Elon Musk Warns: The U.S. Chip Industry Is About to Face a Dark Period, While China May Avoid This Crisis Perfectly!
At the 2026 World Economic Forum in Davos, Switzerland, Elon Musk warned that the U.S. artificial intelligence chip industry is standing on the edge of a crisis of "excess capacity but no use"; while China may escape this storm unscathed due to a completely different energy infrastructure path.
Musk's core argument is surprisingly simple: AI chips are useless without electricity. He pointed out that global AI chip production is surging exponentially - especially NVIDIA's H100 and B100 series GPUs, which have remarkable performance and are in high demand. But the problem is that once these chips are deployed in data centers, each one consumes power like a cow. A 10,000-watt AI training cluster can easily exceed 100 megawatts, equivalent to the power consumption of a medium-sized city.
America is precisely stuck at this "last mile": aging power grids, lengthy approvals, and insufficient transmission and distribution capabilities. States such as California, Texas, and Arizona, where tech companies are concentrated, have frequently experienced power supply shortages in recent years. An insider from the energy industry revealed that NVIDIA's two super-large AI data centers planned in Santa Clara might be delayed for several years due to the local grid's inability to provide stable high-load power. This is not a technical issue, but a hard constraint of infrastructure.
More troubling is that the U.S. power grid system is highly fragmented, operated by hundreds of private and local utility companies, and upgrading and expanding it requires layer-by-layer approval, taking five to ten years. At the same time, residential electricity prices are rising - because grid companies pass on the upgrade costs to users. This creates a vicious cycle: the hotter AI becomes, the stronger the electricity demand, the more the grid lags behind, the higher the electricity bills, ultimately slowing down the entire AI deployment process.
In contrast, as of the end of 2025, China's solar power generation capacity has reached 111.8 million kilowatts, nearly 4.7 times that of the United States' 23.8 million kilowatts. More importantly, China is building large-scale "wind and solar bases" in the western regions - centralized photovoltaic and wind farms, with配套 ultra-high voltage transmission lines directly delivering green electricity to eastern data center clusters.
According to the International Energy Agency, by 2030, global data center electricity consumption will account for more than 4% of total power demand, with AI loads accounting for over 60%. Whoever can ensure power supply while controlling costs will hold the ticket to the AI era. So far, the U.S. still leads in chip design and manufacturing, but in a more practical issue - whether the chips can actually run - it is being dragged into a "dark period" by its own infrastructure shortcomings. Meanwhile, China, with its large-scale deployment of renewable energy and the advancement of a national computing network, may not only keep up in this AI competition but also accelerate by taking advantage of the situation.
Dao Ge believes that Musk's warning is less a warning and more a calm analysis of the reality gap.
Original article: toutiao.com/article/1855175245222921/
Statement: This article represents the views of the author.