# NVIDIA DGX Station for Windows The World’s Most Powerful Deskside AI Supercomputer for Windows, powered by NVIDIA GB300. ## Coming in Q4 Sign up to be notified on availability. [Notify Me](#notify-me) Overview ## A Trillion-Parameter AI Supercomputer on Every Enterprise Desk NVIDIA DGX Station for Windows is the ultimate deskside AI supercomputer to build and run powerful [AI agents](https://www.nvidia.com/en-us/glossary/ai-agents.md) connected to enterprise applications and workflows. Powered by the NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip with up to 748 GB of coherent memory and 20 petaFLOPS of FP4 AI compute, it delivers data-center-level performance, supporting large AI models up to 1 trillion parameters. Configurable with an additional [NVIDIA RTX PRO™ 6000](https://www.nvidia.com/en-us/products/workstations/professional-desktop-gpus/rtx-pro-6000.md) Workstation GPU, DGX Station for Windows pairs frontier AI compute with ray-traced visualization for running agents across productivity, creative, design, and [physical AI](https://www.nvidia.com/en-us/glossary/generative-physical-ai.md) applications. With the NVIDIA AI software stack and WSL support, enterprises can deploy within existing Windows infrastructure and build and run frontier AI from day one. ### Announcing NVIDIA DGX Station for Windows NVIDIA DGX Station for Windows puts a trillion-parameter AI supercomputer on every enterprise desk to build, run, and connect always-on, frontier AI agents to enterprise applications and workflows. [Read the Press Release](https://nvidianews.nvidia.com/news/nvidia-rtx-station-with-windows-puts-a-trillion-parameter-ai-supercomputer-on-every-enterprise-desk) Highlights ## Data Center Performance at Your Deskside ### AI Compute 20 PFLOPS1 (Up to) ### Large Coherent Memory 748GB (Up to) ### Model Capacity 1T(Up to)2 1. FP4 Tensor Core with Sparsity. 2. Up to 1 trillion parameter AI models. Features ## Engineered to Build and Run Frontier Agentic AI ### NVIDIA GB300 Grace Blackwell Desktop Superchip The GB300 Grace Blackwell Ultra Desktop superchip connects the NVIDIA Blackwell Ultra GPU to a high-performance 72-core NVIDIA Grace CPU using NVIDIA NVLink®-C2C interconnect to deliver best-in-class system communication and performance. Build and run AI locally on GB300 and seamlessly scale out to GB300 in the cloud and data center. ### NVIDIA RTX PRO GPU For Visualization and Physical AI DGX Station can be configured with up to one additional NVIDIA RTX PRO™ Blackwell Generation GPU in addition to the GB300 Superchip. This combination gives developers the ability to pair data center-grade AI compute capability with powerful, ray-traced visualization and simulation. ### Windows Operating System Windows runs on DGX Station, so enterprise developers can access GB300 AI compute right where they already work. IT teams deploy within existing Windows infrastructure—no new group policies, no OS changes, no disruption. Develop and deploy powerful, long-running AI agents with frontier intelligence connected to the Windows ecosystem. ### NVIDIA ConnectX®-8 SuperNIC™ The NVIDIA ConnectX®-8 SuperNIC™ supercharges hyperscale AI computing workloads. With up to 800 gigabits per second (Gb/s), ConnectX-8 SuperNIC delivers extremely fast, efficient network connectivity and supports linking multiple DGX Stations to scale model capacity and performance even further. ### NVIDIA AI Software Stack DGX Station supports an NVIDIA AI software stack on Windows. This full-stack environment provides a seamless foundation for agentic AI development, model fine-tuning, inference, and data science workloads. ### Large Coherent Memory DGX Station features up to 748 GB of coherent memory for up to 1 trillion-parameter models to be run efficiently within one memory pool, thanks to the C2C superchip interconnect that bypasses the bottlenecks of traditional CPU and GPU systems. ### IT Manageability DGX Station is built for enterprise