所有标记为「nvidia」的文章
在 Google Cloud Next 上，Google 和 Nvidia 概述了旨在降低大规模 AI 推理成本的基础设施计划，凸显了模型服务的经济性正成为主要战场。
Recursive Superintelligence 成立仅四个月，据报道已融资至少 5 亿美元，目标是构建能够自我改进的 AI 系统。
Nvidia 研究人员表示，Lyra 2.0 能从单张图像生成连贯的 3D 环境，延伸范围约 90 米，并可导出到 Isaac Sim 等仿真工具中用于机器人训练。
Cadence Design Systems借助CadenceLIVE活动扩大了与Nvidia的AI相关合作，并推出了与Google Cloud的新集成，凸显工程软件正围绕AI基础设施重构。
Span 表示，正在与 Nvidia 及合作伙伴推进名为 XFRA 的网络，该网络将把高性能计算节点部署到家庭和小型企业中，并利用智能配电面板把富余电力容量转化为分布式 AI 基础设
总部位于新加坡的Firmus又融资5.05亿美元，投后估值达到55亿美元，这凸显出投资者对AI基础设施以及围绕Nvidia下一代计划建设的数据中心项目的强烈兴趣
中国已批准英伟达向中国客户销售其H200 AI芯片，结束了长期的监管冻结期，同时该公司正在为中国市场开发一款降级的推理芯片。
在GTC 2026大会上,Nvidia宣布了其物理AI平台的重大扩展,包括通过与Uber的合作在洛杉矶部署自动驾驶汽车(从2027年开始)、为FANUC和ABB工业机器人配备Nvidia技术,以及针对人形机器人的新基础模型。
Nvidia将从2027年上半年开始为Uber自动驾驶出租车提供其完整的自动驾驶软件栈，目标是到2028年底在四个大陆的28个全球市场中部署。
MassRobotics, NVIDIA, and AWS have named the nine Physical AI Fellowship startups in their second cohort, providing mentorship, cloud computing resources, and industry connections to early-stage companies building AI-powered physical systems.
Nvidia and former OpenAI CTO Mira Murati's startup announce a long-term AI partnership, signaling a major shift in the competitive AI landscape.
ABB Robotics is integrating NVIDIA Omniverse into RobotStudio, promising physics-based simulation that could cut robot deployment setup times by 80 percent.
Nvidia reported fiscal Q4 revenue of $68.1 billion, a 73% year-over-year surge, as hyperscalers collectively committed $650 billion in AI infrastructure spending for 2026. Despite the blowout results, investors remain cautious about whether the AI boom can sustain its breakneck pace.
Nvidia has reported yet another record-breaking quarter driven by insatiable demand for AI chips, with CEO Jensen Huang declaring that 'the demand for tokens in the world has gone completely exponential.' The results come as hyperscaler capital expenditure on AI infrastructure continues to shatter expectations.
NVIDIA has added Cosmos Policy to its suite of world foundation models, creating a new robot control framework that builds on the Cosmos Predict-2 model. The system is designed to help robots learn complex manipulation tasks by understanding the physical dynamics of their environment.
As AI models grow larger and inference demand scales, the industry's focus is shifting from GPU scarcity to memory constraints. High Bandwidth Memory from SK hynix, Samsung, and Micron is emerging as the critical — and increasingly expensive — component in AI infrastructure.
