China's key weapons in its AI battle with the U.S. — massive Huawei chip clusters and cheap energy

China is making significant strides in AI development despite U.S. restrictions on advanced chips, primarily by leveraging Huawei's Ascend chip clusters and its abundant, cheap energy resources. Huawei compensates for its chips' individual performance gap against Nvidia by interconnecting many Ascend chips into large, high-performance clusters, such as the CloudMatrix 384, which can rival Nvidia's top systems. This strategy, while less power-efficient, is supported by China's cheap energy, including substantial investments in green and nuclear power, and local government subsidies for data centers using domestic chips. However, the long-term sustainability of this approach is challenged by China's domestic chip manufacturer, SMIC, which faces limitations in advancing its technology due to U.S. export controls on critical manufacturing equipment, potentially hindering its ability to keep pace with global leaders like Nvidia and TSMC.

Analysis

China is strategically advancing its AI capabilities despite U.S. export restrictions on advanced semiconductors, primarily by leveraging Huawei's Ascend chip clusters and its abundant, cheap energy resources. Huawei compensates for its individual chips' performance gap against Nvidia by interconnecting numerous Ascend 910C chips into large, high-performance clusters, such as the CloudMatrix 384, which can rival Nvidia's GB200 NVL72 system, albeit using five times more chips. This less power-efficient approach is underpinned by China's significant investments in green and nuclear energy, coupled with local government subsidies that reduce electricity costs for data centers utilizing domestic chips.

However, the long-term sustainability of this strategy faces considerable challenges due to limitations in China's domestic chip manufacturing capabilities. SMIC, Huawei's primary manufacturer, lags generations behind TSMC and is restricted from acquiring critical advanced lithography tools, such as ASML's EUV machines, due to U.S. export controls. While SMIC can produce Huawei's 7-nanometer Ascend chips using older tools, this process is likely costly and inefficient, potentially hindering China's ability to keep pace with the rapid advancements in AI semiconductors from global leaders like Nvidia and TSMC. The overall market sentiment is mixed, reflecting both China's innovative adaptation and persistent technological constraints.

Key Decisions for Investors

• Investors should closely monitor the development of China's domestic semiconductor manufacturing capabilities, particularly SMIC's progress and any shifts in export control policies, as this will be critical for the long-term viability of China's AI self-sufficiency strategy.
• Evaluate the financial implications of China's energy advantage and government subsidies on the operational costs and scalability of its AI infrastructure, considering the higher power consumption of Huawei's chip clusters.
• Assess potential supply chain risks and opportunities for global semiconductor firms like TSM and ASML due to ongoing geopolitical tensions, while observing how companies such as Nvidia adapt their product strategies for restricted markets like China.