NVIDIA DSX platform brings together reference designs, APIs, software libraries, and technologies to design, simulate, build, and operate AI factories. Explore the features and capabilities for each DSX component, and view linked technical resources.
Select a feature/capability tile to view linked resources, select a section title to jump to its detailed tab below.
## AI-Readable DSX Landing Page Architecture And Resource Catalog This catalog is generated from the DSX landing page data files so AI agents can understand the custom React architecture diagram, descriptor mappings, and linked resource cards. Do not edit this block by hand; run `node scripts/generate-landing-ai-catalog.mjs` after changing landing-page descriptors, card data, or diagram groups. ### Architecture Diagram Summary * DSX Sim spans the AI factory lifecycle and represents simulation, digital-twin, SimReady asset, and validated-integration capabilities used to design, validate, build, and operate an AI factory before and after physical deployment. * DSX OS is the operating layer. It contains core Platform Software and Infrastructure Software bands, and it also provides technologies for resiliency, observability, compliance, and security outside those core software bands. * Platform Software covers intelligent scheduling, service orchestration, workload orchestration, and disaggregated inference capabilities for AI services. * Infrastructure Software covers provisioning, lifecycle management, configuration validation, multi-tenant operations, network orchestration, health monitoring, and remediation capabilities for AI factory infrastructure. * DSX MaxLPS leverages software technologies in DSX OS to interface with the DSX Hardware Reference Design and apply dynamic power policies, power allocation, 45°C inlet operation, and advanced performance-per-watt optimizations to AI factory hardware. * DSX Exchange is the IT/OT communications hub. In the diagram it connects with DSX OS, DSX MaxLPS, DSX Flex, Shell, and Power to coordinate compute, network, energy, power, and cooling plant signals. * DSX Flex connects AI factory operations with power-grid signals. It represents flexible power orchestration across utility power, on-site renewable generation, and energy storage. * The diagram includes two reference design sections: DSX Hardware Reference Design and DSX Facilities Infrastructure Reference Design. * The DSX Hardware Reference Design section contains Compute, Networking, and Storage capabilities for validated hardware architecture. * The DSX Facilities Infrastructure Reference Design section describes the site and facility infrastructure required for AI factories, including Land, Connectivity, Shell (Building & Cooling), and Power capabilities for physical site planning, cooling, electrical, and grid-supporting infrastructure. * The Grid Operator node connects to DSX Flex and facilities Power, representing external grid signals such as demand response, load shedding, and pricing events. ### Diagram Sections And Feature Tiles * DSX Sim: High-Fidelity Logical Simulation; Digital Twins; SimReady Assets; Validated Integrations. DSX Sim is the full-lifecycle simulation section of the architecture diagram. * DSX OS - Platform Software: Intelligent Scheduling; Service Orchestration; Workload Orchestration; Disaggregated Inference. Platform Software is one of the core DSX OS software bands. * DSX OS - Infrastructure Software: Provisioning; Lifecycle Management; Configuration Validation; Multi-Tenant Operations; Network Orchestration; Health Monitoring; Remediation. Infrastructure Software is one of the core DSX OS software bands. * DSX OS - Support Functions: Resiliency; Observability; Compliance; Security. These support functions are part of DSX OS and sit outside the core Platform Software and Infrastructure Software bands. * DSX MaxLPS: 45°C Inlet; Dynamic Power Allocation; Advanced Perf/Watt Techniques. DSX MaxLPS spans power-aware optimization capabilities that connect DSX OS with hardware