← Back to Glossary
What Is DGX?_
DGX is NVIDIA's line of fully integrated GPU server systems designed for AI workloads. DGX systems (DGX A100, DGX H100, DGX GH200, DGX GB200) come as complete, validated platforms with GPUs, CPU hosts, networking, storage, and software pre-installed. DGX systems are the reference platform for NVIDIA's AI stack and are often deployed in clusters.
Technical Details
DGX systems are NVIDIA's turnkey AI computing platforms. Each generation builds on the corresponding GPU architecture: DGX A100 (8x A100 GPUs), DGX H100 (8x H100 GPUs), DGX GH200 (Grace Hopper Superchips), and DGX GB200 (part of the NVL72 rack system). DGX systems come with validated hardware configurations, pre-installed NVIDIA AI Enterprise software stack, DGX OS (customized Ubuntu), and NVIDIA support contracts. From an infrastructure perspective, DGX systems have the same physical requirements as their HGX counterparts: power, cooling, network, and rack space. The value proposition is the validated, supported configuration that reduces deployment risk.
How Leviathan Systems Works with DGX
Leviathan Systems deploys DGX systems as part of our GPU rack assembly services, handling physical installation, cabling, and testing while the DGX software configuration is managed by NVIDIA or the end customer.
Related service
GPU Rack Assembly & Rack-and-Stack →Related Terms
Appears In
CUDA Fabric Memory: Allocation Handles, Local Maps and Device AccessWhy Modular Data Centers Can Change the Delivery PlanPartial GPU Deliveries: Install in Phases or Wait for the Complete Kit?NVIDIA-Certified Servers or a Complete Cluster: What Does the Badge Cover?Raw, Protected and Client-Usable Storage: Compare Capacity in a GPU OfferSelf-Encrypting GPU Storage Is Not Evidence of SanitizationGPU Server Substitution After Design Freeze: What to RecheckGPU Server Software Is Preinstalled. What Deployment Work Are You Buying?Who Provides the Server Lift? Equipment-Only and GPU Installation QuotesA GPU Power Cap Is Not a Server Input Budget: Compare the Controlled BoundaryAfter GPU Installation: Dispose of Packaging or Preserve It for Returns?GPU Expansion: Rack, Row or Scalable Unit?Future GPU Expansion: Install the Fiber Now or Add It Later?GPU Deployment Spares: Separate Installed Redundancy from Replacement StockGPU Delivery Damage or a Product Fault? Choose the Right Follow-UpShip GPU Equipment Direct to Site or Consolidate Deliveries First?GPU Cluster Management Handoff: What the Customer Needs to RetainChiplets, 2.5D and 3D Packaging: What Is Inside One AI Accelerator?GPU Server Rail Kits: Rack Fit and Service AccessField Discrepancy or Scope Change? Record the DifferenceGPU Storage Network Handoff: Trace the Physical ConnectionsGPU Management Cabling: Records for BMC and Host ConnectionsData Center Drawing Releases: Keep Revision and Workfront TogetherCustomer-Furnished Materials: An Installation Readiness ChecklistCable Schedule vs BOM: Reconcile Links, Parts and QuantitiesWhite-Space Planning for Liquid-Cooled GPU HallsWarm-Water Cooling for GPU Clusters: Why Hotter Water WinsGPU Breakout Cabling: Map the Cage, Logical Port and Cable LegThermal Soak & Burn-In During Commissioning: What to WatchGPU Rack Handling Plans: Review the Provider's Scope and ReleaseGPU Build RFP and SOW Checklist: Compare the Physical Work PackageReceiving and Staging GPU Equipment: A Custody ChecklistRaised Floor vs Slab for High-Density GPU HallsThe Power-On Walkdown: Evidence Before Electrical ReleaseGPU Hall Power-On Readiness: Who Can Authorize Energization?Scaling from Pilot to Production GPU Cluster: What Breaks and What to PlanMEP Coordination for AI Data Halls: Where Mechanical, Electrical, and IT CollideManifolds & Quick-Disconnects in DLC: Scope and OEM BoundariesLeaf-Spine Cabling: Translate a Topology into Traceable Physical LinksHow to Choose a GPU Deployment Partner: Evidence to RequestH100 to H200 Refresh: Prove Compatibility Before Reusing the RackH100 Installation Checklist: Prepare the Evidence Before Rack WorkGPU Site Survey Checklist: Evidence Before InstallationGPU Data Center Deployment in Phoenix, Arizona: Hiring the Build CrewGPU Data Center Deployment in Ohio: Staffing the Physical BuildGPU Data Center Deployment in Northern Virginia: Who Does It and How to HireGB200 NVL72 Deployment Deep Dive: Liquid Loop, Busbar, and SpineGPU Rack Floor Loading: Read the Rating Before Comparing the WeightGPU De-Racking and Decommissioning: Define the Releases Before RemovalCondensation & Dew-Point Control in Liquid-Cooled GPU HallsB300 Deployment Planning: Identify the System and Its Power OptionB200 Deployment Storage: What a DGX B200 System IncludesGPU Installation Handoff: Records, Revisions and Open Items415V vs 480V Distribution in AI Halls: The High-Density Power DecisionFiber Labeling at Scale: Reconcile IDs Across Halls and FabricsCable Pathways for GPU Rooms: Turn the Route into an Installation InputB200 vs GB200: GPU, HGX, DGX and NVL72 Names ExplainedAMD MI300X vs NVIDIA H100: Compare the Server OptionsNVIDIA GB300 NVL72 Explained: Specs, Power, and What It Takes to DeployGPU Data Center Migration: Plan the Physical HandoffNVIDIA H100 vs H200 vs B200: Compare the Deployment ConfigurationGoogle TPU vs NVIDIA GPU: Compare the Workload and Delivery ModelDirect-to-Chip vs Immersion Liquid Cooling for GPU Data CentersData Center Rack-and-Stack Services for GPU Builds: What's IncludedLiquid Cooling vs Air Cooling in the Data Center: When the Crossover HappensHow to Choose a Data Center Liquid Cooling CompanyWho Deploys GB200 and GB300 NVL72? Assign the Work Before Choosing the TeamWhy Data Center Projects Take Time: Facility Readiness and GPU DeploymentGPU Data Center Deployment in Texas: Who Does It and How to HireGPU Commissioning & Acceptance: What to Demand Before You Sign OffGPU Site Readiness: Inputs and Owners Before Equipment ArrivesRail-Optimized vs Fat-Tree: What the Names Actually SpecifyIn-House or Outsourced GPU Installation: A Work-Package DecisionGPU Rack Assembly Cost Drivers: Build a Comparable ScopeHGX vs DGX Deployment: Start with the Purchased SystemPre-Power Inspection: The Walkdown Before Energizing a GPU HallNVLink Cable Cartridges: Identify the Assembly Before Assigning WorkGB300 NVL72 Deployment: Power, Cooling, and the Cable PlantDGX GB200 vs DGX GB300: NVL72 Network InterfacesGPU Rack Receiving and Movement: Prepare the Review PackageThermal Design Updates During a GPU Refresh: Air and Liquid InterfacesCable Labels and As-Built Records: A Practical GPU Project SchemeLiquid-Cooling Commissioning Versus Physical IT Installation