China’s data centres are accelerating the adoption of domestically produced chips as Beijing pushes to reduce reliance on foreign technology. The move comes in response to tightening US export restrictions on semiconductors, underscoring China’s determination to achieve technological self-sufficiency while supporting the rapid expansion of its artificial intelligence (AI) industry.

Beijing’s Push for Tech Autonomy

According to the South China Morning Post, state-owned computing hubs across the country have been instructed to ensure that over half of their chips are sourced from local manufacturers.

This directive aligns with China’s broader semiconductor strategy, driven by the Ministry of Industry and Information Technology (MIIT) and the National Development and Reform Commission (NDRC).

The new mandate builds on rules introduced in Shanghai in March 2023, which required intelligent computing centres to use at least 50% locally made chips by 2025. Now, these guidelines have been elevated into a nationwide requirement.

Impact of US Export Restrictions

The shift is a direct response to Washington’s tightening export bans on advanced semiconductors, particularly Nvidia’s H100 and H800 GPUs. These chips remain critical for training large-scale AI models, but are increasingly out of reach for Chinese companies due to restrictions.

Although Chinese-made chips still trail behind in terms of performance, they are being deployed more widely in AI inferencing tasks, while Nvidia continues to dominate high-end AI training.

Growth of Intelligent Computing Centres

China is rapidly expanding its intelligent computing infrastructure. According to KZ Consulting, more than 500 new computing centre projects were launched across regions including Inner Mongolia and Guangdong in 2023 and 2024.

These centres are being built to meet the surging demand for AI-driven computing power, particularly as the country ramps up its ambitions in AI growth by 2025.

Adoption of Huawei AI Chips

Several Chinese companies are already embracing local alternatives. iFlytek, a leading AI firm, has started adopting Huawei’s AI chips for training models, signaling a growing confidence in domestic hardware.

However, experts caution that China’s semiconductor industry still faces challenges in fully replacing US technologies for high-performance data centre operations.

The Road Ahead: Self-Sufficiency vs. Global Dependence

While the adoption of domestic chips in Chinese data centres is a major step toward reducing dependence on foreign suppliers, the journey remains complex. For now, local chips are expected to complement—rather than completely replace—foreign hardware in the short term.

The push reflects Beijing’s long-term vision of technological sovereignty, ensuring that China can withstand external pressures while fueling innovation in its AI ecosystem.

Key Takeaways

• China’s data centres instructed to use over 50% domestic chips.
• Policy overseen by MIIT and NDRC to strengthen the local semiconductor industry.
• Triggered by US export restrictions on Nvidia GPUs.
• Over 500 new intelligent computing centres launched across China.
• Companies like iFlytek already deploying Huawei AI chips.
• Full replacement of US semiconductor technology remains uncertain.

For any quarries feels free to contact us