United States chip major Nvidia is reportedly considering a significant increase in the production capacity for its H200 artificial intelligence chips. This decision comes as the demand for these advanced AI chips from Chinese companies has substantially exceeded the current output, according to sources cited by Reuters.

A spokesperson for Nvidia addressed Reuters' inquiries, stating that the company is actively managing its supply chain to ensure that licensed sales of the H200 to authorized customers in China will not adversely impact its ability to supply customers in the United States. While Nvidia confirmed its supply chain management, TSMC (Taiwan Semiconductor Manufacturing Company) and China's Ministry of Industry and Information Technology (MIIT) did not provide immediate responses to queries regarding the matter.

Sources indicate that the demand originating from Chinese companies is so robust that Nvidia may indeed need to add new production capacity. Among the prominent buyers reportedly interested in the H200 chips are China's e-commerce giant Alibaba and ByteDance, the owner of popular platforms like Douyin (TikTok). Earlier reports had suggested that these companies had contacted Nvidia for substantial orders of H200 chips.

This surge in interest follows a decision on December 9 by former US President Donald Trump, who permitted Nvidia CEO Jensen Huang's company to export its second-fastest AI chips, the H200 processors, to China, albeit for a 25% sales fee.

Despite the US authorization, uncertainties persist regarding the final approval from the Chinese government. The report highlights that Chinese officials have yet to greenlight any purchases of the H200. Sources close to the agency revealed that Chinese officials held emergency meetings on December 10 to deliberate on this issue. Historically, only a limited number of H200 chips had been in production as Nvidia had focused its resources on its Blackwell and Rubin chips. Chinese clients have reportedly reached out to Nvidia expressing concerns about the supply, and the company has provided clients with guidance on current supply levels without specifying exact numbers.

The H200 AI chip, launched into mass deployment in 2024, is manufactured by TSMC utilizing its 4nm manufacturing process technology. Its strong demand is largely attributed to its superior performance, being approximately six times more powerful than its predecessor, the H20, which was specifically tailored for the Chinese market in 2023.

Nori Chiou, an investment director at White Oak Capital Partners, commented on the situation, stating that the H200's compute performance is estimated to be two to three times that of the most advanced domestically produced accelerators. Chiou noted, “I'm already observing many CSPs (Cloud Service Providers) and enterprise customers aggressively placing large orders and lobbying the government to relax restrictions on a conditional basis.” He further emphasized that the overall AI demand in China significantly exceeds the capacity of local production.