What Is NVIDIA RTX Spark and What Makes an AI PC?
NVIDIA RTX Spark is a new superchip designed for AI PCs that combines an Arm-based CPU, a Blackwell RTX GPU, and unified memory to run advanced AI models locally on next-generation laptops and desktops without depending on the cloud. AI PCs are computers built around this kind of on-device AI processing, using neural processing units (NPUs) alongside CPUs and GPUs to run assistants, chatbots, and other AI agents directly on the machine. Instead of sending your data to remote servers, many tasks—from summarising documents to generating images—happen on your laptop. Some AI PCs can even train models locally, a task that used to require specialised servers. According to HP, “AI PCs accounted for 44% of its PC shipments in the second quarter,” showing that buyers are already starting to shift towards this new category.
Inside RTX Spark: How the Superchip Powers On-Device AI
RTX Spark is built by NVIDIA with Microsoft and MediaTek to “reinvent the PC” around on-device AI agents. Under the hood, RTX Spark combines up to 20 power-efficient Arm CPU cores with an NVIDIA Blackwell RTX GPU and up to 128 GB of unified memory, reaching about 1 petaflop of AI performance. This layout keeps data and models close to the compute units, cutting delays and making it possible to run large models inside thin-and-light AI PCs. NVIDIA and Microsoft also tuned the Microsoft Power and Thermal Framework (MPTF) for RTX Spark, so these next-generation laptops can maintain performance without overheating or draining the battery too fast. The result is a single superchip that can handle everyday work, gaming, and heavy AI workloads without bouncing between separate processors and memory pools.
Microsoft’s Next-Generation AI PCs and Jensen Huang’s Demo
Microsoft is building its next-generation AI PCs around NVIDIA RTX Spark, aiming to make thin-and-light Windows laptops that are powerful enough to run advanced AI models locally. CEO Jensen Huang introduced RTX Spark ahead of the Computex conference, showing a design that brings AI agents directly onto laptops and compact desktops instead of relying on cloud servers. These new systems will arrive from brands such as ASUS, Dell, HP, Lenovo, Microsoft, and MSI, with Acer and Gigabyte expected to follow. Microsoft’s own Surface Laptop Ultra is one of the first flagship devices, pairing Windows on RTX Spark with support for key apps and AI tools. At the same time, NVIDIA is adding OpenShell on Windows to strengthen security and containment for AI agents, and a Prism emulator so 32-bit and 64-bit x86 apps continue to run on these Arm-based machines.
Why On-Device AI Processing Matters for Everyday Users
On-device AI processing means your AI assistant, chatbot, or creative tool runs on your laptop instead of in a distant data center. That can translate into faster responses, better performance offline, and fewer delays when editing video, generating images, or running AI-powered search. Some experts also argue that keeping more AI work local can help privacy, since personal data does not have to be sent to the cloud for every task. At the same time, there are valid concerns. Microsoft’s earlier “recall” feature—which logged actions on a device for later search—sparked privacy and security criticism until stronger safeguards were added and the feature was made optional. For many buyers, the appeal of AI PCs will depend on whether the benefits of smarter on-device tools outweigh concerns about how much their computers record and store.
Should You Upgrade to an RTX Spark AI PC?
RTX Spark AI PCs promise real gains if you rely on AI-heavy workloads. Creators can speed up video compositing, editing, 3D rendering, and AI-assisted content creation, with native support for tools like Blender, DaVinci Resolve, Photoshop, and Premiere Pro. Gamers benefit from ray tracing and AI-enhanced graphics in titles such as VALORANT, League of Legends, PUBG: Battlegrounds, Alan Wake 2, Naraka: Bladepoint, and War Thunder. Developers can run AI stacks locally using GitHub Copilot, Claude Code, Cursor, ComfyUI, and upcoming support for CUDA-accelerated PyTorch, TensorRT, Hugging Face frameworks, and more. However, supply constraints and rising component costs mean AI PCs may carry higher price tags, and research firm IDC expects global PC shipments to decline even as market value rises past 274 billion. If you want faster, local AI features, RTX Spark next-generation laptops are worth a close look; if you only browse and stream, a standard PC may still be enough.
