2 1 Shi Ji Jing Ji Bao Dao·2025-12-09 04:07Group 1 - The core viewpoint of the article highlights that energy is becoming the "new oil" in the AI era, with concerns about the speed of AI infrastructure development in the U.S. compared to China [2] - NVIDIA's CEO Jensen Huang stated that the U.S. may be outpaced by China in the AI race due to slower infrastructure development, emphasizing that China has double the energy resources of the U.S. [2] - A report from Morgan Stanley warns that by 2028, AI data centers could lead to a potential 20% electricity shortfall in the U.S. due to their high energy consumption [2] Group 2 - John Hopcroft, a Turing Award winner, pointed out that the rapid growth in AI's electricity demand is outpacing the construction of renewable energy facilities like wind power [3] - He provided an example of the Lawrence Livermore National Laboratory's supercomputer, which cost $600 million to build, but its operational electricity costs are significantly higher [3] - Hopcroft also noted that the current high energy consumption model for AI is unsustainable and suggested that AI could find more energy-efficient ways to operate, similar to the human brain [3]