- China plans massive AI computing network backed by domestic chips
- High-bandwidth memory shortages limit production of advanced AI accelerators in China
- Domestic chipmakers still lag behind global leaders by several years
China is developing a plan that could spend about 2 trillion yuan (about $295 billion) on a nationwide AI computer network.
The proposal would connect data centers across the country into a unified computing grid operated largely by state-backed telecommunications companies.
Authorities would like at least 80% of the underlying technology, including AI chips and associated infrastructure, to come from Chinese suppliers.
Massive development centered on local technology
The project is being developed by the National Development and Reform Commission, while major operators including China Mobile and China Telecom are said to oversee the operation.
The network is reportedly linked to a single national IT platform by 2028 through extensive infrastructure deployment.
Financing would rely largely on very long-term sovereign debt and government bonds, while associated improvements to the electricity grid could significantly increase costs.
The total capital requirement could exceed 5 trillion yuan (about $738 billion) when energy infrastructure is included in broader deployment-related estimates.
The plan comes as Beijing continues to tighten restrictions on foreign semiconductor products used in data centers and AI facilities.
By 2025, authorities required data centers to obtain at least 50% of their chips from domestic manufacturers. In November of that year, state-funded projects reportedly faced additional restrictions that excluded foreign accelerators from facilities still under construction.
Officials also imposed compliance measures requiring the removal of Nvidia, AMD and Intel components in projects with less than 30% completion.
These measures have increased opportunities for Chinese chipmakers, including Huawei, while reducing dependence on suppliers such as Nvidia, AMD and Intel.
The policy ensures that critical AI and LLM tools run on hardware developed in China, but replacing imported processors remains a difficult task.
Domestic chip supply remains a major challenge
China’s semiconductor manufacturing capacity comes mainly from SMIC and a small group of state-licensed foundries.
Zhao Haijun, co-general director of SMIC, warned that excessive infrastructure expansion could leave facilities underutilized.
Reports indicate that SMIC’s most advanced stable manufacturing process remains roughly comparable to 7nm technology and is already operating above 93% utilization levels.
With many domestic chip designers competing for the same production resources, it could prove difficult to quickly scale up production given current wafer allocation limits.
High-bandwidth memory also remains a major constraint, limiting the number of advanced accelerators that can be assembled for AI workloads and tool deployment.
Industry estimates suggest that domestic suppliers could only meet about 76% of China’s demand for AI chips by 2030, even as demand grows to a $67 billion market.
Huawei has increased shipments, including about 812,000 chips last year, but supply chain limitations continue to affect production scale.
Chinese industry executives have acknowledged that domestic AI data center chips remain 5 to 10 years behind major international competitors in some categories.
Reports also indicate that DeepSeek has returned to Nvidia hardware for some training tasks after experimenting with Huawei alternatives in heavy workloads.
This suggests that Chinese processors may still struggle with the most demanding AI training environments, despite advances in inference performance.
Via Tom’s material
Follow TechRadar on Google News And add us as your favorite source to get our news, reviews and expert opinions in your feeds.
