NVIDIA announced a major shift in enterprise AI infrastructure: bringing supercomputer-class computing directly to Windows desktops and laptops, enabling businesses to run powerful AI agents locally without relying on cloud services. The company unveiled two flagship products at GTC Taipei 2026 that fundamentally reshape how organizations deploy artificial intelligence, from data centers to individual workstations.

What Is the DGX Station for Windows, and Why Does It Matter?

The DGX Station for Windows is the world's most powerful deskside AI supercomputer, built on the NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip. It can run frontier AI models with up to 1 trillion parameters locally, meaning enterprises no longer need to send sensitive data to cloud services or wait for responses from remote servers. This represents a fundamental change in how enterprise AI workflows operate.

The system features up to 748 gigabytes of coherent memory and delivers up to 20 petaflops of FP4 (fourth-precision floating-point) performance, a technical measure of raw computing speed. It connects a powerful Blackwell Ultra GPU to a 72-core Grace CPU via NVIDIA's NVLink-C2C interconnect, which enables extremely fast communication between the two processors.

"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. DGX Station delivers supercomputing-class AI directly into Windows, where millions already design, engineer, research and create every day," said Chris Marriott, Vice President of Enterprise Platforms at NVIDIA.

Chris Marriott, Vice President of Enterprise Platforms at NVIDIA

The collaboration between NVIDIA and Microsoft addresses a long-standing problem in enterprise computing: for decades, heavy-duty AI workloads required powerful Linux-based systems in data centers, while the vast majority of Fortune 500 companies rely on Windows for everyday work. DGX Station for Windows bridges this gap, bringing data center-class AI infrastructure directly into the Windows ecosystem where enterprises already operate.

How Does NVIDIA's RTX Spark Transform Personal Computers Into AI Machines?

Beyond enterprise desktops, NVIDIA introduced RTX Spark, a superchip designed for consumer laptops and small desktop PCs. This represents NVIDIA's answer to making personal AI agents practical and secure for everyday users. RTX Spark features 6,144 CUDA cores (the parallel processing units that power NVIDIA's AI acceleration), fifth-generation Tensor Cores with FP4 precision, and up to 128 gigabytes of unified memory.

The chip connects an NVIDIA Blackwell RTX GPU to a custom 20-core Grace CPU via NVLink-C2C, delivering up to 1 petaflop of AI compute performance. MediaTek, a leading ARM-based chip designer, collaborated with NVIDIA on the custom CPU design to optimize power efficiency and connectivity. Microsoft's Surface Laptop Ultra will be the first device equipped with RTX Spark, with additional models from Lenovo, Dell, HP, ASUS, and Acer launching later in 2026.

Jensen Huang, NVIDIA's founder and CEO, framed this shift as a fundamental reimagining of the personal computer itself. "The PC is being reinvented," Huang stated. "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".

What Can These AI Supercomputers Actually Do?

Both the DGX Station for Windows and RTX Spark are designed to handle demanding AI workloads that previously required cloud services or specialized hardware. The capabilities span multiple use cases across enterprise and consumer markets:

• AI Agent Development: Build and run multiple frontier AI agents in parallel, connecting them directly to enterprise applications and workflows without sending data to external servers.
• Large Language Model Inference: Run language models with up to 120 billion parameters and context windows of up to 1 million tokens locally, enabling real-time responses without cloud latency.
• 3D Rendering and Video Editing: Render ultralarge 3D scenes exceeding 90 gigabytes in size and edit 12K 4:2:2 video with NVIDIA's DLSS technology, which uses AI to enhance visual quality.
• High-Performance Gaming: Play AAA (major studio) games at 1440p resolution with over 100 frames per second while using ray tracing, DLSS, and Reflex technologies for enhanced visuals and responsiveness.
• Data Science and Analytics: Ingest large datasets into up to 748 gigabytes of coherent memory, eliminating data movement bottlenecks that slow down machine learning workflows.
• Physical AI and Simulation: Pair the GB300 Superchip with an additional RTX PRO Blackwell Workstation GPU to combine frontier AI compute with ray-traced visualization for agents that perceive and interact in virtual-to-physical environments.

How Do NVIDIA and Microsoft Address Security and Privacy Concerns?

Running powerful AI agents locally raises critical questions about security and privacy. NVIDIA and Microsoft developed a comprehensive security framework to address these concerns. The solution combines new Windows security primitives with NVIDIA OpenShell, an open-source runtime specifically designed for autonomous agents.

NVIDIA OpenShell creates an individual, isolated sandbox for each agent, separating application-layer operations from infrastructure-layer policy enforcement. This means security and privacy policies are enforced at the system level, preventing agents from overriding them or leaking credentials and private data. Rather than relying on behavioral system prompts that agents might circumvent, OpenShell uses Windows security primitives to enforce constraints on the environment itself.

The new Windows security infrastructure delivers identity, containment, policy, and end-to-end security capabilities to build and run agents natively. NVIDIA OpenShell provides additional policy capabilities, allowing users to define what agents can and cannot do, intelligently route queries to local models based on privacy preferences, and disguise personal information in queries sent to cloud models.

"We are strong supporters of deploying agents like OpenClaw securely into the Windows ecosystem. Running solutions like OpenShell and the Microsoft security primitives on RTX Spark will enable users to leverage a fully integrated stack for private, personal agents running on device," stated Vincent Koc, Chief Architect at the OpenClaw Foundation.

Vincent Koc, Chief Architect at the OpenClaw Foundation

What Does This Mean for Enterprise IT and Software Developers?

For enterprise IT teams, DGX Station for Windows brings a secure, managed platform to GB300 deployments, extending the same Windows security, compliance, and fleet management infrastructure organizations already rely on. AI agents deploy and operate within this managed environment, governed through familiar Microsoft tools. Linux workloads receive the same level of manageability support through Windows Subsystem for Linux, helping organizations maintain security, compliance, and operational readiness across their fleets.

Software developers and creative professionals are embracing the RTX Spark platform. Over 100 Windows software providers, including Adobe, Blackmagic Design, Blender, CapCut, and ComfyUI, are optimizing their applications for RTX Spark. Adobe is re-engineering Photoshop and Premiere specifically for the platform, claiming up to a 2-fold increase in AI and graphics performance. Game developers such as KRAFTON, NetEase, Remedy Entertainment, Riot Games, and XBOX are also embracing the new RTX Spark platform.

The availability timeline shows DGX Station for Windows is expected to arrive in Q4 2026, with manufacturers including ASUS, Dell Technologies, GIGABYTE, HP, and MSI preparing to offer the system. RTX Spark-equipped devices are launching throughout 2026 from major PC manufacturers.

"Our goal is to deliver unmetered intelligence to every home and every desk with Windows," said Satya Nadella, Chairman and CEO of Microsoft.

Satya Nadella, Chairman and CEO of Microsoft

This announcement represents a significant expansion of NVIDIA's influence beyond GPU manufacturing. By integrating CPUs, GPUs, networking, storage, cooling systems, security, and Windows integration into a unified narrative around "Agent AI," NVIDIA is attempting to define what data centers and personal computers should look like in the AI era. The shift from cloud-dependent AI to local, on-device inference marks a fundamental change in how enterprises and consumers will interact with artificial intelligence in the coming years.