Cheyenne Suarez

Nvidia shakes up CPU market with new chip designed for Windows, AI agents

Article (news)

Arthur Fuentes

ABS-CBN News

<div><img src="https://od2-image-api.abs-cbn.com/prod/editorImage/1780383647885nvidia rtx spark edit.jpg" style="width:633px" class="fr-fic fr-dib" alt="Nvidia unveils the RTX Spark ARM-based Windows CPU at Computex. Arthur Fuentes, ABS-CBN News"/>Nvidia unveils the RTX Spark ARM-based Windows CPU at Computex. Arthur Fuentes, ABS-CBN News</div>TAIPEI — After dominating the graphics card market, Nvidia appears poised to carve out a slice of the pie traditionally shared by Intel, AMD, and, to a certain extent, Qualcomm. At Computex, Nvidia unveiled RTX Spark, an ARM-based system-on-a-chip, which the company said will power a new class of Windows PCs purpose-built for the exploding ecosystem of personal AI agents.Nvidia said the RTX Spark combines a 20-core Nvidia Grace CPU, up to 6,144 Blackwell-based GPU cores, and support for up to 128GB of unified memory. The company said it collaborated with mobile chip maker, MediaTek, on the custom CPU design, to ensure power efficiency, performance and connectivity.The new chip is set to debut in slim Windows laptops promising all-day battery life and small, desktop PCs, Nvidia said. What sets apart the RTX Spark from other CPUs is that it is designed to address the needs of users and developers of AI agents, Nvidia said. <div><img src="https://od2-image-api.abs-cbn.com/prod/editorImage/1780384714143asus rtx spark.jpg" style="width:633px" class="fr-fic fr-dib" alt="Nvidia RTX Spark in an upcoming laptop. Arthur Fuentes, ABS-CBN News"/>Nvidia RTX Spark in an upcoming laptop. Arthur Fuentes, ABS-CBN News</div>Driven by the rapid adoption of open-source GitHub projects like OpenClaw and Hermes, developers and users alike are shifting toward agents that adapt to individual workflows. These local assistants can manipulate applications, automate multi-step tasks, and generate content—all without sending sensitive data to the cloud.However, running AI agents securely on-device demands serious hardware. Nvidia said its RTX Spark answers the call, packing a massive 1 petaflop of AI compute and 128GB of unified memory. With 128GB of unified memory, Nvidia said users can render massive 90GB 3D scenes, edit 12K video, run 120-billion-parameter LLMs locally, and play AAA games at 1440p/100+ FPS with ray tracing.Over 100 top software providers and game studios—including Adobe, Blackmagic Design, Remedy, and Xbox—are already adopting RTX Spark to supercharge over 1,000 applications, the company added.<div><img src="https://od2-image-api.abs-cbn.com/prod/editorImage/1780384836293acer dgx spark.jpg" style="width:633px" class="fr-fic fr-dib" alt="Nvidia DGX Spark in an Acer workstation. Arthur Fuentes, ABS-CBN News"/>Nvidia DGX Spark in an Acer workstation. Arthur Fuentes, ABS-CBN News</div>Nvidia’s local AI push isn&#x27;t just for consumers. Scaling its Windows partnership to the enterprise sector, the company also introduced the Nvidia DGX Station for Windows. Billed as the ultimate deskside supercomputer for professionals, this powerhouse brings data-center-class GPU and CPU inference capabilities directly to an office desk, utilizing Windows for seamless enterprise manageability, compatibility, and security.With the launch of the RTX Spark and DGX Station ecosystems, NVIDIA is drawing a line in the sand: the future of AI productivity won&#x27;t just live in massive data centers, but right on your desk and in your lap.

After dominating the graphics card market, Nvidia appears poised to carve out a slice of the pie traditionally shared by Intel, AMD, and, to a certain extent, Qualcomm.

After dominating the graphics card market, Nvidia appears poised to carve out a slice of the pie traditionally shared by Intel, AMD, and, to a certain extent, Qualcomm.