China’s campaign to expand its power grid dates back to the 1970s. Concerned that electricity shortages could hinder national development, Chinese leaders directed state-owned enterprises to build hundreds of coal-fired power plants.
Later, the country placed a major bet on renewable energy, channeling massive investment into giant hydropower projects, vast solar farms, and large-scale wind power installations.
Because favorable locations for power generation are often far from population centers in eastern China, the country has also built the world’s largest ultra-high-voltage (UHV) transmission network, investing more than USD 50 billion since 2021, according to domestic media.
Today, China has a total power generation capacity of around 3.75 terawatts—more than double that of the United States. According to the World Nuclear Association, Beijing is currently constructing 34 nuclear reactors, with nearly 200 more planned or proposed. In Tibet, China is building the world’s largest hydropower project, which could generate three times as much electricity as the Three Gorges Dam.
According to China’s National Energy Administration, data centers in the country can now purchase electricity at prices as low as around 3 cents per kilowatt-hour through long-term contracts. In contrast, operators in U.S. markets such as Northern Virginia typically pay between 7 and 9 cents per kilowatt-hour.
However, China’s surge in energy spending is accelerating its debt burden, which is increasingly becoming a drag on the economy. At State Grid Corporation of China, debt and other liabilities rose by more than 40% from early 2019 to the end of 2024, reaching approximately USD 450 billion.
The city of Ulanqab and the neighboring county of Horinger in Inner Mongolia, about 320 kilometers northwest of Beijing, have been designated as one of eight hubs under the government’s “Eastern Data, Western Computing” initiative. Many of these areas were chosen for their access to low-cost electricity. The program also aims to attract investment to China’s poorer inland regions.
Chinese leaders have required companies to build new data centers only in designated areas, while fast-tracking approvals and land clearance for these projects. Some data centers pay only half of their electricity bills, with the remainder covered by government subsidies.
In Ulanqab, one side of the road is lined with nearly abandoned buildings, with only a few small eateries and tobacco shops remaining. On the other side stands a data center operated by Centrin Data, which has also built wind turbines and solar panels in the area. This data center, which became operational just 16 months after construction began, provides cloud computing services to customers in Beijing.
Apple, Alibaba, and Huawei also operate data centers there, while companies such as electric vehicle maker XPeng use the area to train AI models and process AI workloads.
In Horinger, about a two-and-a-half-hour drive from Ulanqab, state-owned power group China Huadian began developing power infrastructure last year for a data center cluster. This year, the USD 230 million project started supplying electricity to data centers operated by China’s largest telecom companies and cloud computing startup Paratera.
To compensate for domestic chips that are less advanced than U.S. chips, Huawei, Alibaba, Baidu, and other Chinese companies are seeking to boost computing power by linking systems that combine hundreds or even thousands of Chinese-made chips.
According to insiders at Chinese technology firms, the biggest obstacle to this strategy is China’s inability to manufacture its most advanced chips quickly enough.
Meanwhile, U.S. export controls also limit access to advanced chipmaking equipment. Chinese companies have sought ways around these restrictions, including purchasing U.S. chips through informal channels and using chips located at data centers outside China.
Analysts predict that shortages of high-end chips will persist for at least several more years. It remains unclear how many Nvidia H200 chips China will be able to acquire, and how changes to export restrictions might alter the competitive landscape.
“In the short term, China’s lack of leading-edge chip manufacturing capacity is a bigger constraint than the U.S. power bottleneck,” said Qingyuan Lin, a semiconductor analyst at Bernstein. However, “the longer the AI race lasts, the more opportunities China will have to narrow the gap.”
(Source: vneconomy.vn)