Key Points
- Nvidia has verbally instructed partner data centers worldwide to ensure at least 51 percent of their service capacity is leased to non-Chinese firms under new contracts.
- The restriction closes a major workaround used by Chinese tech companies that leased overseas cloud infrastructure to train AI models despite U.S. chip export bans.
- Cloud operators across Asia, especially in Southeast Asia where Chinese enterprises are anchor clients, face significant commercial fallout from the new mandate.
Summary Nvidia has reportedly instructed its international data center partners to curb their business with Chinese tech clients, escalating the global struggle over high-performance artificial intelligence chips. According to a report by the Financial Times, Nvidia recently notified cloud operators that going forward, at least 51 percent of their computing capacity must be allocated to non-Chinese enterprises.
This directive marks an unprecedented restriction on cloud infrastructure. Rather than issuing formal written contracts, Nvidia conveyed the guidelines verbally through its sales teams and regional representatives worldwide. The policy applies specifically to new agreements, while leaving existing contracts intact. In its official response, Nvidia emphasized that data centers and cloud services built with American chips serve as the foundational backbone allowing U.S. artificial intelligence firms to compete globally, aligning with long-standing U.S. policy to expand American technological infrastructure worldwide.
For Chinese tech companies, this development closes a crucial operational workaround. Strict export bans already bar Chinese firms from acquiring state-of-the-art U.S. AI chips directly. To circumvent these restrictions, many companies leased computing power from overseas data centers located in neutral hubs across Southeast Asia to train complex AI and large language models. Under the Trump administration, U.S. regulatory authorities have intensified enforcement against gray-market intermediaries inside China while simultaneously moving to plug the loophole of offshore data center leasing.
The sudden mandate has caused significant disruption for global cloud operators. Regional data center hubs, notably in Southeast Asia, rely heavily on Chinese tech conglomerates as their primary clients. As one Singapore-based cloud operator noted, finding sufficient non-Chinese clients outside the United States presents a severe hurdle, warning that strict enforcement of the 51 percent quota will deal a major blow to their business model. For tech observators in East Asia and beyond, the move highlights how geopolitical supply chain controls are reshaping the digital landscape, forcing international cloud providers to choose between American silicon access and their primary commercial customers.
Sponsored · Ad