I have a slightly different opinion. China isn't facing a nationwide power shortage, but it is experiencing a structural power shortage. What China lacks right now isn't the total annual electricity generation, but electricity during peak hours, electricity for load centers, stable and dispatchable power, and power capacity that can quickly connect to data centers and industrial projects. From a regional perspective, western China has power, but eastern China needs power. There's solar power during the day, but electricity is needed at night. It's not tight during normal times, but peak times are very tight. There's a lot of installed capacity for renewable energy, but there's insufficient stable and dispatchable power sources. Especially when it comes to AI data centers, which are different from regular industrial electricity needs. AI training and inference require continuous and stable power supply, high-power-density server rooms, low-latency networks, cooling systems, backup power, and substation capacity. Solar power during the day is helpful, but AI data centers don't only operate during the day. At night, on cloudy days, during extreme weather, summer peaks, and winter peaks, stable power sources are essential. Developing AI isn't just about having electricity; it's about having stable, scalable, sustainable, and data-center-grade power resources that are close to the demand for computing power and can quickly connect to the grid. This is why I say that future competition won't just be about computing power—it will be about power resources.