TAIPEI, Taiwan (NVIDIA GTC), June 8, 2026 – NVIDIA brought its Grace Blackwell AI infrastructure to Windows DGX Station for Windows, a deskside system built to run enterprise AI agents locally. The system supports AI models of up to 1 trillion parameters and connects always-on agents to Windows applications, design tools and engineering workflows.
DGX Station for Windows addresses a gap between enterprise AI infrastructure and everyday business software. Training, fine-tuning, large-scale inference and multi-agent development have relied on Linux-based data center systems, while many Fortune 500 companies run daily productivity, creative, design and engineering work on Windows.
DGX Station for Windows applies the NVIDIA DGX Station system design to Windows-based enterprise environments. NVIDIA describes it as the first deskside AI supercomputer to bring GB300 Grace Blackwell-class AI infrastructure to Windows users.
“As enterprises scale AI agents across their organizations, they need AI infrastructure that can connect directly to the applications and workflows that power their business,” said Chris Marriott, vice president of enterprise platforms at NVIDIA. “DGX Station delivers supercomputing-class AI directly into Windows, where millions already design, engineer, research and create every day.”
“For decades, Microsoft and NVIDIA have partnered to advance the most powerful computing platforms in the world,” said Pavan Davuluri, executive vice president of Windows + Devices at Microsoft. “Today, we’re taking that collaboration to the next level, scaling the full power of Windows from thin-and-light PCs to data-center-class workstations with DGX Station powered by GB300. This unlocks a new class of AI performance on Windows, the platform enterprises trust for security, manageability and compatibility.”
Introducing DGX Station for Windows
At the hardware level, the DGX Station uses the NVIDIA GB300 Grace Blackwell Ultra desktop superchip. It connects NVIDIA Blackwell Ultra GPU with a 72-core NVIDIA Grace CPU through the NVIDIA NVLink-C2C interconnect for high-speed communication between system components.
The system includes up to 748GB of coherent memory and up to 20 petaflops of FP4 performance. It can also be paired with an NVIDIA RTX PRO 6000 Blackwell Workstation GPU for AI compute, ray-traced visualization and simulation workloads.
NVIDIA also integrates the ConnectX-8 SuperNIC into DGX Station for Windows. The network interface supports speeds of up to 800Gb/s, allowing data transfer for AI workloads and high-speed connectivity between multiple DGX Station systems.
A Window Into the World of Always-On AI Agents
Developed with Microsoft, DGX Station for Windows provides local infrastructure for enterprises building and running AI agents on Windows. NVIDIA said the system can run hundreds of agents on tasks at the same time while supporting frontier-scale models locally.
These agents can connect to 3D design and engineering applications used by developers, designers and engineers. Within supported software environments, they can work with application tools and process repetitive workflow steps.
For enterprise IT teams, DGX Station for Windows brings GB300 deployments into existing Windows security, compliance and fleet management systems. AI agents run inside an environment governed through Microsoft administration tools.
The system also supports Linux workloads through Windows Subsystem for Linux. Enterprise features, including deployment controls and system updates, help organizations maintain security, compliance and operational readiness across DGX Station fleets.
NVIDIA and Microsoft are extending the same Windows agent strategy to smaller PC systems through NVIDIA RTX Spark. That work brings NVIDIA AI capabilities to slim laptops and small desktop PCs for personal agent use cases.
Secure Agent Development With NVIDIA OpenShell
For secure agent development, NVIDIA OpenShell provides an open-source runtime for autonomous agents running on DGX Station. The runtime governs how agents act, operate tools and interact with the rest of the system before deployment at data center scale.
NVIDIA OpenShell uses new Windows security and containment primitives to create a separate isolated sandbox for each agent. It separates application-layer operations from infrastructure-layer policy enforcement, keeping security and privacy policies outside the agent’s control and applying restrictions at the system level to limit policy overrides, credential exposure and private data leakage.
DGX Station for Windows Workflows
DGX Station for Windows is built to run enterprise AI workflows on local Windows systems, from autonomous agent deployment and frontier model development to high-throughput inference, data science and physical AI. It combines NVIDIA AI hardware with Microsoft and Windows enterprise management tools.
- AI Agents: Build and run multiple frontier agents in parallel, and connect them to enterprise applications and workflows.
- AI Development: Pretrain, fine-tune and iterate on large AI models within a Windows environment, with access to Linux AI toolchains through Windows Subsystem for Linux.
- Data Science: Ingest large datasets into coherent memory, reducing data movement bottlenecks across data preparation, machine learning and analytics workflows.
- AI Inference: Run high-throughput inference workloads on AI models.
- Physical AI: Pair the GB300 Superchip with an additional NVIDIA RTX PRO Blackwell Workstation GPU to combine AI compute with ray-traced visualization and simulation in a single deskside system. The configuration supports agents working across virtual-to-physical environments.
DGX Station for Windows can operate as an AI supercomputer or as a shared local compute node. Workloads can scale to GB300 systems in the data center or cloud.
Availability
DGX Station for Windows is expected to be available from ASUS, Dell Technologies, GIGABYTE, HP, MSI and Supermicro, coming in Q4 this year.
Source: NVIDIA
About NVIDIA
NVIDIA, founded in 1993 and headquartered in Santa Clara, CA, designs and manufactures graphics processing units, systems on chips, networking hardware, and AI intelligence software such as CUDA. Its products serve industries including gaming, data centers, autonomous vehicles, professional visualization, robotics, health care, and energy. The company introduced the GPU in 1999 and later expanded into accelerated computing and AI infrastructure. In gaming, its GPUs support high-performance rendering, while in AI and high-performance computing, its systems provide the infrastructure for training and deploying large-scale models. NVIDIA also develops tools for robotics and autonomous driving.