Nvidia has reportedly instructed its data center and cloud computing partners to cap their business with Chinese clients at 49%, requiring the majority of new contracts be sold to non-Chinese companies. This verbal directive represents a fresh restriction in the global AI infrastructure landscape.
- New contracts must allocate majority cloud services to non-Chinese customers
- Restriction applies only to new business; existing contracts are unaffected
- Aligns with US policy to maintain AI infrastructure leadership over China
What happened
Nvidia, a key supplier of AI chips and data center technology, has told its cloud computing partners to limit the portion of business conducted with Chinese customers to no more than 49%. This instruction targets new contracts and has been communicated verbally via Nvidia's sales and regional teams rather than through formal written mandates. Existing agreements remain in place without such limits.
The company has not officially confirmed or denied the reported directive but justified the policy by emphasizing the need for American AI infrastructure to dominate globally. This step is unprecedented for Nvidia and comes amid intensifying geopolitical concerns about AI technology transfer to China.
Why it matters
This limitation introduces an additional barrier for Chinese technology firms that depend heavily on cloud-based computing resources outside China to train advanced AI models. The restriction fits into a wider US strategy aimed at preventing access to cutting-edge semiconductors and AI infrastructure that could enhance Chinese AI capabilities.
Despite existing US export controls and previous restrictions on Nvidia’s highest-end chips, substantial amounts of AI technology still reach China, sometimes through smuggling and indirect channels. Nvidia’s new directive could tighten control over these supply chains by curbing the availability of cloud services that facilitate AI research and development in China.
What to watch next
Industry observers will be monitoring how cloud service providers implement Nvidia’s guidance and whether the rule influences contract negotiations or cloud service allocations globally. It may also impact Chinese AI firms’ reliance on foreign infrastructure, potentially accelerating Beijing’s move toward domestic alternatives.
Regulators and policymakers will continue to evaluate how these restrictions shape technology competition between the US and China, while enforcement challenges such as smuggling remain key concerns. Market reactions from Nvidia partners and customers may also hint at broader industry shifts in response to tightening geopolitical technology controls.