Considering higher production to meet rising Chinese demand, NVIDIA reviews expanding H200 chip output globally and supply planning.
NVIDIA is evaluating an increase in production capacity for its H200 artificial intelligence chips after demand from Chinese customers exceeded current output levels, according to sources familiar with the matter.
The potential move follows comments by US President Donald Trump that Washington would permit NVIDIA to export the H200, its second fastest AI processor, to China, subject to a 25% fee on sales. Chinese demand has surged since the announcement, prompting NVIDIA to review whether additional capacity is required.
NVIDIA has informed Chinese clients that it is assessing options to expand output, the sources said, adding that discussions remain private. An NVIDIA spokesperson said the company was managing its supply chain to ensure licensed sales to authorised Chinese customers would not affect its ability to supply clients in the United States.
Major Chinese technology firms, including Alibaba and ByteDance, have already contacted NVIDIA about placing large H200 orders, Reuters reported earlier. However, uncertainty remains as Chinese regulators have yet to approve any purchases. Officials held emergency meetings to decide whether shipments of the chip would be allowed into the country, according to people briefed on the talks.
Supply of the H200 is currently limited, as NVIDIA prioritises production of its more advanced Blackwell chips and its upcoming Rubin line. The H200, part of NVIDIA’s Hopper generation, entered mass deployment last year and is produced by TSMC using a four nanometre manufacturing process.
Chinese companies are seeking clarity on availability as the H200 represents the most powerful AI chip they can currently access, amid Beijing’s push to strengthen its domestic AI industry.
