Jensen Huang ComputeX 2025 Key Note
Here is the translation:
Summary
On May 19, 2025, NVIDIA CEO Jensen Huang held a keynote speech at Taipei’s Nangang Exhibition Center. This lecture may signal the direction of the industry for the next decade. Jensen Huang described NVIDIA’s transformation from a chip company to an AI infrastructure giant and systematically outlined the technical framework and industrial logic of the intelligent revolution era.
Main Contents
NVIDIA’s 30-year development history is marked by three key transformations. Founded in 1993, NVIDIA was initially a technology company focused on graphics chips. In 2006, CUDA parallel computing architecture was born, and NVIDIA first showcased the general computational ability of GPUs to the public. In 2016, the tide of deep learning began to emerge, and NVIDIA keenly captured the revolutionary demand for AI in computing architectures.
In 2025, NVIDIA completed its third strategic upgrade, transforming from an “AI technology supplier” to a “builder of intelligent infrastructure”. Jensen Huang compared this transformation to GE and Siemens during the electric revolution. He announced that NVIDIA’s mission has shifted from manufacturing chips to building the underlying architecture that supports global smart economies.
Key Points
- NVIDIA has completed three key transformations and become the core engine of the global AI revolution.
- Jensen Huang divided AI development into three stages: perception AI, generation and reasoning AI, and physical AI.
- NVIDIA introduced two key technologies: Newton physical engine and Isaac Groot platform.
- NVIDIA is trying to reduce the barrier to entry for AI development through a consistent architecture from data centers to desktops.
- NVIDIA proposed the concept of “AI native architecture”, reconfiguring the three-layer architecture of computing, storage, and networking.
- NVIDIA plans to build its overseas headquarters, “NVIDIA Constellation”, in Taipei’s Beitou Shilin Science Park.
- Jensen Huang called on citizens to support the construction of the new headquarters.
- NVIDIA is trying to build a vertically integrated AI infrastructure empire that spans chips, systems, software, and services through an open ecosystem + dual integration strategy.
Translation
Summary
2025年5月19日，英伟达CEO黄仁勋在台北流行音乐中心举办了一场主题演讲。这场演讲可能预示着未来十年的产业走向。黄仁勋描绘了英伟达从芯片公司向AI基础设施巨头的转型路径，并首次系统性地阐述了智能革命时代的技术框架与产业逻辑。
Main Contents
英伟达30年发展历程以三次关键转型为切入点。1993年成立之初，英伟达是专注于图形芯片的技术公司。2006年CUDA并行计算架构诞生后，英伟达首次将GPU的通用计算能力展现出世人面前。2016年深度学习的浪潮初现，英伟达敏锐捕捉到了AI对计算架构的颠覆性需求。
2025年，英伟达完成第三次战略升级，从“AI技术的供应商”转型为了“智能基础设施的构建者”。黄仁勋将这个转变类比为电力革命中的GE与西门子。他宣布英伟达的使命从制造芯片转变为构建支撑全球智能经济的底层架构。
Key Points
- 英伟达完成了三次关键转型，成为全球AI革命的核心引擎。
- 黄仁勋将AI发展划分为三个阶段：感知AI、生成与推理AI和物理AI。
- 英伟达推出了两大关键技术：Newton物理引擎和Isaac Gr00t平台。
- 英伟达正在试图通过“从数据中心到桌面”的一致性架构来降低AI的开发门槛。
- 英伟达提出了“AI原生架构”的理念，重构计算、存储、网络的三层架构。
- 英伟达计划在台北北投士林科技园区建造“英伟达Constellation”海外总部。
- 黄仁勋呼吁支持市民支持新总部建设。
- 英伟达正在试图通过开放生态+垂直整合的双重策略来构建一个横跨芯片、系统、软件、服务的AI基础设施帝国。
Reference:
https://www.youtube.com/watch?v=TLzna9__DnI