NVIDIA and Microsoft unveiled RTX Spark, a new Windows PC chip platform built to run personal AI agents locally, as the companies move to turn personal computers into more active AI systems rather than devices driven mainly by apps.
Announced at NVIDIA GTC Taipei, RTX Spark pairs NVIDIA’s Blackwell RTX GPU technology with a Grace CPU and Windows-native agent tools, giving laptop and desktop makers a new platform aimed at AI developers, creators and users who want more AI processing handled on their own devices.
Microsoft and NVIDIA target local AI agents
The core focus of the partnership is on personal AI agents that can run directly on Windows PCs, rather than relying only on cloud-based systems.
NVIDIA and Microsoft said they are working on a Windows-native framework for agents that combines new security features with NVIDIA OpenShell, a runtime designed to help control what agents can access and how they operate on a user’s device.
The companies said the system is intended to support agents that can work across Windows applications, search local files, generate images and video, write code and handle cross-app workflows.
NVIDIA said the chip can deliver up to 1 petaflop of AI performance and support up to 128GB of unified memory.
“For forty years, you launched apps. Click. Type. With RTX Spark and Microsoft Windows, you ask — and the PC does the work. RTX Spark brings everything NVIDIA has built — CUDA, RTX, our AI platform — into a single superchip. Local agents. Frontier models. Creative workflows. RTX games. All on a laptop. This is the new PC. The personal AI computer,” said Jensen Huang, founder and CEO of NVIDIA.
RTX Spark brings large models to PCs
RTX Spark is designed to handle demanding local AI workloads, including large language models and long-context processing. NVIDIA said the superchip can run 120-billion-parameter models with up to a 1 million-token context window, while also supporting 4K AI video generation, large 3D scenes and advanced creator workflows.
The chip combines a Blackwell RTX GPU with 6,144 CUDA cores, fifth-generation Tensor Cores and a 20-core Grace CPU, connected through NVIDIA’s NVLink-C2C interconnect.
Adobe tools get AI-native upgrades
Adobe is also working with NVIDIA and Microsoft to rebuild parts of Photoshop and Premiere for RTX Spark, with a focus on AI-assisted creative work. NVIDIA said Adobe Premiere will use RTX Spark’s unified memory, Blackwell GPU and TensorRT software for real-time editing, color correction and AI acceleration, while Photoshop will be optimized for GPU-based compositing and AI-native workflows.
NVIDIA said RTX Spark-powered laptops and compact desktops are expected this fall from ASUS, Dell, HP, Lenovo, Microsoft Surface and MSI, with Acer and GIGABYTE models to follow.
Personal AI moves closer to Windows users
The launch puts AI agents at the center of the next generation of Windows PCs, as NVIDIA and Microsoft expect more users to integrate AI tools into everyday work across files, apps and local workflows.
The partnership gives Microsoft another path to expand AI across Windows, while helping NVIDIA push its AI hardware beyond servers and high-end workstations into everyday PCs.