Nvidia于台湾Computex科技展上发表其RTX Spark晶片，将统一记忆体、RTX显示卡及全新N1 ARM架构CPU整合为单一「超级晶片」。不同于Microsoft在2024年推出的Copilot+ PC——搭载NPU与16 GB记忆体却缺乏在本地端运行大型语言模型的效能——RTX Spark笔记型电脑提供高达128 GB的统一记忆体以及媲美独立RTX 5070的GPU效能。这些规格搭配Nvidia的CUDA软体平台，使RTX Spark成为首批真正配得上「AI PC」称号的Windows装置，直接挑战此前唯一能在本地端运行基础级AI模型的Apple MacBook Pro。
包括HP、Asus、Dell与Lenovo在内的主要OEM厂商将生产RTX Spark笔记型电脑，而Microsoft的Surface Laptop Ultra——配备38.1公分（15吋）Mini-LED显示萤幕——则是这一新产品线的旗舰机型。高阶配置估计售价将超过4,000美元，与同等规格的MacBook Pro机型相当。该晶片支援高达20核心的CPU，并规划推出较低阶配置以触及包括游戏玩家和创作者在内的更广泛受众。Nvidia亦瞄准紧凑型桌上型电脑市场推出小型化（SFF）PC，目标直指Apple Mac Mini——后者正面临延长的出货等待时间，Apple将此归因于出乎意料的快速AI应用普及。
RTX Spark将CUDA——Nvidia在资料中心广泛使用的成熟AI开发平台——整合至消费级PC中，可望带来远超当前Windows笔记型电脑的AI推理效能。Intel今年发表的Core Ultra Series 3晶片使Dell XPS 14等笔记型电脑的电池续航力更接近MacBook Pro水准，但这些配置的记忆体上限仅为64 GB。透过在ARM架构中统一CPU、GPU与记忆体，RTX Spark消除了长期困扰Windows笔记型电脑的独立GPU在电池续航力与风扇噪音方面的取舍问题。其结果是催生出全新类别的Windows装置，可能对Intel、AMD和Qualcomm的市场地位产生重大影响。
Nvidia announced its RTX Spark chips at the Computex tech expo in Taiwan, combining unified memory, RTX graphics, and the new N1 ARM-based CPU into a single "superchip." Unlike Microsoft's Copilot+ PCs introduced in 2024—which featured NPUs and 16 GB of RAM but lacked the power to run large language models locally—RTX Spark laptops offer unified memory up to 128 GB and GPU performance comparable to a discrete RTX 5070. These specifications, paired with Nvidia's CUDA software platform, position RTX Spark as the first Windows devices capable of genuinely earning the "AI PC" label, directly challenging Apple's MacBook Pro as the only prior option for running foundation-level AI models locally.
Major OEMs including HP, Asus, Dell, and Lenovo will produce RTX Spark laptops, while Microsoft's Surface Laptop Ultra—featuring a 38.1-centimetre (15-inch) Mini-LED display—stands as the flagship of the new lineup. High-end configurations are estimated to exceed $4,000, comparable to similarly equipped MacBook Pro models. The chips support up to a 20-core CPU, and lower-tier configurations are planned to reach broader audiences including gamers and creators. Nvidia is also targeting the compact desktop segment with small-form-factor PCs, aiming at Apple's Mac Mini, which faces extended shipping delays that Apple attributes to unexpectedly rapid AI adoption.
RTX Spark's integration of CUDA—Nvidia's established AI development platform widely used in data centres—into consumer PCs could yield AI inference performance far beyond current Windows laptops. Intel's Core Ultra Series 3 chips, released this year, brought laptops like the Dell XPS 14 closer to MacBook Pro-level battery life, but those configurations cap at 64 GB of memory. By unifying CPU, GPU, and memory in an ARM-based architecture, RTX Spark eliminates the discrete-GPU trade-offs in battery life and fan noise that have long plagued Windows laptops. The result is an entirely new class of Windows device that could significantly impact Intel, AMD, and Qualcomm's market positions.