在全球芯片短缺愈演愈烈之际，韩国三星电子公司的劳资纠纷可能给这场"芯片荒"再添一把火。 据三星最大的工会组织"全国三星电子工会（NSEU）"威胁称，该公司正就三星史上最大规模的罢工计划进行投票——一旦本周三投票通 过，将在5月中断芯片生产，这可能给三星公司带来数百亿美元的损失。 三星韩国员工总数为12.5万人，其中约9万名已加入NSEU工会，并且都将参与此次投票。 据工会负责人Choi Seung-ho表示，如果工人与集团未能达成协议，他们计划从5月2日开始进行18天的罢工， 这可能会对三星位于首尔南 部平泽的半导体工厂的约一半产量造成影响。 三星作为全球最大的存储芯片制造商，如果其发生罢工事件，可能会对三星半导体业务造成重大冲击，并加剧全球半导体供应的瓶颈问题， 从而抑制从汽车、电脑到智能手机等行业的半导体供应。 三星史上最大规模罢工迫在眉睫 本月早些时间，全国三星电子工会（NSEU）正式启动为其十天的罢工投票程序，投票结果将于本周三出炉。 这场投票标志着三星工会与三星集团长达八个月的薪酬谈判最终破裂，并为可能发生的三星历史上最大规模罢工铺平道路。 "我预计会存在生产中断的情况，"Choi Seung-ho ...

Group 1 - Nvidia achieved a significant lobbying victory as the proposed GAIN AI Act, which aimed to restrict the sale of advanced AI chips to "concerned countries," was not included in the upcoming defense bill [3][9] - The GAIN AI Act intended to prioritize U.S. customers for high-end AI chips before allowing sales to countries like China, but Nvidia argued that such restrictions would hinder global competition and were unnecessary [5][9] - The decision not to include the GAIN AI Act in the defense bill was seen as a wise choice by Nvidia's CEO Jensen Huang, who highlighted the potential greater harm of the act compared to previous legislation [9] Group 2 - The GAIN AI Act, officially known as the "Guaranteeing Access and Innovation for National Artificial Intelligence Act of 2025," was proposed to address the perceived imbalance between U.S. demand for high-end AI chips and supply capabilities [13] - The act included a certification requirement for companies wishing to sell advanced AI chips to "concerned regions," mandating proof of priority purchasing rights for U.S. buyers and no backlog of orders [14] - The Brookings Institution expressed concerns that the GAIN AI Act could undermine U.S. AI leadership due to its vague language and potential compliance issues [15][16] Group 3 - Despite the temporary setback for the GAIN AI Act, the U.S. chip industry remains on alert as hawkish lawmakers are reportedly working on a new proposal, the SAFE Act, to solidify existing chip export bans into law [11] - The focus on chip restrictions may overlook critical factors affecting AI development, such as power supply limitations, which have been identified as a growing bottleneck in the U.S. [19] - Goldman Sachs analysts predict that by 2030, the electricity demand from data centers could nearly double, potentially straining the U.S. power grid [22]