我国氮化镓电源芯片实现新突破，AI算力中心年省3亿度电

Group 1: Industry Insights - The successful development of a new GaN-based power module by Jiufengshan Laboratory can save nearly 300 million kWh of electricity annually when deployed in a 1 GW AI computing center, translating to approximately 240 million yuan in electricity costs [1] - The construction of a 1 GW AI computing center is expected to consume 8.76 billion kWh of electricity annually, equivalent to the full-load output of a large nuclear power plant [1] - The demand for high-efficiency and high-density power devices is expected to surge due to AI computing, with server rack power exceeding 100 kW and power density requirements surpassing 100 W per cubic inch [2] Group 2: Company Developments - Xinyi Technology is leveraging its expertise in power electronics to enter the AI server power supply market, emphasizing its core advantages in SiC and GaN technologies, which are crucial for achieving high efficiency and power density [3] - Jingfang Technology's investment in Israeli company VisIC, a leader in GaN device design, is focused on developing silicon-based GaN technology for inverter solutions rated at 400V, 650V, and 800V [3]