【环球网科技报道 记者 李文瑶】1月7日消息，在日前开幕的2026年国际消费电子展（CES 2026）上，物理AI与边缘AI的融合落地成为贯穿全场 的主线。从自动驾驶、机器人到个人电脑、可穿戴设备，智能化正深入各类终端。 Arm方面表示，Arm 计算平台作为核心算力基石，提供关键算力支撑，助力各类设备实现感知、推理与执行的全链路能力。其中，在今年CES 2026 上亮相的全新 Arm 技术驱动的端点平台，如 Alif Ensemble E8，也预示着环境感知 AI 任务即将在超低功耗的边缘节点上实现稳定运行。 机器人技术从实验室走向规模化商用，成为物理AI落地的重要体现。轮足机器人、清洁配送机器人及人形机器人均在展会上展示了在复杂环境中 自主作业的能力。这些系统的实时感知、决策与控制，同样依赖于Arm架构的高能效计算平台。 在消费电子领域，端侧AI已成为PC、笔记本电脑及平板的标配特性。Windows on Arm生态快速发展，预计2026年将有超百款相关机型上市。同 时，基于Arm架构的苹果MacBook、谷歌Chromebook及小米平板等设备，展示了在本地高效完成AI任务的同时，兼得高性能与长续航的可行 ...

这一轮暴涨并非偶然，而是市场对AI发展阶段认知的深刻修正。美银美林分析师Wamsi Mohan在最新的报告中指出，2026年将是企业级和边缘AI 的拐点。随着多模态AI（包含文本、图像、视频）的普及，数据生成量将呈指数级增长，这将推动硬件支出周期的延续。 美银认为，AI投资的主题正在从资本支出驱动的模型训练，转向以投资回报率（ROI）为核心的AI推理阶段，这一转变将使存储、边缘设备和网 络连接厂商成为继GPU之后的新一轮受益者。 AI推理与多模态：数据爆炸的真正推手 随着AI浪潮从训练阶段向推理应用大规模迁移，被视为"AI工作记忆"的存储板块正迎来前所未有的价值重估。 周二美股市场存储概念股集体飙升，闪迪暴涨27.56%，创下自2月以来的最佳单日表现。西部数据和希捷科技紧随其后，分别大涨16.77%和 14.00%。 这一轮行情的直接催化剂来自英伟达CEO黄仁勋在CES上的讲话。他直言："就存储而言，这目前是一个完全未被开发的市场。这是一个从未存在 过的市场，而且很可能成为全球最大的存储市场，基本上承载全球AI的工作记忆（working memory）。"与此同时，英伟达在CES上展示了针对 代理AI推理优化 ...

在今天的CES展会上，AMD嵌入式平台介绍了两大系列产品：锐龙AI嵌入式P100系列以及锐龙AI嵌入 式X100系列产品。主要应对边缘AI市场的需求。覆盖人工智能、汽车辅助驾驶、医疗、机器人等等相 关领域。提供高性能、高精度、快速响应与低功耗的使用场景。 根据AMD的介绍，锐龙AI嵌入式P100系列在今天正式发布，该系列采用4纳米制程工艺搭载，CPU方面 能够提供4-6个核心，配备1MB L2缓存及8MB L3缓存，实现可预测时延包含512位宽的AVX-512指令集 以及VNN1/BF16，可用于AI运算，集成10GbE w/TSN与Infinity Fabric。同时功耗只有15-54W，并且拥 有极长的寿命，能够在24×7的全天候高强度运行支持十年生命周期。 性能上，锐龙AI嵌入式P100系列的CPU得益于Zen5架构的加持，在SPECrate@2017的测试中，单线程多 线程性能相较于上代提升2倍以上，支持可靠多任务处理，适配汽车、工业、客户端等场景。 GPU方面锐龙AI嵌入式P100系列搭载了RDNA 3.5架构，渲染性能提升35%，支持4K-8K高保真显示， 低时延流式播放，减轻CPU负担，适配高密 ...

