What RTX Spark PCs Are and Why They Matter
RTX Spark PCs are a new class of AI Windows PCs that combine NVIDIA’s Blackwell architecture, Arm-based CPUs, and unified memory to run advanced AI agents and generative models locally without depending on the cloud for every task. Designed with Microsoft, these RTX Spark systems fuse a Blackwell RTX GPU, a 20-core NVIDIA Grace CPU developed with MediaTek, and up to 128GB of unified LPDDR5X memory into an AI-first platform capable of delivering up to 1 petaflop of AI performance. According to NVIDIA CEO Jensen Huang, “The PC is being reinvented. With RTX Spark and Microsoft Windows, you ask, and the PC does the work.” That reinvention shifts the PC from a static box of apps to an active assistant that can act on your behalf, using local AI inference to search files, generate content, automate workflows, and respond faster than cloud-bound tools.
Inside the Blackwell Architecture: Data Center DNA in a Laptop
The Blackwell architecture at the heart of RTX Spark PCs is the same technology family that powers NVIDIA’s latest data center AI products, now scaled down into thin-and-light Windows machines. Flagship configurations combine up to 20 Arm-based CPU cores with 6,144 Blackwell GPU cores and up to 128GB of unified memory, giving these AI Windows PCs more in common with an AI workstation than a traditional notebook. NVIDIA brings its wider software stack along for the ride, including CUDA, TensorRT, RTX graphics, DLSS, OptiX, Reflex, and G-SYNC, so developers and enterprises can run on the same foundations used in NVIDIA-powered servers but on a personal device. In practice, that means local AI inference for large models, faster content generation, and responsive AI agents Windows users can rely on even when network connections are weak or cloud services are busy.
AI Agents on Windows: Local, Private, and Always-On
RTX Spark PCs are built specifically for AI agents Windows users can run locally, turning the operating system into a hub for task automation instead of a patchwork of standalone apps. Microsoft and NVIDIA describe a future where these agents live inside Windows experiences and third-party applications, able to search local files, generate images or video, and coordinate tasks across workflows without sending every request to remote servers. Local AI inference on Blackwell hardware keeps sensitive data on-device and improves responsiveness. NVIDIA OpenShell adds security and control by defining what agents are allowed to do, deciding when to use local models, and masking personal information before anything goes to the cloud. Microsoft says its goal is to deliver “unmetered intelligence” on every Windows PC, and RTX Spark marks a concrete step toward always-on, privacy-aware assistants that feel more like co-workers than chatbots.
Creators, Gamers, and Developers: New Workflows on AI Windows PCs
For creators and gamers, RTX Spark PCs promise workstation-like performance in portable machines. With ray tracing, AI-enhanced graphics, and 1 petaflop of AI compute, these systems accelerate video compositing, high-resolution video editing, 3D rendering, and AI-assisted content creation while keeping workloads local. Microsoft and NVIDIA have optimized RTX Spark for popular apps such as Blender, DaVinci Resolve, Photoshop, and Premiere Pro, and they integrated the Microsoft Power and Thermal Framework so thin-and-light designs can sustain high performance on the go. Gamers benefit from RTX features like DLSS, Reflex, and G-SYNC, plus improved support for Windows on Arm games. Developers can build and test AI agents, generative tools, and coding assistants directly on RTX Spark hardware, using the same CUDA and TensorRT-based stack found in data centers, but in a personal, always-available form factor.
Beyond Copilot+ PCs: The Next Wave of AI-First Personal Computing
By bringing data center-grade AI capabilities into consumer Windows PCs, RTX Spark expands the AI PC market beyond the current wave of Copilot+ PCs. NVIDIA plans more than 30 laptops and 10 desktops based on its new superchip over time, partnering with Microsoft, Dell, HP, Lenovo, ASUS, and MSI to ship systems starting this fall. RTX Spark PCs include the Prism emulator to run 32-bit and 64-bit x86 apps, which helps bridge existing Windows software into this new Arm-based era. The result is a PC that feels less like a static productivity device and more like a platform for AI-native computing, where agents, models, and traditional applications share the same desktop. As AI demand continues to rise, RTX Spark positions personal computers as first-class AI devices, giving everyday users, creators, and developers access to local AI agents that previously required data center infrastructure.
