据悉，GTC 2026大会的主题演讲将于3月15日在加利福尼亚州圣何塞举行，核心聚焦AI基础设施竞赛的 新时代。黄仁勋坦言，这些全新芯片的研发极具挑战，"所有技术都已逼近极限"，但结合其过往履约记 录，业界对英伟达此次新品充满期待。 （原题为：重磅预告！黄仁勋将在GTC 2026发布"世界前所未见"芯片） 2月19日消息，据外媒wccftech报道，英伟达首席执行官黄仁勋在接受媒体采访时，对即将到来的GTC 2026大会进行预热，明确表示将在会上揭晓"世界前所未见"的全新芯片，引发业界广泛关注。作为AI芯 片领域的领军者，英伟达此次重磅预告，被认为将进一步巩固其在AI基础设施领域的领先地位。 来源：环球网 目前，新品具体型号尚未披露，但外界普遍猜测，大概率出自两大芯片系列：一是Rubin系列的衍生产 品（如此前曝光的Rubin CPX），该系列已于2026年CES大会上亮相，包含6款全新设计芯片，目前已全 面量产；二是下一代Feynman系列芯片，该系列被称为"革命性"产品，英伟达正探索以SRAM为核心的 广泛集成，或通过3D堆叠技术整合LPUs，不过相关细节尚未确认。 值得注意的是，英伟达正适配AI算力需求 ...

Group 1 - Nvidia and Meta have announced a multi-year, multi-billion dollar chip procurement agreement, where Meta will purchase millions of Nvidia's latest AI acceleration chips, including the upcoming "Vera Rubin" series [1][3] - Meta's CEO Mark Zuckerberg stated that the company plans to invest up to $135 billion in AI infrastructure by 2026, nearly double its spending in 2025 [3] - Despite challenges in its own AI chip development, Meta continues to rely on Nvidia's established solutions due to technical challenges and deployment delays in its self-developed AI chip projects [3] Group 2 - The transaction highlights a shift in AI computing focus from "model training" to the "inference" phase, with Nvidia's CEO Jensen Huang adjusting the product strategy to sell CPUs as standalone products rather than integrated with GPUs [3] - The growing demand for efficient, low-latency inference capabilities is driving this change in hardware architecture, as noted by industry analysts [3]

Core Viewpoint - Nvidia is attempting to challenge Tesla and a few other companies in the autonomous vehicle sector, while Tesla's CEO Elon Musk expresses no concern about this competition [1][4]. Group 1: Nvidia's Developments - Nvidia launched a series of open-source autonomous driving models named Alpamayo, which are designed to make "human-like" decisions and handle rare or tricky scenarios through reasoning rather than traditional rule-based methods [5][6]. - The new model series is expected to enhance Nvidia's influence in the growing autonomous vehicle industry, which the company views as a significant opportunity [6]. - Nvidia's CEO Jensen Huang stated that the company has been developing autonomous vehicle technology for eight years and has thousands of employees dedicated to this sector [6]. Group 2: Industry Reactions and Competitors - Companies like Uber, Jaguar Land Rover, and Lucid Group are interested in using Alpamayo to develop fully autonomous vehicles, and they have previously collaborated with Nvidia on autonomous technology [6]. - Mercedes-Benz announced plans to launch a CLA model equipped with advanced driver assistance features similar to those offered by Tesla later this year [6]. - Lucid and Uber also unveiled a near-production vehicle designed for autonomous taxi services this week [6]. Group 3: Market Impact - Following Nvidia's announcements, its stock rose by 1.6%, while Tesla's stock faced selling pressure, dropping approximately 3% [7]. - Nvidia introduced its next-generation AI, the Vera Rubin series of chips, which is expected to significantly improve performance and cost-effectiveness compared to the current Blackwell platform [7]. - Tesla plans to reduce its reliance on Nvidia chips in its data centers, while xAI is collaborating with Nvidia to build a large data center in Saudi Arabia, expected to utilize the Rubin platform [7][3].

Core Viewpoint - NVIDIA's GTC 2025 conference highlighted advancements in AI technologies, particularly focusing on Agentic AI and its implications for global data center investments, which are projected to reach $1 trillion by 2028 [3]. Group 1: AI Development and Investment - Huang Renxun introduced a three-stage evolution of AI: Generative AI, Agentic AI, and Physical AI, positioning Agentic AI as a pivotal phase in AI technology development [3]. - The scaling law indicates that larger datasets and computational resources are essential for training more intelligent models, leading to significant investments in data centers [3]. Group 2: Product Launches and Innovations - The Blackwell Ultra chip, designed for AI inference, is set to be delivered in the second half of 2025, with a performance increase of 1.5 times compared to its predecessor [4]. - NVIDIA's Quantum-x CPO switch, featuring 115.2T capacity, is expected to launch in the second half of 2025, showcasing advanced optical switching technology [5]. - The introduction of the AI inference service software Dynamo aims to enhance the performance of Blackwell chips, alongside new services for enterprises to build AI agents [6].