deployment. Out-of-band telemetry via the BMC, and Enterprise Manageability gives IT teams full visibility, monitoring, and remote management capabilities. Hardware root of trust and enterprise secure boot help keep every system in your fleet protected and compliant. Workloads ## Build and Run AI on Windows #### Build and Run AI Agents on Windows Build and deploy powerful autonomous AI agents connected to Windows apps. DGX Station for Windows lets you develop and run powerful AI agents that connect to enterprise applications and workflows, including productivity, design, and engineering tools. These frontier agents are powered by GB300 on DGX Station, offloading costs with local token generation. #### AI Inference Serve powerful AI inference locally DGX Station for Windows accelerates local [inference](https://www.nvidia.com/en-us/glossary/ai-inference.md) for large AI models—delivering fast token generation, data analysis, and agentic AI on Windows. #### Data Science Accelerate enterprise data science workflows. Ingest massive data lakes with up to 748 GB of coherent memory to accelerate end-to-end data science workflows. #### Frontier AI Development Develop and run AI models up to 1 trillion parameters locally. With up to 748 GB of coherent memory and up to 20 petaFLOPS of FP4 compute, DGX Station delivers the memory and performance to develop, fine-tune, and run frontier AI models and agents at your deskside. #### Physical AI Pair GB300 AI compute with RTX PRO for physical AI. Combine the GB300 Superchip with the NVIDIA RTX PRO 6000 for best-in-class AI training, ray-traced simulation, and visualization for building and optimizing autonomous AI systems and physical AI applications. #### Windows Subsystem for Linux (WSL) for AI Toolchains Run Linux AI frameworks on Windows via WSL. WSL on DGX Station gives developers access to Linux AI toolchains on Windows. ## DGX Station with Ubuntu and NVIDIA AI Developer Tools DGX Station featuring Ubuntu with NVIDIA AI Developer tools lets developers, researchers, and data scientists rapidly develop, fine-tune, and inference with large AI models and long-running AI agents locally and seamlessly deploy to the data center or cloud. [Learn More](https://www.nvidia.com/en-us/products/workstations/dgx-station.md) Specifications ## Technical Specifications | | | | --- | --- | | NVIDIA GPU | 1x NVIDIA Blackwell Ultra | | NVIDIA CPU | 1x Grace 72-Core Neoverse V2 | | GPU Memory | 252 GB HBM3e | 7.1 TB/s | | CPU Memory | 496 GB LPDDR5X | 396 GB/s | | NVLink-C2C | 900 GB/s | | Networking | Peak Bandwidth | NVIDIA ConnectX-8 SuperNIC | Up to 800 Gb/s | Ethernet | | Supported OS | Microsoft Windows | | FP4 Tensor Core | 20 | 153 PFLOPS | | FP8 / FP6 Tensor Core | 10 PFLOPS | | Int8 Tensor Core | 330 TOPS | | FP16/BF16 Tensor Core | 5 PFLOPS | | TF32 Tensor Core | 2.5 PFLOPS | | FP32 | 80 TFLOPS | | FP64 / FP64 Tensor Core | 1.3 TFLOPS | | Total System Power | 1,600 W | | Decoders | 7 NVDEC 7 nvJPEG | | Supported RTX PRO GPU4 | NVIDIA RTX PRO 6000 Workstation Edition RTX PRO 6000 Blackwell Max-Q Workstation Edition RTX PRO 4000 Blackwell SFF Edition RTX PRO 2000 Blackwell | | Storage | 4x M.2 Gen 5 Slots | | Ethernet Ports | 2x QSFP112 (400 Gbs per port) 1x RJ45 10GbE 1x RJ45 1GbE (BMC for system management) | | PCIe Slots | 1x PCIe Gen 5 x16 2x PCIe Gen 5 x16 (x8 electrical) | | Additional Ports and Connectors4 | Front IO: 2x USB Type C (USB3.1), 2x USB Type A (USB3.1), Audio Rear IO: 4x USB Type A, USB Micro-B (BMC) mDP5 (BMC), Audio | | 1. Peak rates are based on GPU boost clock. 2. All Tensor Core specifications are with sparsity unless otherwise noted. 3. Without sparsity. 4. Support may vary by system. Check with your OEM to confirm specific support and configurations. 5. Mini DisplayPort is intended only for system management. | | Partners ## Built by Global OEM Partners [Notify Me](#notify-me) ## Email Me When Available Welcome back. Not you? Log Out Welcome back. Not you? Clear form