Group 1: TI's Automotive Innovations - TI launched three powerful automotive products at CES: the TDA5 series SoC, AWR2188 radar transmitter, and DP83TD555J-Q1 Ethernet PHY [1][4][7] - The TDA5 SoC features a maximum performance of 1200 TOPS and an energy efficiency of over 24 TOPS/W, with a 12-fold increase in AI computing power compared to previous generations [1] - AWR2188 is the industry's first single-chip 8x8 radar solution, enhancing performance by 30% and achieving high-precision detection for targets over 350m [4] - The DP83TD555J-Q1 Ethernet PHY supports nanosecond-level time synchronization and can transmit power and data over the same line, reducing cable design complexity and costs [7] Group 2: ADI's Diverse Solutions - ADI showcased various solutions in automotive, consumer, and robotics sectors, highlighting the A²B 2.0 solution with four times the bandwidth of its predecessor [10] - The automotive solutions include advanced lighting control and ADAS systems utilizing machine vision inputs [10][11] Group 3: NXP's High-Integration Processor - NXP introduced the S32N7 processor series, which integrates multiple vehicle functions on a single chip, potentially reducing total cost of ownership (TCO) by up to 20% [12][15] Group 4: Microchip's Demonstrations - Microchip presented demos including the ASA Motion Link for Qualcomm's Ride platform and a software-free intelligent headlight system using 10BASE-T1S technology [17][18] Group 5: Silicon Labs' New SDK - Silicon Labs launched a new Simplicity SDK for Zephyr, enhancing support for embedded systems and showcasing advancements in Bluetooth wireless technology [19] Group 6: Infineon's Development Kit - Infineon and Flex unveiled a modular development kit for regional control units, aimed at accelerating the development of software-defined vehicle architectures [20] Group 7: ST's Automotive Gateway - ST displayed the Osdyne Automotive Gateway, which enhances vehicle communication and security while reducing wiring complexity [22] Group 8: Ambarella's AI Vision Chip - Ambarella released the CV7 AI vision SoC, built on a 4nm process, achieving over 20% power reduction and more than 2.5 times the AI performance of its predecessor [25] Group 9: NVIDIA's Revolutionary Products - NVIDIA introduced the Rubin platform with six new chips and launched the Alpamayo series for AI-assisted driving development [26][28] Group 10: AMD's AI Innovations - AMD announced several new products, including the MI455X GPU and Ryzen AI 400 series processors, emphasizing its comprehensive AI capabilities [29][30] Group 11: Arm's Technology Trends - Arm focused on five key technology trends at CES, including advancements in autonomous driving, robotics, and smart home devices [31][32] Group 12: Industry Trends - The CES highlighted three major trends: the penetration of AI across all technology layers, the shift towards centralized and software-defined automotive electronics, and the importance of ecosystem collaboration over isolated technology competition [33]

如果您希望可以时常见面，欢迎标星收藏哦~ 我 们 在 谈 到 AMD 的 时 候 ， 更 多 的 时 候 是 关 注 他 们 在 PC 和 数 据 中 心 市 场 的 表 现 。 诚 然 ， 凭 借 Ryzen、EPYC和Instinct系列等CPU和GPU产品的优越表现，AMD在这些市场已然获得了傲人的 成绩。 财务数据显示，AMD在2025年第三季度再次取得亮眼业绩，录得了超出华尔街预期的92亿美元 GAAP营收，这主要得益于市场对高性能计算产品的广泛需求。来到净利润方面，AMD在Q3同比 增长 61% 至 12 亿美元，毛利率因产品组合的丰富而提升至 52%。 然而，正如AMD自适应和嵌入式产品市场管理高级总监Sumit Shah所说，除了上述广为人知的产 品外，AMD的芯片还被广泛应用到多种终端。也正是得益于这些投入，催生了AMD的第三个支柱 ——嵌入式产品。 嵌入式需求变了 关于什么是嵌入式，按照维基百科的说法，所谓嵌入式系统（Embedded system），是一种嵌入机 械或电气系统内部、具有专一功能和实时计算性能的计算机系统。嵌入式系统常被用于高效控制许 多常见设备，被嵌入的系统通常是包含数字 ...

