What RTX Spark Is and Why It Matters
NVIDIA RTX Spark is an Arm-based Windows PC platform co-developed with Microsoft that combines a MediaTek-designed CPU, a Blackwell-based GPU and unified memory to bring AI-accelerated computing to client devices for creators, gamers and developers. Announced by Jensen Huang at major industry keynotes, RTX Spark marks NVIDIA’s most direct entry into the PC client market after years of focusing on graphics and data centers. The platform is built on the GB10 “Grace Blackwell” Superchip, adapted from NVIDIA’s DGX Spark AI systems into a configuration suited for laptops and desktops. More than a single SoC, RTX Spark comes with a multi-generation roadmap, signaling that NVIDIA and Microsoft see Arm-based Windows PCs as a long-term bet rather than another short-lived “Windows on Arm” experiment. That strategic intent, plus a broad OEM lineup, positions RTX Spark as a serious alternative to traditional x86 PCs.
Inside the RTX Spark Platform: Arm CPU, Blackwell GPU and Unified Memory
At the heart of the RTX Spark platform is the GB10 Superchip, a multi-die package built on TSMC’s 3 nm process. One dielet houses a 20-core Arm v9.2 CPU cluster custom-designed by MediaTek, while the other integrates a GB100 Blackwell GPU with 5th-generation Tensor Cores, RTX ray tracing and up to 31 TFLOPs of FP32 performance. Unified LPDDR5X memory provides up to 128 GB and 301 GB/s of bandwidth, shared across CPU and GPU through NVIDIA’s NVLINK-based C2C interface, with aggregate access up to 600 GB/s. This design gives RTX Spark systems a single memory pool that Windows can dynamically allocate, with NVIDIA stating that a 128 GB system can dedicate up to 111 GB to the GPU without entering firmware menus. The SoC also includes 16 MB of system cache, support for up to four high-resolution displays, and security features such as dual secure roots and TPM support.
Target Users: Creators, Gamers and AI Developers on Windows
RTX Spark is tailored for three overlapping groups: content creators, gamers and AI developers who want AI computing on Windows without depending on the cloud for every task. For creators, the Blackwell GPU’s Tensor Cores, CUDA stack and unified memory aim to speed up video editing, 3D rendering and generative media workflows that need access to large models and datasets. Gamers gain features such as DLSS, Reflex and RTX ray tracing integrated into an on-package GPU, with support for up to 8K output via HDMI 2.1a and multiple 4K displays over DisplayPort. AI developers get a client platform aligned with NVIDIA’s broader software ecosystem, from TensorRT to vLLM, making it easier to move workloads between RTX Spark PCs, DGX Spark workstations and cloud infrastructure. According to NVIDIA, RTX Spark systems can reach “up to 1 petaflop of AI performance,” underscoring the focus on on-device AI.
Microsoft and MediaTek: The Partnerships Behind RTX Spark
RTX Spark is the product of a three-year collaboration among NVIDIA, Microsoft, Arm, MediaTek and other ecosystem partners, aimed at redefining what a Windows PC can do for AI workloads. Microsoft’s role is central on both the platform and software sides, ensuring that Windows on Arm, new security features and AI agent frameworks are tuned for RTX Spark hardware. A notable addition is NVIDIA OpenShell, a runtime that controls how AI agents access local data, interact with applications and connect to cloud services, aligning with Microsoft’s push for on-device AI assistants. MediaTek contributes its experience building Arm-based system-on-chips, designing the 20-core N1X CPU at the heart of RTX Spark’s GB10 package. This shared effort gives NVIDIA a ready-made entry point into premium Windows laptops and desktops from OEMs including Microsoft Surface, Dell, HP, ASUS, Lenovo and MSI, with more than 30 laptops and 10 desktops expected at launch.
Challenging Intel and AMD’s x86 PCs With a Multi-Generation Roadmap
RTX Spark arrives at a moment when Arm-based Windows PCs and on-device AI are reshaping expectations for personal computers. For decades, Intel and AMD defined the Windows PC processor landscape with x86 architectures, while NVIDIA supplied discrete GPUs. RTX Spark changes that balance by putting a full Arm-based SoC—with CPU, GPU and unified memory—at the center of an AI-focused Windows PC platform. Unlike earlier Windows on Arm attempts that lacked clear continuity, NVIDIA and Microsoft are backing RTX Spark with a multi-generation roadmap, signaling ongoing performance and efficiency gains over time. The platform’s design, borrowing features like NVFP4 precision, NVLINK C2C and high-bandwidth unified memory from data center products, brings server-grade ideas into client systems. As more OEMs adopt RTX Spark and developers optimize for Arm-based Windows, NVIDIA is setting up a direct, long-term NVIDIA–Intel–AMD competition inside the PC client market.
