800, is likely to be included.
Despite the introduction of these stricter measures, the U.S. emphasized that consumer-grade chips, predominantly used in products like laptops, would remain exempt from the new regulations. However, companies dealing with such chips will be required to report their activities to ensure compliance with national security objectives. Additionally, the removal of certain parameters used to restrict chip exports may lead to a decrease in the communication speed of AI chips, rendering AI development more challenging and cost-intensive.
While the regulatory focus remains primarily on AI chip exports, the U.S. administration clarified that no immediate restrictions on access to U.S. cloud computing services or those of its allies were anticipated. Nonetheless, the government intends to solicit feedback on the associated risks and potential mitigation strategies. This comprehensive approach underscores the U.S.’s commitment to managing the complexities arising from the intersection of technology and national security, with the ultimate goal of fostering stability in U.S.-China relations.