and power operations. * DSX Exchange: IT/OT Communications Hub. DSX Exchange is the communications hub that connects IT and OT signals across compute, network, energy, power, and cooling systems. * DSX Flex: Flexible Power Orchestration. DSX Flex represents grid-aware and renewable or hybrid power orchestration capabilities. * DSX Hardware Reference Design: Compute; Networking; Storage. This is a reference design section in the diagram and contains the hardware architecture capabilities for compute, networking, and storage. * DSX Facilities Infrastructure Reference Design: Land; Connectivity; Shell; Power. This is a reference design section in the diagram and contains the facilities infrastructure capabilities for land, connectivity, shell, and power that support AI factory site planning, facility design, cooling, electrical infrastructure, and grid integration. ### Resource Cards By Landing Section #### DSX Sim A suite of simulation technologies that together enable partners to design, validate, and operate AI factories before and after physical deployment. * [NVIDIA DSX Air](https://docs.nvidia.com/networking-ethernet-software/nvidia-air/): High-fidelity logical simulation of NVIDIA hardware and software infrastructure — GPUs, SuperNICs, DPUs, switches — plus validated integrations with storage, security, and orchestration partners. * Feature/capability tags: High-Fidelity Logical Simulation, Validated Integrations. * [NVIDIA Omniverse DSX Blueprint for AI Factories](https://docs.omniverse.nvidia.com/dsx/latest/index.html): An end-to-end framework to design, simulate, build, and operate gigawatt-scale AI factory digital twins. * Feature/capability tags: Digital Twins. * Related links: [Browse the DSX Blueprint GitHub repo](https://github.com/NVIDIA-Omniverse-blueprints/omniverse-dsx-blueprint-for-ai-factories) or get started on [build.nvidia.com](https://build.nvidia.com/nvidia/omniverse-dsx-blueprint-for-ai-factories). * [AI Factory Digital Twin Pipeline Samples](https://nvidia-omniverse.github.io/aif-pipeline-samples/): Sample scripts and presets for creating DSX SimReady USD assets, covering CAD ingestion, optimization, validation, and metadata workflows for digital twin and AI factory applications. * Feature/capability tags: SimReady Assets. * Related links: [Browse the AIF Pipeline Samples GitHub repo](https://github.com/NVIDIA-Omniverse/aif-pipeline-samples). #### DSX OS Open-source, modular software for building, operating, and scaling AI factory infrastructure, with composable components for lifecycle management, runtime consistency, health automation, resiliency, multi-tenant operations, AI platform services, and more. * [NVIDIA Run:ai](https://run-ai-docs.nvidia.com/): Kubernetes-native platform for AI workload and GPU orchestration, designed to maximize accelerated infrastructure utilization. * Feature/capability tags: Intelligent Scheduling, Workload Orchestration. * [NVIDIA Cloud Functions (NVCF)](https://docs.nvidia.com/nvcf): Unified API layer for scaling inference and simulation workloads across one or more Kubernetes clusters. * Feature/capability tags: Service Orchestration, Workload Orchestration. * Related links: [Browse the NVCF GitHub repo](https://github.com/NVIDIA/nvcf). * [KAI Scheduler](https://github.com/NVIDIA/KAI-Scheduler): Open source Kubernetes-native scheduler for AI workloads with topology-aware placement and resource allocation. * Feature/capability tags: Intelligent Scheduling. * [Grove](https://github.com/ai-dynamo/grove): Modular component of NVIDIA Dynamo — provides a Kubernetes API for defining and scaling multi-component AI inference workloads. * Feature/capability tags: Disaggregated Inference. * [Dynamo](https://docs.nvidia.com/dynamo): Distributed inference-serving framework built to deploy models in multi-node environments. * Feature/capability tags: Disaggregated Inference, Advanced Perf/Watt Techniques. * Related links: [Browse the Dynamo GitHub repo](https://github.com/ai-dynamo/dynamo). * [NVIDIA Fleet