技术封锁终将成自主创新“磨刀石”

Core Viewpoint - The recent U.S. policy allowing Nvidia to sell H200 AI chips to "approved customers" in China, with a 25% revenue share to the government, raises questions about the value of this transaction for Chinese buyers, highlighting issues of cost-effectiveness, reliability, and competition in the tech sector [1][2][3][4] Cost-Effectiveness - The policy shift indicates a move from a complete blockade to a limited opening with economic ties, as the U.S. previously halted H20 chip exports, causing Nvidia's market share in China to plummet from 95% to 0% [1] - The H200 chip, while being a higher-performance product, is still considered a "second-tier" option compared to Nvidia's Blackwell series, which remains banned for sale in China, thus reducing the cost-effectiveness for Chinese users [1] Reliability Concerns - The temporary lifting of the ban on the H200 chip does not restore trust, as Chinese companies fear potential supply disruptions or scrutiny after purchasing the chips [2] - There are widespread suspicions regarding the possibility of "backdoors" in the Nvidia chips, particularly after reports of advanced tracking and remote shutdown capabilities, which further erodes market confidence [2] Competition Landscape - Domestic alternatives to Nvidia chips are emerging, with Chinese AI chips now included in official procurement lists, indicating a shift towards a self-sufficient technology ecosystem [3] - The support from over a trillion yuan in industrial funds and collaborative innovation efforts are enabling domestic chips to compete in various sectors, including government and finance [3] - The market is expected to evolve into a dual-track model of "diverse procurement + self-reliance," where imported chips may be used in specific scenarios while prioritizing domestic technology development [3][4]