文章

AI算力格局重塑 · 辽宁舰远海演训

#834 · 2026-06-23 · 21ZHAO Blog
Reading Path / ARTICLE 先抓主张，再转成行动 #834 · 21ZHAO Blog · 读完进入产品或下一篇

AI算力格局重塑 · 辽宁舰远海演训

一、 权威必看

EN: The PLA Navy Liaoning carrier group has safely returned to port after completing its longest open-sea combat training in the vessel’s history. Spanning nearly two months from late April to mid-June, the exercise covered the South China Sea and Western Pacific, integrating air-sea system confrontation and multi-batch day-night aircraft attacks. This deployment highlights a strategic shift towards sustained power projection capabilities beyond traditional coastal defense zones. 中： 据央视新闻6月22日报道，海军辽宁舰编队圆满完成远海实战化训练并安全返回母港。此次训练自4月20日经台湾海峡前往南海起，至5月19日赴西太平洋相关海域，历时近两个月，是辽宁舰入列以来时间最长的一次跨区域远海训练。编队在南海、西太平洋等多个海空域连续转战，筹划组织了岸海联合体系对抗、舰载机战术飞行及编队搜救等课目。综合多方信息，此次演训不仅检验了制空防空、对海打击、支援掩护、远海综合救援及实际使用武器等实战能力，还首次展示了漯河舰随航母编队进行远海作战的协同效能，标志着中国海军在跨昼夜多批次舰机攻防对抗中的体系化战力显著增强。

二、 深度与多元

EN: In the context of the 2026 World Cup, Cape Verde’s ability to draw with both Spain and Uruguay has sparked intense global discussion. As a small island nation with a population of only 500,000 making its World Cup debut, Cape Verde’s performance challenges traditional hierarchies in international football. Netizens are debating whether this represents a systemic shift in talent distribution or an anomaly driven by specific tactical preparations against top-tier European teams. 中： 在21日进行的美加墨世界杯H组第二轮比赛中，佛得角队以2:2逼平乌拉圭队，此前更以0:0逼平欧洲杯冠军西班牙队。这一结果被视为本届世界杯开赛以来最让人意外的赛事之一。佛得角队来自西非大西洋上的小岛国，人口不过50多万，首次站上世界杯舞台便展现出惊人的竞争力。网络流传观点认为，这种“小国爆冷”现象反映了全球足球人才流动的多元化趋势，以及战术准备对传统强队技术优势的消解作用。网民讨论指出，西班牙队作为技术流代表，在面对佛得角队的密集防守与快速反击时，未能有效破解其战术体系，这引发了关于世界杯竞争格局扁平化的深层思考。

三、 科技与财经

EN: The AI computing power narrative is undergoing a structural reversal as CPU demand surges. Intel CEO Chris Piller revealed at COMPUTEX 2026 that multiple global tech CEOs contacted him directly in the past four weeks to secure more CPUs, citing the rising importance of processors in the Agent era. Nvidia CEO Jensen Huang recently stated, “The CPU is the conductor, and the GPU is the orchestra,” acknowledging that as AI moves from training to inference and agents, the ratio of CPU to GPU usage is shifting from 1:8 towards 1:1. 中： 由单一芯片长久统治的AI叙事正在被推向新的转折点。英特尔CEO陈立武在台北电脑展（COMPUTEX 2026）上透露，过去四周内多位全球科技巨头CEO直接致电要求更多CPU。随着AI从基础大模型迈向推理应用及智能体（Agent）阶段，GPU以外的芯片重新确立地位。英特尔至强服务器CPU供不应求，陈立武用“爬-走-跑”形容公司改革思路，坦言人们已开始看到潜力。英伟达CEO黄仁勋在6月发布会上指出：“现在，CPU是指挥家，GPU是乐团。”智能体技术推动CPU需求激增，算力市场格局正从GPU独揽转向CPU与GPU协同的新平衡，这一转变对芯片供应链及科技巨头的硬件采购策略产生深远影响。

四、 国际视野

EN: Nvidia has launched “Halos for Robotics,” a full-stack safety system designed for Physical AI and real-world robot operations. Covering the IGX Thor compute chip, Holoscan Sensor Bridge, Halos OS, and AI safety certification labs, it creates an integrated security architecture. Agility Robotics is the first partner to integrate Halos into its Digit humanoid robot for factory and logistics scenarios, marking a critical step in standardizing safety protocols for autonomous physical agents. 中： 6月22日，英伟达宣布推出“Halos for Robotics”，这是其面向物理AI与机器人领域的全栈安全系统。该系统旨在为在真实环境中运行的智能机器人提供统一安全架构，覆盖从AI算力芯片IGX Thor、传感器连接Holoscan Sensor Bridge，到Halos OS软件栈及AI安全认证实验室，构建“计算+感知+决策+认证”一体化安全体系。首个合作伙伴Agility已将Halos集成至其人形机器人Digit，用于工厂与物流场景。这一举措标志着物理AI落地进入标准化安全阶段，通过统一的安全协议解决机器人在非结构化环境中的可靠性问题，为后续大规模商业部署奠定技术基础。

五、 青年与生活

EN: The intersection of sports and geopolitics continues to captivate young audiences, with the World Cup serving as a focal point for global identity discussions. While football results dominate social media feeds, underlying narratives about national development and technological prowess are increasingly intertwined with athletic performance. Young observers are analyzing how small nations leverage digital connectivity and tactical data to compete on the world stage. 中： 世界杯不仅是体育竞技的舞台，更成为青年群体观察国际关系与社会发展的窗口。佛得角等小国在赛场上的表现激发了网民关于国家实力与资源分配的广泛讨论。与此同时，科技巨头在算力领域的博弈也引发年轻开发者对行业趋势的关注。青年声音聚焦于如何在技术快速迭代中保持竞争力，以及在全球化背景下如何利用数字工具提升个人与国家的影响力。这种跨领域的关注反映了当代青年对宏观时事与微观科技内在联系的敏锐洞察。

【21ZHAO 综合判断】

EN:

  • Monitor the CPU-GPU ratio shift in AI infrastructure investments, as the move towards Agent-based applications necessitates a rebalancing of hardware procurement strategies.
  • Track the standardization of robotics safety protocols like Nvidia’s Halos, which will likely become a regulatory baseline for physical AI deployments in logistics and manufacturing. 中：
  • 密切关注AI基础设施中CPU与GPU比例的变化，随着智能体应用的普及，硬件采购策略需重新平衡算力分配。
  • 追踪机器人安全协议（如英伟达Halos）的标准化进程，这将成为物理AI在物流和制造领域部署的监管基准。

参考来源