Nvidia (NASDAQ:NVDA) has reportedly instructed its data center partners to reduce their exposure to Chinese customers, introducing new requirements that could make it more difficult for Chinese technology companies to access advanced artificial intelligence computing capacity through overseas cloud services.
According to a Financial Times report published Tuesday, citing five people familiar with the matter, Nvidia now requires cloud computing customers to generate at least 51% of their service sales from companies outside China.
The reported policy represents a change from Nvidia’s previous approach, which did not impose comparable restrictions on the geographic composition of its cloud partners’ customer bases.
Nvidia Introduces New Requirements for Cloud Partners
Under the reported requirements, Nvidia’s cloud computing partners must ensure that non-Chinese companies account for at least 51% of their service sales.
The Financial Times said the instructions were communicated verbally through Nvidia’s sales representatives and regional teams rather than through formal written documentation.
The company had not previously imposed such limits, according to the report.
The new requirement could affect cloud providers that rely heavily on Chinese customers seeking access to Nvidia-powered computing infrastructure.
Partners with substantial exposure to Chinese businesses may need to adjust their customer mix to comply with the reported policy.
The Financial Times attributed the information to five people with knowledge of the matter.
Chinese AI Developers Face Additional Computing Restrictions
The reported measures could create further challenges for Chinese technology companies developing artificial intelligence systems.
Many Chinese businesses have relied on overseas cloud computing services to access advanced processing capacity, particularly for computationally intensive activities such as training AI models.
These arrangements have allowed companies to rent computing resources located outside China rather than purchasing and importing restricted hardware directly.
Nvidia’s reported requirements could limit the ability of overseas data center operators to concentrate their services on Chinese customers.
The restrictions may therefore affect access to advanced AI infrastructure for companies that depend on international cloud providers.
However, the report does not establish that all Chinese customers would be prohibited from using Nvidia-powered cloud services.
Instead, the reported requirement would limit the proportion of services that participating cloud providers could sell to Chinese companies.
Washington Continues Efforts to Restrict Advanced Chip Access
The development comes as the United States continues efforts to prevent Chinese companies from obtaining advanced American semiconductor technology through alternative channels.
Washington has maintained restrictions on sales of certain high-performance chips to China since at least 2022, targeting technologies that can support sophisticated artificial intelligence applications.
Overseas cloud computing arrangements have become an area of concern because they can provide access to advanced processing capabilities without requiring restricted chips to be physically delivered to China.
The reported Nvidia policy would introduce an additional commercial restriction alongside existing US export controls.
By limiting Chinese customers’ share of cloud service sales, the requirements could reduce the concentration of Chinese demand among Nvidia’s international data center partners.
The measures also highlight the increasing importance of cloud infrastructure in the broader debate over access to advanced AI computing technology.
Nvidia Hardware Smuggling Cases Add to Regulatory Concerns
US authorities have also investigated attempts to move restricted Nvidia technology into China despite export controls.
Reports of Nvidia chips being smuggled into the country have emerged in recent years, raising questions about enforcement of restrictions on advanced semiconductor exports.
Earlier this year, several senior employees at server manufacturer Super Micro were charged in connection with the alleged smuggling of approximately $2.5 billion worth of Nvidia AI servers to China.
Super Micro assembles servers incorporating Nvidia processors, making its equipment part of the infrastructure used for advanced computing workloads.
The charges concern alleged misconduct and do not establish guilt.
The case illustrates the challenges facing US authorities as they seek to enforce restrictions across complex international semiconductor supply chains.
Implications for Nvidia’s Data Center Business
Nvidia’s reported policy could have consequences for cloud computing providers that operate AI infrastructure using its processors.
Companies with significant Chinese customer bases may face additional requirements when selling computing services, potentially affecting their commercial strategies.
For Chinese AI developers, the restrictions could make overseas computing arrangements less accessible, particularly where providers have historically relied on demand from China.
The reported measures also demonstrate how access to advanced AI technology is increasingly being influenced by restrictions extending beyond direct semiconductor exports.
Nvidia’s data center partners may need to consider customer geography alongside existing regulatory requirements when planning future sales and infrastructure capacity.
Outlook
Nvidia’s reported decision to require cloud partners to sell at least 51% of their services to non-Chinese companies would represent a further restriction on Chinese access to advanced AI computing resources.
The Financial Times report suggests that the chipmaker is taking steps to reduce Chinese customer exposure among its international data center partners as Washington continues tightening controls over advanced semiconductor technology.
The measures could complicate overseas cloud computing arrangements used by Chinese AI developers, while creating additional compliance considerations for Nvidia’s partners.
Investors will be monitoring whether the reported requirements are formally confirmed, how they are implemented and whether they affect demand for Nvidia-powered infrastructure.