Intelligence](https://docs.nvidia.com/fleet-intelligence/latest/index.html): Agent-based managed service offering real-time insights into GPU fleet health and integrity. * Feature/capability tags: Health Monitoring. * Related links: [Browse the Fleet Intelligence Agent GitHub repo](https://github.com/NVIDIA/Fleet-Intelligence-Agent). * [NVIDIA Infra Controller (NICo™)](https://docs.nvidia.com/infra-controller): Bare-metal provisioning and secure lifecycle management for multi-tenant GPU infrastructure. * Feature/capability tags: Provisioning, Lifecycle Management, Multi-Tenant Operations. * Related links: [Browse the NICo GitHub repo](https://github.com/NVIDIA/infra-controller-core). * [NVIDIA Switch Infrastructure - Config Manager](https://docs.nvidia.com/switch-infrastructure/config-manager): Open-source network automation and configuration management platform for large-scale datacenter operations. * Feature/capability tags: Network Orchestration. * Related links: [Browse the Config Manager GitHub repo](https://github.com/nvidia/nv-config-manager). * [AI Cluster Runtime](https://docs.nvidia.com/aicr): Canonical, continuously validated definition of the NVIDIA-accelerated Kubernetes runtime. * Feature/capability tags: Configuration Validation. * Related links: [Browse the AICR GitHub repo](https://github.com/NVIDIA/aicr). * [NVSentinel](https://docs.nvidia.com/nvsentinel): Open-source, Kubernetes-native GPU monitoring and fault remediation. * Feature/capability tags: Remediation. * Related links: [Browse the NVSentinel GitHub repo](https://github.com/NVIDIA/NVSentinel). * [NVIDIA DOCA Platform Framework (DPF)](https://github.com/NVIDIA/doca-platform): Orchestration system to build, deploy, and operate BlueField-accelerated infrastructure services. * Feature/capability tags: Network Orchestration, Multi-Tenant Operations. #### DSX MaxLPS & Flex DSX MaxLPS is a suite of technologies to maximize AI factory compute throughput and tokens per watt within a fixed power budget, by applying intelligent optimizations and dynamically enforcing power policies at GPU, rack, and workload level. DSX Flex enables renewable and hybrid power orchestration across utility, on-site renewables, and storage, receiving grid signals (load shedding, demand response, pricing events) and adapting AI workloads dynamically. * [NVIDIA Domain Power Service](https://docs.nvidia.com/datacenter/dps): Dynamic power allocation software that continuously monitors GPU and rack-level power consumption, reallocating it where needed based on defined power budgets to unlock stranded capacity and optimize overall utilization. * Feature/capability tags: Dynamic Power Allocation. * [Power Management for NVIDIA Vera Rubin Data Center Systems](https://partners.nvidia.com/DocumentDetails?DocID=1153760): (NVOnline #1153760) * Access: NVOnline access required. * Feature/capability tags: Advanced Perf/Watt Techniques. * [Workload Power Profiles](https://developer.nvidia.com/blog/optimize-data-center-efficiency-for-ai-and-hpc-workloads-with-power-profiles/): Learn more about workload-aware optimized power profiles for improving AI factory energy efficiency and performance. * Feature/capability tags: Advanced Perf/Watt Techniques. * [NvGrid](https://docs.nvidia.com/datacenter/dps/versions/0.8/guides/concepts/nvgrid/): Provides grid integration capabilities that enable AI factory power management in response to utility grid signals. * Feature/capability tags: Flexible Power Orchestration. #### DSX Exchange The IT/OT integration and communications hub, coordinating compute, network, energy, power, and cooling plant signals in the AI factory. * [DSX Event Bus](https://docs.nvidia.com/dsx-exchange): NATS-based event bus for AI factory communications and operations. * Feature/capability tags: IT/OT Communications Hub. * Related links: [Browse the DSX Exchange GitHub repo](https://github.com/nvidia/dsx-exchange). * [AsyncAPI Schemas](https://docs.nvidia.com/dsx-exchange/schema): Event bus schemas for DSX data exchange