Nvidia Reportedly Tells Data Centre Partners to Reduce Exposure to Chinese Customers

Nvidia, the US technology company specialising in graphics processing units (GPUs) and semiconductor technology, has reportedly instructed its global data centre and cloud computing partners to reduce their exposure to Chinese customers, according to the Financial Times (FT).

The move could make it harder for Chinese technology companies to access advanced AI computing capacity through overseas providers. Nvidia's position in the AI hardware market has also made its commercial decisions increasingly significant for companies building and operating AI infrastructure.

According to the FT's November 3, 2026 report, the chipmaker has instructed cloud computing customers to direct at least 51 per cent of their services towards non-Chinese customers.

The instructions have reportedly been communicated verbally through Nvidia's sales teams and regional representatives, rather than through formal written agreements. They apply to future contracts and do not affect existing leasing arrangements, according to people familiar with the matter.

Reuters reported the development but said it could not immediately verify the FT's findings. The reported guidance applies to Nvidia's global data centre partners, although its impact on individual providers remains unclear.

How Nvidia's Restrictions Could Affect Chinese AI Development

The move comes as the United States continues to tighten restrictions on China's access to advanced semiconductor technology. Nvidia GPUs are widely used to train and run AI models, making them an important resource for companies developing AI systems. US export controls have barred some of the company's most advanced processors from reaching Chinese customers, prompting some Chinese technology firms to seek alternative computing resources.

One option has been to lease computing power from overseas data centres and cloud providers that already have access to the hardware. This allows companies to build AI with advanced processors without having to buy and import the physical chips themselves.

The practice has caught the eye of US policymakers who are worried about access to cutting-edge computing technology abroad. The Congressional Research Service, in a report in April 2026, discussed the potential for foreign companies to remotely access advanced computing chips through cloud services. It warned that these arrangements could erode the effectiveness of present export controls.

The report also addressed proposed legislation to impose export controls on remote access to certain computing resources. These proposals reflect concerns that restrictions on physical chip sales may not be sufficient to control how advanced computing technology is accessed and used.

Nvidia's guidance could make overseas computing arrangements more difficult for Chinese firms by encouraging data centre partners to allocate a larger proportion of their services to non-Chinese customers. The FT reported that operators in Southeast Asia could be particularly affected, given the importance of Chinese technology companies to some regional markets.

One Singapore-based cloud provider reportedly warned that finding enough non-Chinese customers could be difficult. Strict enforcement of the guidance could therefore affect its business.

Data Centre Operators Face New Compliance Challenges

The guidance also presents challenges for data centre operators that have invested in Nvidia-powered infrastructure to serve international customers. If they have to restrict the proportion of services provided to Chinese customers, operators may need to find alternative sources of demand to support their investments.

The FT also reported that Nvidia has already reduced the number of authorised customers in Asia by more than half following the introduction of stricter compliance checks. These checks are intended to prevent advanced AI chips from being diverted to unauthorised buyers or reaching customers in restricted markets.

The latest guidance would extend this approach beyond direct hardware sales, placing greater attention on how computing services are distributed among end users. This could require data centre operators to introduce additional customer verification and compliance procedures.

Nvidia Responds to Reported Changes

Nvidia has not publicly confirmed the specific 51 per cent customer threshold described by the FT. In a statement quoted by the publication, the company said that data centres and cloud services built on American chips were important in helping US AI model and application developers compete globally. It also said that longstanding US policy supported the expansion of American technology infrastructure worldwide.

A separate report by The Information on November 2 provided further context, describing Nvidia's instructions as guidance intended to encourage cloud providers, including Nvidia-backed neocloud companies, to diversify their customer base.

According to three people familiar with the matter cited by the publication, the guidance was not mandatory and had not been enforced as a strict requirement preventing cloud providers from accepting Chinese customers. The publication also reported that Nvidia was asking cloud customers to identify the end users of its chips.

The practical impact will therefore depend partly on how providers interpret and apply the guidance.

US-China Technology Restrictions Continue to Evolve

The move underscores the increasing complexity of US efforts to restrict China’s access to sophisticated artificial intelligence (AI) technology. Historically, export controls have centered on semiconductor sales, but overseas cloud computing arrangements have added new headaches for regulators.

The latest guidance shows the growing importance of monitoring not just where advanced chips are sold, but how their computing capacity is accessed and deployed.

For Nvidia, the approach could be another way to strengthen oversight of its technology. But it also could spell trouble for cloud providers that rely on international customers to help pay the bills for their data center investments. The trade-off between technology access, commercial interests and national security will likely continue to be a key consideration as US-China technology restrictions evolve.

It’s not yet clear how fully the changes will be implemented and, in particular, how consistently the guidance will be applied across Nvidia’s data center partner network. Now the development adds another layer to the restrictions on advanced AI computing technology and the commercial relationships that support its global distribution.