China has added homegrown AI chips to its ‘secure and reliable’ government procurement list for the first time, formally cementing a strategic pivot away from Nvidia and other Western suppliers. The move represents a watershed moment in Beijing’s effort to build a self-reliant AI chip ecosystem, one that sidesteps U.S. export controls and positions state-backed chipmakers as the default choice for public-sector purchasing.
Key Takeaways
- Nine domestic AI chips were certified for China’s ‘secure and reliable’ procurement system in a historic first.
- The certifications create a new category specifically for ‘AI training and inference chips’.
- Certifications are valid for three years and apply to government agencies, state-owned enterprises, and state-linked customers.
- The list reportedly includes suppliers such as Huawei and Cambricon.
- The move could generate billions of dollars in additional revenue for domestic chipmakers.
How China’s Domestic AI Chips Procurement List Works
The domestic AI chips procurement list operates within China’s broader Xinchuang initiative, a state-backed framework designed to replace foreign technology with homegrown alternatives across critical sectors. The certifications were issued jointly by the China Information Technology Security Evaluation Centre and the National Secrecy Science and Technology Evaluation Centre, lending official government weight to the selections. These certifications carry three-year validity periods, meaning suppliers must undergo recertification to maintain their approved status.
The procurement guidance has been distributed to ministries, state-owned companies, and public institutions across China. Crucially, this marks the first time these entities received formal written instructions to prioritize domestic suppliers in their AI chip purchasing decisions. The list itself is typically not publicly released, making the exact composition of all nine approved chips difficult to verify independently. However, reporting suggests that Chinese suppliers including Huawei and Cambricon secured positions on the roster.
The framework applies specifically to government agencies, central state-owned enterprises, and other state-linked customers, creating a captive market for domestic chipmakers. This is not a voluntary guideline—it is procurement policy with teeth, affecting billions of dollars in annual technology spending across China’s public sector.
Domestic AI Chips vs. Nvidia: A Structural Shift in Procurement
The inclusion of domestic AI chips on the procurement list directly challenges Nvidia’s dominance in China’s AI accelerator market. Where Nvidia has traditionally supplied H100 and other high-end GPUs to Chinese data centers and AI research institutions, the new procurement framework creates institutional pressure to source domestically instead. This is not about technical parity—it is about state control and reducing exposure to U.S. export restrictions.
The timing reflects Beijing’s frustration with American semiconductor controls. U.S. export curbs have already blocked the sale of Nvidia’s most advanced chips to China, forcing Chinese AI developers and researchers to find alternatives. By formalizing domestic chip procurement, China is converting a constraint into a strategic advantage: it locks government spending into a domestic supply chain that cannot be interrupted by Washington.
Competitors like AMD and Intel face similar headwinds in China’s public sector, though the procurement list specifically targets AI accelerators rather than general-purpose CPUs. The Xinchuang initiative previously addressed x86-replacement processors and Windows alternatives, establishing a proven model for technology substitution. Domestic AI chips represent the next frontier of that substitution strategy.
Why This Matters for China’s AI Ambitions
The domestic AI chips procurement policy signals that Beijing is willing to trade short-term performance gains for long-term strategic autonomy. Chinese chipmakers like Huawei and Cambricon have been developing AI accelerators for years, but they have struggled to compete with Nvidia on raw performance and software ecosystem maturity. By guaranteeing government procurement, China eliminates the chicken-and-egg problem: suppliers now have a massive, stable customer base that can fund R&D and manufacturing scale-up.
The move could generate billions of dollars in additional revenue for local chipmakers, according to reporting on the procurement framework. That capital can be reinvested in chip design, manufacturing partnerships, and software optimization. Over time, this creates a virtuous cycle where domestic chips improve through volume and investment, reducing China’s dependence on foreign suppliers even further.
The geopolitical dimension is equally important. By standardizing domestic AI chips across government and state-owned enterprises, China reduces the leverage that U.S. export controls can exert. If Nvidia chips are banned, the government’s AI infrastructure does not collapse—it simply runs on domestic alternatives that were pre-approved and already integrated into mission-critical systems.
What Happens to Foreign Suppliers?
The procurement list does not explicitly ban Nvidia, AMD, or Intel chips from China. Rather, it creates a strong institutional preference for domestic options through the Xinchuang framework. Government agencies and state-owned enterprises will face pressure—both formal and informal—to source from approved domestic suppliers when feasible. Private companies and academic institutions not bound by procurement rules may continue purchasing foreign chips, but the largest, most resource-rich customers in China will increasingly turn domestic.
This tiered approach is strategically clever. It avoids overt violation of trade agreements while systematically shifting a massive chunk of China’s AI chip market away from foreign suppliers. Nvidia and others can still sell into China’s private sector, but they have lost the state-backed anchor tenant that typically drives adoption and sets industry standards.
FAQ
What chips are on China’s domestic AI chips procurement list?
The exact composition of the nine certified chips is not publicly available, as China’s procurement list is typically not released. However, reporting indicates that suppliers including Huawei and Cambricon secured positions on the roster. The specific models and performance characteristics of each approved chip remain largely opaque to outside observers.
How long are the certifications valid?
The certifications issued for domestic AI chips are valid for three years. After that period, suppliers must undergo recertification to maintain their approved status on the procurement list.
Does this ban Nvidia from selling chips to China?
The procurement list does not explicitly ban Nvidia or other foreign suppliers. Instead, it creates a strong institutional preference for domestic alternatives within government and state-owned enterprises. Private companies and academic institutions may still purchase foreign chips, but the largest state-backed customers now face pressure to source domestically.
China’s addition of domestic AI chips to its official procurement framework marks a decisive shift in how the world’s second-largest economy sources critical technology. By formalizing government preference for homegrown accelerators, Beijing is building a parallel AI ecosystem that can operate independently of U.S. export controls. Whether domestic chips can match Nvidia’s performance is almost beside the point—the real goal is strategic autonomy, and that goal just moved significantly closer.
Edited by the All Things Geek team.
Source: Tom's Hardware
