China’s AI Race: How Huawei Chips and Cheap Energy Are Powering the Future

China’s artificial intelligence (AI) industry continues to make big strides, even without access to the world’s most advanced American chips. While Chinese-made semiconductors still lag behind Nvidia’s powerful GPUs, local companies are finding new ways to stay competitive. The secret lies in China’s access to cheap energy and the powerful chip clusters built by its tech champion, Huawei.

Experts say this strategy is part of China’s broader goal to become self-sufficient in AI technology. “China is striving for self-sufficiency across the AI stack, as it sees AI as a strategic technology for national and economic security,” said Wendy Chang, a senior analyst at the Mercator Institute for China Studies (MERICS).

China’s AI Race: How Huawei Chips and Cheap Energy Are Powering the Future

The U.S. has placed strict restrictions on exporting advanced technology to China. Nvidia, known for its world-leading AI chips, is no longer allowed to sell its most advanced GPUs to Chinese companies. To fill this gap, Nvidia introduced a downgraded version called the H20 chip, designed specifically for China. But Beijing has encouraged local firms to avoid Nvidia products and instead support domestic alternatives.

This is where Huawei steps in. The company produces the Ascend series of chips, which power many of China’s AI projects. While a single Huawei chip cannot match the performance of Nvidia’s best GPUs, Huawei’s real strength lies in connecting hundreds of these chips together to form massive AI clusters.

One of Huawei’s latest innovations is the CloudMatrix 384 system, which links 384 Ascend 910C chips. This setup can deliver performance levels that rival Nvidia’s top-end GB200 NVL72 system, which uses only 72 GPUs. By linking more chips together, Huawei can overcome performance gaps, though at the cost of higher power usage.

That’s where China’s energy advantage makes a difference. According to MERICS’ Chang, “Solutions like the CloudMatrix are less power efficient than Nvidia systems, but here China benefits from its abundance of cheap energy.” The country has made huge investments in renewable energy, such as solar, wind, and nuclear power. This gives it a steady and low-cost supply of electricity for running energy-hungry AI data centers.

Local governments are also playing a role. Cities like Shanghai and Shenzhen are offering subsidies and “vouchers” to companies that use domestic chips, helping them reduce computing and electricity costs. Some regions even cut power bills for data centers running on Chinese-made hardware.

Still, questions remain about how sustainable this approach is. Huawei’s chips are made by Semiconductor Manufacturing International Company (SMIC), China’s largest chipmaker. But SMIC lags several generations behind Taiwan’s TSMC, which manufactures chips for Nvidia and other global tech giants. Because of export bans, SMIC cannot buy key chipmaking equipment, including advanced lithography machines from Dutch firm ASML.

See also: Digital Silk Road: Pakistan and China to Invest in 5G, Robotics, and Biotech

Huawei’s chips are built using a 7-nanometer process — older and less efficient than TSMC’s cutting-edge technology. This makes production costly and complex. Yet, China’s focus on scale and energy efficiency has allowed it to continue growing in the AI sector despite these hurdles.

Experts say China’s long-term challenge is whether it can keep pace as AI chips continue to evolve. “China still lags significantly due to export controls on semiconductor manufacturing equipment,” said Hanna Dohmen from Georgetown’s Center for Security and Emerging Technologies.

For now, China’s combination of homegrown innovation, cheap power, and government support is helping it compete in the global AI race, proving that it doesn’t always take the most advanced chips to stay in the game.