公众号记得加星标⭐️，第一时间看推送不会错过。 日前，TI发布了使用5nm工艺打造的自动驾驶汽车的"大脑"TDA5，也是德州仪器(TI)全新解决方案 的核心。应用这款芯片，即可构建"边缘AI"环境，将每秒运算速度从10万亿次(1 TOPS；1 TOPS为 每秒1万亿次运算)提升至高达1200万亿次(1200 TOPS)。TI表示，这使得车辆即使在面对复杂多变的 道路环境时，也能快速分析数据并做出响应，从而实现L3级自动驾驶。 能效也是一大优势。该芯片每瓦功耗 (W) 可支持 24 TOPS 的计算能力。德州仪器 (TI) 处理器产品 机构部门负责人（副总裁）Roland Schupfli 表示："对于电动汽车而言，单次充电续航里程是一项关 键指标，因此需要功耗更低、性能更高的芯片。"他补充道："TDA5 拥有业界最佳的能效。" 为了实现低功耗、高性能的 TDA5 芯片，德州仪器集成了其神经处理单元 (NPU) 产品 C7。副总裁 Schupfli 表 示 ： " 我 们 在 保 持 功 耗 相 近 的 情 况 下 ， 实 现 了 比 上 一 代 产 品 高 出 12 倍 的 AI 计 算 性 能。"他还补充道 ...

据港交所12月29日披露，国民技术（300077）股份有限公司(简称：国民技术)(300077.SZ)向港交所主板提交上市申请，中信证券为其独家 保荐人。 公司简介 据招股书，国民技术是一家平台型集成电路设计公司，致力于为各类智能终端提供控制芯片与系统解决方案。公司在消费电子、工业控 制和数字能源、智慧家居、汽车电子与医疗电子多个关键领域实现了产品的广泛应用，成功构建了多元化场景覆盖的产品矩阵。 根据灼识咨询的资料，按2024年收入计，国民技术在全球平台型微控制器单元市场中，在中国企业中名列前五，而在全球32位平台型 MCU市场中，在中国企业中名列前三。根据灼识咨询的资料，按2024年收入计，公司在内置商业密码算法模块的中国MCU市场中，排名 第一。 自2000年成立以来，国民技术从专业市场芯片向通用MCU，再到边缘AI计算等高端产品逐步实现跨越式发展，2023年，公司也延伸出 BMS芯片、射频芯片等更丰富的产品体系，并于2024年开始产生收入。自2018年明确平台型集成电路设计公司发展方向后，公司先后推 出多款基于Cortex-M0至M7的32位MCU产品，并持续优化芯片尺寸、功耗及性能，实现了从嵌入式控制 ...

TPU TPU 、、 LLPUPU 、、 GGPUPU AAII芯芯⽚⽚的的过过去去、、现现在在与与未来未来 历史演进 从图形处理到AI基⽯的华丽转⾝ 架构对⽐ 专⽤化与通⽤性的技术权衡 未来展望 异构计算与边缘AI的新时代 核⼼洞察 在⼈⼯智能浪潮席卷全球的今天，算⼒已成为驱动技术⾰命的核⼼引擎。在这场算⼒竞赛中，图形处理 器（GPU）、张量处理器（TPU）和语⾔处理器（LPU）等专⽤芯⽚扮演了⾄关重要的⻆⾊。 GPU凭借NVIDIA的CUDA⽣态，从图形渲染领域华丽转⾝，成为AI训练的基⽯； TPU源于⾕歌对内部算 ⼒危机的"未⾬绸缪"，以专⽤架构重塑了计算效率； LPU则由前TPU团队再创业⽽⽣，精准切⼊推理市 场，以确定性执⾏架构挑战传统范式。 这三款芯⽚的诞⽣与发展，共同谱写了AI硬件从通⽤到专⽤、从训练到推理的演进史诗，并将在未来持 续塑造AI技术的边界与格局。 1. 回顾历史：AI芯⽚的诞⽣与初⼼ 1.1 GPU：从图形处理到AI基⽯的华丽转⾝ ⻩仁勋的远⻅：CUDA⽣态的构建 在⼈⼯智能浪潮席卷全球之前，NVIDIA的核⼼业务聚焦于为电⼦游戏提供⾼性能的图形处理器。然⽽， 公司创始⼈兼CEO⻩ ...

