Markets were rocked this week after DeepSeek claimed its large language model outperforms OpenAI’s but cost a fraction of the price to train. However, questions were soon raised over the provenance of the semiconductors used to build DeepSeek’s R1 reasoning model given U.S. restrictions on exporting advanced AI chips in China.
A Nvidia spokesperson told CNBC Monday that the chips used by DeepSeek were fully export-compliant. DeepSeek was not immediately available for comment when contacted by CNBC.
“We expect US companies, like Nvidia, to comply with US export controls and our domestic legislation. Our customs and law enforcement agencies will continue to work closely with their US counterparts,” MTI said in its statement.
“We have always upheld the rule of law, and acted decisively and firmly against individuals and companies that flout the rules.”

MTI cited Nvidia’s comments in its Saturday statement and said the chipmaker said there was no reason to believe that DeepSeek had obtained any export-controlled products via Singapore.
“Singapore is an international business hub. Major US and European companies have significant operations here. Nvidia has explained that many of these customers use their business entities in Singapore to purchase chips for products destined for the US and other Western countries,” MTI added.
— CNBC’s Ryan Browne contributed to this report.
NVIDIA Corp,Singapore,United States,Breaking News: Technology,Technology,Breaking News: Politics,Politics,business news
#DeepSeek #Nvidia #chip #questions #Singapore #responds