Extended Reality
Extended Reality and Spatial Computing: The Interface for Physical and Agentic AI
Extended reality (XR) is an umbrella term for immersive technologies such as virtual reality (VR), augmented reality (AR), and mixed reality (MR) that blend the physical and digital worlds, while spatial computing is the broader paradigm in which digital intelligence understands and interacts with the physical world in real time.
NVIDIA’s comprehensive offerings in this space bridge the gap between heavy computational requirements and lightweight mobile devices through a suite of advanced hardware and software solutions. By centralizing the heavy lifting on NVIDIA RTX™-powered servers and workstations, NVIDIA allows you to bypass the processing limits of mobile hardware and stream complex, photorealistic environments with zero model decimation. These technologies leverage the dedicated ray-tracing and AI cores of the RTX architecture to power full-fidelity XR and spatial computing experiences—delivering unmatched visual realism and physically accurate simulations to any untethered device.
Explore XR Technologies
NVIDIA CloudXR
Stream high-fidelity XR content from RTX-powered servers to an open ecosystem of devices. Built as a universal OpenXR bridge, NVIDIA CloudXR™ 6.0 delivers native visionOS integration and frictionless web-based access via CloudXR.js.
NVIDIA Omniverse
NVIDIA Omniverse™ is a collection of libraries and microservices for developing industrial digital twins and physical AI simulation applications. Stream complex OpenUSD stages directly to Apple Vision Pro with full RTX visual fidelity.
NVIDIA Professional VR Ready Solutions
NVIDIA VR Ready systems deliver unparalleled performance for smooth, immersive virtual reality experiences.
NVIDIA VR Capture and Replay
NVIDIA Virtual Reality Capture and Replay (VCR) enables developers to accurately capture and replay VR content for performance testing, scene quality control, and more.
NVIDIA VRWorks
NVIDIA VRWorks™ enables VR application and headset developers to create amazing virtual reality experiences, improving VR performance and visual quality.
NVIDIA Warp
Warp enables Python developers to create GPU-accelerated, 3D simulation workflows that drive ML pipelines in PyTorch, JAX, PhysicsNeMo, and NVIDIA Omniverse.
NVIDIA XR AI
NVIDIA XR AI is a framework that connects XR devices—such as lightweight AR/AI glasses or head-mounted displays—to your organization’s full computational power, enabling spatially aware, intelligent agents to operate seamlessly across cloud, data center, workstation, and edge deployments.
XR Learning Resources
- FeaturedSessionCosmos
An Introduction to NVIDIA Cosmos for Physical AI
In this introductory, hands-on course, explore NVIDIA Cosmos, a platform of generative WFMs, advanced tokenizers, guardrails, and an accelerated data processing and curation pipeline.
- FeaturedBlogCosmos|DRIVE|DRIVE AV|Omniverse
AV Simulation with Neural Reconstruction and Cosmos
Create realistic, diverse training data from real-world environments with AV simulation using neural reconstructions and Cosmos models.
- FeaturedSessionCosmos
Beyond Rendering: Why Graphics is Needed For Physical AI
Machines need more than perception—they need understanding. NVIDIA Cosmos WFMs help developers predict, transfer knowledge, and reason for autonomous systems.
- FeaturedGuideCosmos|Isaac Sim
Cosmos Cookbook
Read the Cosmos Cookbook for an open-source guide offering workflows and examples to deploy and post-train NVIDIA Cosmos world foundation models for real applications.
- FeaturedModelCosmos
Demo Cosmos Reason Model on NVIDIA Build
NVIDIA Cosmos Reason is a 7B-parameter vision language model that lets robots and AI agents reason and plan actions using knowledge, physics, and common sense.
- FeaturedDemoCosmos|Isaac Sim
Synthetic Data From Neural Simulation for Robot Training
This blueprint helps overcome the data challenge for training robot policy models like GR00T N - and enables generalist robots reason, plan, and execute tasks.