On September 14, Chen Liwu, the CEO of Intel, shared insights at an industry summit, noting that the burgeoning growth of AI agents has led to a significant uptick in CPU demand. Presently, the supply can only cater to approximately 50% of the needs of Intel's leading customers. Several heads of major tech firms have reached out, urging for an increase in supply, a situation for which Chen Liwu expressed his regrets. The scarcity of CPUs is largely attributed to the soaring demand in AI inference applications, where CPUs hold distinct advantages over GPUs, particularly in areas such as agent orchestration, making them hard to substitute. Regarding process technology, Intel's 18A node has already commenced mass production. Furthermore, the company has plans to initiate production of the 14A node in the first quarter of the upcoming year. The innovative architecture is anticipated to slash inference power consumption to a mere 1/15th of that required by GPUs.