services. * Feature/capability tags: IT/OT Communications Hub. * Related links: [Browse the AsyncAPI Schemas GitHub repo](https://github.com/NVIDIA/dsx-exchange/tree/main/schemas). * [BMS Companion Guide](https://docs.nvidia.com/dsx-exchange/bms-integration): This guide is for System Integrators and BMS contractors who will configure a Building Management System to publish facility data to the DSX Event Bus. It covers the concepts, topic structure, metadata fields, and implementation steps needed to complete a BMS integration. * Feature/capability tags: IT/OT Communications Hub, Shell. #### DSX Reference Designs Generation-specific, validated AI factory architectures covering compute, networking, storage, facilities infrastructure, and hardware cluster design. * [NVIDIA Vera Rubin NVL72 Reference Design](https://partners.nvidia.com/DocumentDetails?DocID=1151654): (NVOnline #1151654) * Access: NVOnline access required. * Feature/capability tags: Compute, Networking, Storage. * [NVIDIA DSX - Vera Rubin Facilities Infrastructure Reference Design](https://partners.nvidia.com/DocumentDetails?DocID=1145739): (NVOnline #1145739) * Access: NVOnline access required. * Feature/capability tags: Land, Connectivity, Shell, Power. * [NVIDIA DSX Facilities Infrastructure Design Guide](https://partners.nvidia.com/DocumentDetails?DocID=1152370): (NVOnline #1152370) * Access: NVOnline access required. * Feature/capability tags: Land, Connectivity, Shell, Power, 45°C Inlet. * [BESS Self-Qualification Guidelines](https://docs.nvidia.com/datacenter/dsx/BESS-Self-Qualification-Guidelines.html): Defines the partner-run qualification process for a Battery Energy Storage System (BESS) intended to support AI load buffering, demand response (DR), and low voltage ride-through (LVRT) in grid-connected and islanded on-site generation modes. * Feature/capability tags: Flexible Power Orchestration, Power. * [NVIDIA DSX AI Factory Marketplace](https://marketplace.nvidia.com/en-us/enterprise/dsx-infrastructure/): Products that have been validated to meet NVIDIA functional requirements for AI factory applications. * Feature/capability tags: Shell. #### NVIDIA Cloud Partner Resources * [NVIDIA Cloud Partner Software Reference Guide](./ncp/software-reference-guide/introduction): Infrastructure-native reference for building AI cloud services — software components, architecture patterns, and deployment guidance. * [NVIDIA Inference Reference Architecture](./ncp/inference-ra): Software architecture to help operators build performant, cost-effective inference solutions on NVIDIA platforms. * [NVIDIA Requirements for AI Clouds](./ncp/nvidia-requirements-for-ai-clouds/introduction): Primary requirements document covering the full stack of AI cloud infrastructure services — SLAs, security, networking, storage, and operations. * [NVIDIA Exemplar Cloud](https://www.nvidia.com/en-us/data-center/ai-cloud-performance/): Improves performance per TCO, security, and reliability for cloud providers with hardware and software recipes, tools, capabilities, and references. * Related links: [Browse the NVIDIA Performance Benchmarking recipes](https://github.com/NVIDIA/dgxc-benchmarking). * [NVIDIA AI Cloud-Ready Validation Initiative](https://www.nvidia.com/en-us/data-center/isv-validation-program/): Qualifies and validates AI infrastructure and platform software for deployment on NVIDIA-accelerated cloud infrastructure. * Related links: [Browse the ISV NCP Validation Suite](https://github.com/NVIDIA/ISV-NCP-Validation-Suite). ### Feature/Capability Tags To Linked Resources * High-Fidelity Logical Simulation: [NVIDIA DSX Air](https://docs.nvidia.com/networking-ethernet-software/nvidia-air/). * Digital Twins: [NVIDIA Omniverse DSX Blueprint for AI Factories](https://docs.omniverse.nvidia.com/dsx/latest/index.html). * SimReady Assets: [AI Factory Digital Twin Pipeline Samples](https://nvidia-omniverse.github.io/aif-pipeline-samples/). * Validated Integrations: [NVIDIA DSX