According to an OilPrice.com analysis, China’s artificial intelligence industry may hold a structural edge over U.S. rivals, largely because of lower electricity costs and a more integrated power system. The article argues that this gap could increase pressure on the stock-market valuations of American AI companies.
The piece cites Mehrdad Emadi, head of risk analysis and energy derivatives markets consultancy Betamatrix, who told OilPrice.com that major Chinese AI players currently achieve around 90% of the performance of their U.S. competitors at roughly 10% of the cost. In his view, as much as half of U.S. AI firms’ expenses is tied to powering data centers, and that burden is likely to rise.
The article also highlights structural differences between the U.S. and Chinese grids. Steve Keen, honorary professor at University College London, told OilPrice.com that China benefits from a centralized power network built around a unified engineering approach, while the U.S. grid is fragmented and less effective at moving large volumes of electricity over long distances.
OilPrice.com says China’s grid operates under a national strategy controlled by State Grid Corporation, allowing electricity to move across thousands of kilometers with relatively low losses. By contrast, the U.S. system is split into three isolated interconnections — Eastern, Western, and Texas/ERCOT — which, the article says, makes power transfers more difficult and more expensive.
The publication also points to technical differences. It says China relies on 800-1,100 kilovolt Ultra-High Voltage direct-current lines that can move up to 12 gigawatts through a single transmission corridor over roughly 3,000 kilometers with almost no losses. In the U.S., according to the article, the backbone remains 345-500 kilovolt High-Voltage Alternating Current trunk lines, which do not match that scale or efficiency.
The central conclusion is that the U.S.’s aging and fragmented grid could become a physical constraint on how quickly American technology companies expand their AI capacity, while China faces fewer such limitations.
OilPrice.com also examines whether new generation capacity can solve the problem. It says solar and wind still do not provide the always-on supply that AI data centers require. Nuclear energy, including small modular reactors, has drawn interest from hyperscalers such as Microsoft, Amazon, and Google, but the article notes long lead times and the risk of multi-billion-dollar additional costs.
The author says traditional U.S. nuclear projects have often faced cost overruns running into billions of dollars and development timelines of up to 15 years, while SMR projects may still require construction windows of as much as seven years. The article also notes the need for regulatory approvals, including from the U.S. Nuclear Regulatory Commission.
The piece is presented as an interpretive analysis rather than a confirmed forecast. Still, its main argument is that electricity prices, grid infrastructure and capital intensity could become some of the most important constraints on the U.S. AI boom.
