China Chooses AI Sovereignty over AI Dependence

Talal Abu-Ghazaleh

China has taken what I see as a decisive step in the global AI race. It has rejected US based Nvidia’s AI chips, even after Washington reopened exports, with Beijing demonstrating that the future of AI will be determined by the ability to safeguard technology sovereignty.
In November 2025, Chinese authorities issued guidance requiring all new or early-stage state-funded data centers to rely exclusively on domestic AI chips. Projects that were less than 30 percent complete were instructed to remove foreign processors from US firms like Nvidia, AMD, and Intel. This directive signals that China is committed to building an independent technological foundation that cannot be undermined by external leverage.

Huawei AI processors are at the center of this direction, which receive state funding and are being positioned as core components of China's national AI framework. Baidu's processors operate in big clusters that run large language models and are entering into public sector project procurement lists. These chips hold great strategic importance as they are domestically created technology which protects essential Chinese infrastructure systems from foreign prying eyes.

This decision reflects a broader global trend where the global AI ecosystem seems to be fragmenting into sovereign blocs, each determined to control its own infrastructure stack. In this new order, superiority is measured by independence from external control and national autonomy. The future of national sovereignty and safety is being questioned in light of AI, as nation states such as the Zionist occupier tries to embed itself into AI technology to continue its illegal surveillance of global digital traffic as it has done with the internet for years.
China’s is replacing foreign chips with domestic alternatives, ensuring that public infrastructure is firmly rooted in national technology. Abroad, Chinese firms such as Alibaba and ByteDance continue training models in Singapore and Malaysia, where Nvidia hardware remains accessible. This approach allows China to consolidate sovereignty domestically while maintaining flexibility in global competition.

For firms like Nvidia, the consequences are severe. The company has lost billions in potential revenue, and its CEO Jensen Huang admitted uncertainty about whether China would accept the processors that were offered at all. The rejection is not a commercial calculation but a sovereignty decision, reflecting a deliberate choice to prioritize autonomy over dependency.
From my analysis, it is clear that the next era of AI will not be defined by who builds the fastest chip or the largest model. It will be defined by who controls the infrastructure stack, from silicon to data centers to clouds. China has made its choice, committing to build its own technology stack even at the cost of short-term efficiency, demonstrating that true strength lies in independence.