Air](https://docs.nvidia.com/networking-ethernet-software/nvidia-air/). * Intelligent Scheduling: [NVIDIA Run:ai](https://run-ai-docs.nvidia.com/); [KAI Scheduler](https://github.com/NVIDIA/KAI-Scheduler). * Service Orchestration: [NVIDIA Cloud Functions (NVCF)](https://docs.nvidia.com/nvcf). * Workload Orchestration: [NVIDIA Run:ai](https://run-ai-docs.nvidia.com/); [NVIDIA Cloud Functions (NVCF)](https://docs.nvidia.com/nvcf). * Disaggregated Inference: [Grove](https://github.com/ai-dynamo/grove); [Dynamo](https://docs.nvidia.com/dynamo). * Provisioning: [NVIDIA Infra Controller (NICo™)](https://docs.nvidia.com/infra-controller). * Lifecycle Management: [NVIDIA Infra Controller (NICo™)](https://docs.nvidia.com/infra-controller). * Configuration Validation: [AI Cluster Runtime](https://docs.nvidia.com/aicr). * Multi-Tenant Operations: [NVIDIA Infra Controller (NICo™)](https://docs.nvidia.com/infra-controller); [NVIDIA DOCA Platform Framework (DPF)](https://github.com/NVIDIA/doca-platform). * Network Orchestration: [NVIDIA Switch Infrastructure - Config Manager](https://docs.nvidia.com/switch-infrastructure/config-manager); [NVIDIA DOCA Platform Framework (DPF)](https://github.com/NVIDIA/doca-platform). * Health Monitoring: [NVIDIA Fleet Intelligence](https://docs.nvidia.com/fleet-intelligence/latest/index.html). * Remediation: [NVSentinel](https://docs.nvidia.com/nvsentinel). * 45°C Inlet: [NVIDIA DSX Facilities Infrastructure Design Guide](https://partners.nvidia.com/DocumentDetails?DocID=1152370). * Dynamic Power Allocation: [NVIDIA Domain Power Service](https://docs.nvidia.com/datacenter/dps). * Advanced Perf/Watt Techniques: [Dynamo](https://docs.nvidia.com/dynamo); [Power Management for NVIDIA Vera Rubin Data Center Systems](https://partners.nvidia.com/DocumentDetails?DocID=1153760); [Workload Power Profiles](https://developer.nvidia.com/blog/optimize-data-center-efficiency-for-ai-and-hpc-workloads-with-power-profiles/). * IT/OT Communications Hub: [DSX Event Bus](https://docs.nvidia.com/dsx-exchange); [AsyncAPI Schemas](https://docs.nvidia.com/dsx-exchange/schema); [BMS Companion Guide](https://docs.nvidia.com/dsx-exchange/bms-integration). * Flexible Power Orchestration: [NvGrid](https://docs.nvidia.com/datacenter/dps/versions/0.8/guides/concepts/nvgrid/); [BESS Self-Qualification Guidelines](https://docs.nvidia.com/datacenter/dsx/BESS-Self-Qualification-Guidelines.html). * Compute: [NVIDIA Vera Rubin NVL72 Reference Design](https://partners.nvidia.com/DocumentDetails?DocID=1151654). * Networking: [NVIDIA Vera Rubin NVL72 Reference Design](https://partners.nvidia.com/DocumentDetails?DocID=1151654). * Storage: [NVIDIA Vera Rubin NVL72 Reference Design](https://partners.nvidia.com/DocumentDetails?DocID=1151654). * Land: [NVIDIA DSX - Vera Rubin Facilities Infrastructure Reference Design](https://partners.nvidia.com/DocumentDetails?DocID=1145739); [NVIDIA DSX Facilities Infrastructure Design Guide](https://partners.nvidia.com/DocumentDetails?DocID=1152370). * Connectivity: [NVIDIA DSX - Vera Rubin Facilities Infrastructure Reference Design](https://partners.nvidia.com/DocumentDetails?DocID=1145739); [NVIDIA DSX Facilities Infrastructure Design Guide](https://partners.nvidia.com/DocumentDetails?DocID=1152370). * Shell: [BMS Companion Guide](https://docs.nvidia.com/dsx-exchange/bms-integration); [NVIDIA DSX - Vera Rubin Facilities Infrastructure Reference Design](https://partners.nvidia.com/DocumentDetails?DocID=1145739); [NVIDIA DSX Facilities Infrastructure Design Guide](https://partners.nvidia.com/DocumentDetails?DocID=1152370); [NVIDIA DSX AI Factory Marketplace](https://marketplace.nvidia.com/en-us/enterprise/dsx-infrastructure/). * Power: [NVIDIA DSX - Vera Rubin Facilities Infrastructure Reference Design](https://partners.nvidia.com/DocumentDetails?DocID=1145739); [NVIDIA DSX Facilities Infrastructure Design Guide](https://partners.nvidia.com/DocumentDetails?DocID=1152370); [BESS Self-Qualification Guidelines](https://docs.nvidia.com/datacenter/dsx/BESS-Self-Qualification-Guidelines.html).