Core Insights - The Chinese semiconductor industry is experiencing a significant transformation, marked by the successful IPOs of domestic GPU companies such as MuXi and Moer Thread, reflecting a collective investor sentiment towards "computing power autonomy" [1][2] - The value focus in the semiconductor sector is shifting from centralized computing power to "ubiquitous intelligence," indicating a profound change in the IoT semiconductor landscape [1] Group 1: Current Trends - The listings of MuXi and Moer Thread occur during a unique time when AI applications are rapidly emerging, making computing power a fundamental resource [2] - The expansion of computing power is moving from data centers to edge and terminal devices, driven by the maturation of AI technology and decreasing costs [3][4] Group 2: Predictions for IoT Semiconductor Industry - Prediction 1: The integration of edge AI will accelerate, leading to widespread application of IoT devices equipped with edge AI capabilities by 2026 [7] - Prediction 2: Modular design (Chiplet) and open architecture (RISC-V) will see significant growth, allowing for more flexible and cost-effective chip designs [10][11] - Prediction 3: Carbon footprint tracking will become a critical design constraint, alongside performance, power consumption, and cost [12][13] - Prediction 4: Localized production of IoT chips will increase, driven by government incentives and the need for supply chain security [16][17] - Prediction 5: AI will evolve from a supportive tool to a co-pilot in the design process, automating routine tasks and allowing engineers to focus on higher-level decisions [18][19] - Prediction 6: Security design will transition from an optional feature to a regulatory requirement, necessitating built-in hardware protections [20][21] Group 3: Strategic Recommendations - For chip design companies: Shift focus from parameter competition to scenario differentiation and evaluate the availability of lightweight NPU reserves [24] - For equipment manufacturers (OEMs): Avoid reliance on cloud-based intelligence and seek SoC suppliers that support edge inference capabilities [27] - For industry investors: Look for companies providing AI EDA plugins, compliance automation tools, and Chiplet interconnect IP, as they are likely to be the winners in the upcoming market [29] Group 4: Future Landscape - The semiconductor industry's value will transition from "peak computing power" to "intelligent density per unit energy consumption" and "lifecycle security compliance" [31] - The global semiconductor supply chain will evolve from a unipolar dominance to a multipolar coexistence, emphasizing the importance of regional chip ecosystems [32]

Group 1: Semiconductor Market Trends - Severe memory shortages are driving a reevaluation of semiconductor equipment targets, with DRAM spot prices soaring by 300%-400% in the past three months, reaching $17 per GB for DDR4 and $13-14 per GB for DDR5 [2] - NAND flash market is experiencing similar trends, with core benchmark products seeing a 200% price increase over the last three months, and contract prices rising by 20%-60% [2] - The memory shortage is expected to continue until at least 2027, leading to significant increases in wafer fab equipment spending, particularly benefiting companies like ASML, VAT Group, and SUSS MicroTec [3][4] Group 2: AI and Component Supply Challenges - AI investments are crowding out supply for non-AI components, leading to potential shortages in memory, passive components, and optical components, which could impact consumer electronics, smartphones, PCs, and automotive electronics [4] - The automotive electronics sector is less affected due to dedicated production lines for automotive-grade products [5] Group 3: Optical and Testing Innovations - AI data centers are driving a surge in bandwidth demand, leading to advancements in optical components and the transition to higher-speed pluggable optical devices [3] - The testing sector is undergoing a structural transformation due to increased chip complexity and rising failure costs, with companies like Technoprobe expanding testing coverage to improve quality [6] Group 4: GaN and Power Semiconductor Opportunities - The shift to 800V architecture in AI data centers, driven by Nvidia, is creating opportunities for GaN technology, similar to the impact of SiC in Tesla applications [8] - AI processor power consumption is projected to grow from 7GW in 2023 to over 70GW by 2030, creating significant market opportunities for suppliers addressing power challenges [9] Group 5: Edge AI and Local Processing - Edge AI is gaining traction, with companies like AMD noting its growth potential, although it remains in the experimental phase [10] - Ambarella anticipates that its defined "edge AI" market will account for 80% of its total revenue by 2025, covering various applications [10] Group 6: Localization of Semiconductor Production in China - There is a significant shift in China's semiconductor capabilities, with local manufacturers facing increased pressure for domestic procurement and improving their scale and quality [11] - The year 2026 is expected to be pivotal as the market recognizes the potential shrinkage of Western companies' market size in China [11][12]