NextFin

China First Adds Domestic AI Chips to Procurement List as Beijing Weighs Nvidia H200 Purchases Following Trump's Approval

China has reportedly added domestic artificial intelligence (AI) chips to an official government procurement list for the first time, a move that comes as Beijing grapples with whether to allow companies to purchase Nvidia Corp.'s advanced H200 processors following U.S. President Donald Trump's decision to permit their export.

AI Generated Image

AI Generated Image

The Ministry of Industry and Information Technology (MIIT), which oversees the tech industry, recently included AI chips from Chinese manufacturers including Huawei and Cambricon on its government-approved supplier list, according to the Financial Times, citing people familiar with the matter on Wednesday. The step was designed to enhance use of domestic semiconductors in China's public sector and could generate billions of dollars in new sales for local chipmakers.

The development follows a series of conflicting signals on China's stance toward Nvidia's H200 chips. On Tuesday, the Financial Times reported that Beijing was set to limit access to the processors despite Trump's export approval. On Wednesday, The Information reported that Chinese regulators convened emergency meetings with representatives from companies including Alibaba Group, ByteDance and Tencent Holdings to assess their demand for the H200, with officials promising to inform them of the government's decision soon.

The rapid succession of moves highlights Beijing's dilemma between supporting AI development that requires powerful chips China cannot yet produce domestically and pushing adoption of homegrown alternatives to achieve technological self-sufficiency.

Domestic Chips Gain Official Status

The inclusion of Chinese AI processors on the procurement list marks a significant milestone in Beijing's campaign to reduce reliance on American technology. The Information Technology Innovation List, known as Xinchuang in Chinese, serves as guidelines for government agencies, public institutions and state-owned companies that spend billions annually on IT products.

While these entities have previously been encouraged to support local chipmakers, this represents the first time public-sector groups have received written instructions to purchase domestic AI chips, according to sources cited by the Financial Times. The list has not been made public, but several government agencies and state-owned companies have already received the guidance document.

The move forms part of Beijing's broader strategy to phase out foreign technology products following Washington's export controls. In recent years, domestic microprocessors to replace those made by AMD and Intel, as well as operating systems to substitute Microsoft's Windows, have been added to the list.

China has also increased subsidies that cut energy bills by up to half for some of the country's largest data centers, helping tech giants such as Alibaba and Tencent offset higher electricity costs associated with using less efficient domestic semiconductors.

Nvidia Export Approval Sparks Urgent Government Consultations

Trump announced Monday in a Truth Social post that he had told Chinese President Xi Jinping the U.S. would allow Nvidia to ship H200 products to "approved customers in China" under conditions allowing for "continued strong National Security." Trump added that 25% would be paid to the United States, though the mechanism remains unclear.

Within hours of the announcement, Chinese officials moved swiftly to assess the implications. The National Development and Reform Commission and the Ministry of Industry and Information Technology, which oversee local tech companies' AI chip procurement, sent letters to major tech firms requesting information on their H200 requirements for AI model development, according to The Information, citing people involved in the correspondence.

On Wednesday, these agencies convened emergency meetings with company representatives to gather demand assessments, with officials promising a government decision would be communicated soon after compiling responses, the report said.

According to Reuters, citing four sources briefed on the matter, ByteDance and Alibaba have asked Nvidia about purchasing the H200 and are keen to place large orders should Beijing approve, though they remain concerned about supply. The H200 is almost six times as powerful as the H20, which was previously the most advanced chip legally exportable to China.

Beijing Faces Complex Calculus on Chip Approval

Chinese authorities are weighing multiple factors as they consider whether to permit H200 purchases. According to the Financial Times, regulators have been discussing ways to permit limited access, with buyers probably required to go through an approval process, submitting purchase requests and explaining why domestic providers cannot meet their needs. No final decision has been made.

The bottom line for allowing H200 purchases is that it would not derail China's long-term goal of technological self-sufficiency, according to three people familiar with the Chinese government's thinking cited by The Information. Officials view buying high-end American chips as a temporary measure to address the shortage of powerful training chips while Chinese firms catch up.

One option under consideration is imposing caps on Nvidia chip purchases relative to the amount of domestic processors companies already have and plan to buy, though whether the ratio would be measured by dollar value or chip count remains unclear, according to two people close to the Chinese government cited by The Information.

The government is also likely to impose restrictions in strategically sensitive industries such as finance and energy, according to one person cited in the report. Any policy would be communicated through "window guidance," an informal method regulators use to convey policy expectations privately rather than through public rule-making.

Technical Gaps Fuel Demand Despite Self-Sufficiency Push

China can currently produce sufficient chips for inference tasks, when AI models generate answers or perform tasks, but still lacks the know-how and equipment to manufacture more powerful processors for training, in which models ingest vast amounts of data. This technical gap makes Nvidia's offerings attractive.

"The training of leading Chinese AI models still relies on Nvidia cards," Zhang Yuchun, a general manager at Chinese cloud service provider SuperCloud's solution and ecology units, told Reuters. "I expect the leading Chinese tech companies to buy a lot although in a low-key manner."

However, the transition to domestic alternatives faces practical challenges. An executive from a state-owned financial institute told the Financial Times that while his organization had allocated Rmb100 million to buy domestic AI chips this year, most purchased Chinese processors now sit idle. The company's quantitative trading models were built on Nvidia hardware, and switching to Huawei processors would require significant adaptation work, including rewriting code in an unfamiliar language.

A Chinese policymaker acknowledged such reluctance was common during transition phases but stressed the need for greater technological independence. "The growing pains are unavoidable," they told the Financial Times. "But we have to get there."

Very limited quantities of H200 are currently in production, according to two people familiar with Nvidia's supply chain cited by Reuters, as the company has focused instead on its most advanced Blackwell and upcoming Rubin lines.

Explore more exclusive insights at nextfin.ai.