The U.S. administration under Donald J. Trump has given the go‑ahead for Nvidia to sell its H200 artificial‑intelligence chips to approved customers in China. The decision marks a significant reversal from earlier restrictions, and Trump has justified the move by saying it supports American jobs, boosts manufacturing, and benefits U.S. taxpayers.
However, the approval comes with strings attached: the chips can only go to vetted customers, and a “25% fee” on these sales will go to the U.S. — a detail Trump highlighted in his statement. Importantly, the deal does not cover Nvidia’s latest-generation chips (such as the “Blackwell” models) or the upcoming “Rubin” line, which remain off‑limits for China.
Despite U.S. approval, obstacles remain. According to a recent report, regulators in Beijing are reportedly planning to limit access to the H200 chips — potentially allowing only restricted or limited use inside China. This could undermine the very business case that motivated Nvidia and the U.S. decision in the first place.
For Nvidia and other U.S. chipmakers, the broader implications are uncertain. On one hand, the export permission could open a major market; on the other, China’s hesitancy and its push for domestically built AI hardware may limit demand. As of now, analysts warn that the approval alone may not significantly boost Nvidia’s China business unless more advanced chips are allowed for export in the future.
