Following a simultaneous surge in the stock prices and performance metrics of AI-concept stocks, the market has started to express concerns over potential bubble risks. These encompass overvalued assessments, technological bottlenecks, and speculative capital maneuvers. Amidst these apprehensions, NVIDIA and OpenAI unveiled an industry-shaking investment initiative. The duo plans to collectively invest hundreds of billions of dollars in constructing a 10-gigawatt super AI data center, equipped with millions of GPUs to facilitate the training of next-generation large language models. This strategic move has further ignited market enthusiasm for the future prospects of AI computing capabilities.

Several public fund managers have highlighted that the global AI landscape has entered a novel phase, marking the official commencement of an AI infrastructure race defined by 'ultra-large scale, ultra-high energy consumption, and ultra-high investment.' From semiconductor chips to liquid cooling technology, and from computing clusters to energy infrastructure, the entire global AI computing power industry chain is experiencing a comprehensive activation. Concurrently, the competitive focus among nations is pivoting from 'model rivalry' to the more fundamental 'computing power clash.'