Nvidia's annual revenue at its Singapore subsidiary hit $104 billion in 2025 before the company stopped disclosing those figures, illustrating the massive scale of regional compute demand now cut off by regulatory restrictions. Acting CEO Colette Kress stated Tuesday that Nvidia will no longer sell its H200 chips in China, dismantling a mechanism that previously allowed hardware shipments under a 25% tariff structure. This operational pivot follows the departure of co-founder Jensen Huang last week, which occurred in part over an affair with the CEO of a company under investigation for smuggling chips into China. To manage the geopolitical fallout, Nvidia announced it was creating a board-level committee to oversee national security policies, export control compliance, and global trade risks.
This retreat marks a decisive end to complex regulatory accommodations, shifting capital flows directly toward alternative regional hardware solutions. The abandoned H200 sales route had been brokered by Huang and President Donald Trump, requiring the chips to be shipped through the U.S., allowing the government to collect a 25% tariff. Huang long argued that it was more important for the U.S. to compete in China, telling podcaster Dwarkesh Patel in April that conceding that market for the United States technology industry is a disservice to the country. The abrupt termination of that policy leaves a massive revenue channel exposed.
The Cost of Offshore Compute Bans
The financial reckoning extends well beyond direct physical shipments into mainland facilities. New rules, issued Monday, bar advanced U.S. chips from foreign data centers that serve Chinese customers remotely. This broad restriction follows earlier tightening on November 4, 2026, when Nvidia restricted Chinese AI access via cloud providers under tightening international limits. How deep will the resulting contraction run in the hardware supplier's quarterly balance sheets?
"Investment bank Jefferies forecast that those rules might trim between $44 billion and $62 billion from Nvidia’s data center revenue next year."
Investment bank Jefferies forecast that those rules might trim between $44 billion and $62 billion from Nvidia’s data center revenue next year. Critics of Nvidia’s dependence on China claim that as much as 20% of its revenue comes from demand linked to the country — annual revenue at its Singapore subsidiary hit $104 billion in 2025 before Nvidia stopped disclosing those figures. Over the longer horizon, the company may face the loss of potential customers in Singapore, Malaysia, Thailand, and Indonesia.
Backlog Assumptions Meet Cloud Enforcement
Nvidia maintains that surging global demand will absorb the shortfall without damaging top-line performance. Referencing the new export controls, Kress stated that the backlog of demand for advanced chips like Vera Rubin means the rules would have no impact on the company’s top-line revenue. Corporate leadership is betting that unfulfilled orders for the upcoming Vera Rubin architecture will immediately substitute for eliminated international contracts.
Yet the structural reality inside regional infrastructure reveals major friction. Even before Nvidia decided to stop shipping H200s, China had largely refused to accept their delivery, with Georgetown's Center for Security and Emerging Technology researcher Jacob Feldgoise noting that the government would prefer to generate demand for the country’s domestic chip industry, even if major AI compute users want the still-superior Nvidia equipment. Meanwhile, all of China’s leading AI models show evidence of having been trained on Nvidia Grace Blackwell chips that are prohibited from being exported to China.
"If they can’t provide that certification, the new rules may force them toward other chip options and further eat into Nvidia’s market share."
If large data centers suited for training AI models want to buy Nvidia chips, they will have to certify they aren’t serving any Chinese customers, as export control attorney Dan Fisher-Owens told TechCrunch. Data center operators unable or unwilling to sign those guarantees are being pushed toward non-U.S. hardware alternatives. Nvidia projects zero top-line damage based on an uninterrupted queue for Vera Rubin silicon, while its largest international hosting hubs are legally barred from taking its hardware if they service one of the world's most compute-intensive client bases.
What came before
The 25-percent tariff scheme for shipping H200 hardware through the US survived solely on Jensen Huang's personal maneuvering, but following his abrupt exit, that compromise vanished entirely. Interim leader Colette Kress insists that Vera Rubin backlogs will easily soak up the blow, brushing aside the immediate evaporation of a massive regional client pool. Will the Vera Rubin backlog absorb Jefferies' projected hit of up to $62 billion in data center revenue once access from Singapore, Malaysia, and beyond is barred?
Our stories: Nvidia Board Probes CEO Jensen Huang Over Megaspeed Ties (November 9).