Things are getting testy again between China and the U.S.
As the two nations continue their spectacular competition for global AI supremacy, the Chinese government is reportedly leaning into a familiar tactic: Made in China.
According to a new Reuters report, Beijing regulators will require new data center projects that have received state funds to use domestically-made AI chips.
The mandate affects cloud computing facilities that are less than 30% complete “to remove all installed foreign chips”—that is, those made by AMD, Broadcom, Intel, and Nvidia—“or cancel plans to purchase them,” according to the report.
Anything further along will be sorted on an individual basis.
Aggressive? You betcha—particularly since it’s been just one week since Xi Jinping and Donald Trump met in South Korea and agreed to ease some of the trade hostilities between the two nations.
This is, of course, precisely what Jensen Huang had feared. The Nvidia CEO argued earlier this year that U.S. restrictions on the export of the most advanced U.S.-made AI chips to China would backfire by encouraging the Chinese government to invest in domestic firms.
A Chinese sort-of ban on coveted American-made chips—including products the U.S. has itself barred from export for competitive reasons—just adds more fuel to the fire.
—AN