At the recent GTC conference, NVIDIA CEO Jensen Huang unveiled six new chips in one go, drawing significant attention from the tech industry. These chips span multiple cutting-edge domains including AI training and inference, autonomous driving, robotics, and high-performance computing, showcasing NVIDIA’s comprehensive strategy in the AI hardware ecosystem. Among them is the next-generation Blackwell architecture GPU, optimized for large-scale AI model training; the Jetson Thor chip for edge AI inference, offering several times the performance of its predecessor; and a new platform combining the Grace CPU with the Hopper GPU for data centers. Huang also introduced DRIVE Thor, designed specifically for autonomous vehicles, and the GB200 superchip for general-purpose accelerated computing. Beyond raw computational power, these new chips emphasize energy efficiency and seamless integration with software ecosystems, highlighting NVIDIA’s strategic shift from hardware provider to full-stack AI solutions company. Huang stressed that AI will be ubiquitous in the future, with chips serving as its core engine. This concentrated product launch not only reinforces NVIDIA’s leadership in AI chips but also offers developers and enterprises worldwide a broader range of technological options.
在近期的GTC大会上，英伟达CEO黄仁勋一口气发布了六款全新芯片，引发业界广泛关注。这六颗芯片涵盖人工智能训练、推理、自动驾驶、机器人及高性能计算等多个前沿领域，展现了英伟达在AI硬件生态中的全面布局。其中包括新一代Blackwell架构GPU，专为大规模AI模型训练优化；用于边缘AI推理的Jetson Thor芯片，性能较前代提升数倍；以及面向数据中心的Grace CPU与Hopper GPU融合的新平台。此外，黄仁勋还介绍了专为自动驾驶设计的DRIVE Thor芯片和用于通用加速计算的GB200超级芯片。这些新品不仅在算力上实现飞跃，更强调能效比与软件生态协同，凸显英伟达从硬件到全栈AI解决方案的战略转型。黄仁勋强调，未来AI将无处不在，而芯片是其核心驱动力。此次密集发布，不仅巩固了英伟达在AI芯片领域的领导地位，也为全球开发者和企业提供了更丰富的技术选择。
原创文章，作者：admin，如若转载，请注明出处：https://avine.cn/9361.html