The US Commerce Department notified Nvidia last week that the hardware would now require clearance for export — a rule Nvidia said will apply “for the indefinite future.”

Writing in its Bog, Nvidia said: “The [government] indicated that the license requirement addresses the risk that the covered products may be used in, or diverted to, a supercomputer in China.”

The restrictions arrive as part of an intensifying scrap between Washington and Beijing over tech supremacy, with semiconductors at the centre of a bitterly escalating trade war. Both sides have thrown up tariffs across a growing list of goods.

Shares in Nvidia dropped nearly six per cent in after-hours trading following the announcement. The firm said the multi-billion-dollar charge was tied to inventory, purchase commitments, and reserves linked to the H20 chips.

Counterpoint Research analyst Marc Einstein said the figure was expected. “While this is certainly a lot of money, this is something Nvidia can bear,” he said.

Einstein added, “As we have seen in the last few days and weeks, this may largely be a negotiating tactic. I wouldn't be surprised to see some exemptions or changes made to tariff policy in the near future.”

The company took a blow in January when Chinese developers behind the DeepSeek AI chatbot claimed it was built for a fraction of the cost of Western rivals — a sign, some said, that the US had underestimated Beijing’s speed.

Tech Buzz China founder Rui Ma said the long-term outlook could see a full tech break-up. “It doesn't make any sense for any Chinese customer to be dependent on US chips.” She added that China is already flush with excess data centre capacity.

She warned that if the restrictions continue, AI semiconductor supply chains will likely become “fully decoupled.”