The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →ScaleMatrix’s “no data center” claim described where its 2019 AI system could be deployed—not a machine that ran without data-center-grade power, cooling, or site planning. The system paired NVIDIA DGX-2 compute with NetApp storage and Cisco networking in a purpose-designed 45U cabinet. ScaleMatrix said the equipment occupied about 36U and drew approximately 40kW.
What was ScaleMatrix’s DGXPod?
Announced on March 27, 2019, the configuration combined three vendors’ equipment in one rack-scale system: NVIDIA DGX-2 AI systems for compute, a NetApp A800 All Flash array for storage, and Cisco 3232C 100GB switches for networking. ScaleMatrix called the integrated configuration “#DGXPOD” and described it as NetApp ONTAP AI. ScaleMatrix’s announcement said it was installed in a 45U S-1052 Dynamic Density Control (DDC) cabinet.
| Published figure | What it refers to |
|---|---|
| 45U | The height of the ScaleMatrix S-1052 DDC cabinet in the 2019 announcement. |
| About 36U | The rack space ScaleMatrix said the equipment used. |
| Approximately 40kW | The reported power requirement for the described configuration; the announcement does not specify a measurement method or operating conditions. |
These are ScaleMatrix’s historical figures, not independently measured results. They describe a dense enterprise deployment, not a single self-contained consumer appliance. The announcement did not publish enough detail to establish the configuration’s performance, total cost, or present-day availability.
Did it really need no data center?
No—not if “no data center” is taken to mean no facility infrastructure. ScaleMatrix’s 2019 message was that a conventional data-center room was not the only possible deployment location. The company promoted its own colocation facilities as well as cabinets or enclosures deployed at customer premises. Its GTC 2019 blog described ruggedized S-Series and R-Series enclosures for challenging settings, including outdoor locations, and said a data center was “no longer a requirement.”
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Scan for outdated or missing drivers - takes under a minute3Clear out junk files and repair common Windows errors#1 Best Overall
- Custom Fit Compatibility: Specifically designed rack mount bracket for Nvidia DGX SparkNano, ensuring precise alignment in standard 10 inch rack systems for stable and secure installation.
- Space-Saving Design: Compact 1.5U rack mount profile allows efficient use of limited rack space, ideal for network cabinets, lab setups.
- Mounting Stability: Engineered rack shelf structure provides balanced weight distribution, helping keep equipment level and properly supported during operation.
- Durable Structural: Rack bracket frame construction enhances strength, offering dependable mounting performance.
- Fast Installation: Rackmount holder design allows straightforward setup using standard rack hardware, minimizing installation time.
That flexibility depended on an enclosure designed to manage the environment around dense computing. ScaleMatrix’s materials describe cooling, airflow, fire suppression, and security features. Those provisions do not eliminate the need to plan for adequate electrical service, heat removal, physical protection, and a suitable installation site. “Outside a data center” means outside a conventional data-center room, not outside the need for facility engineering.
What did the facility requirements look like?
For context, ScaleMatrix wrote in January 2019 that NVIDIA’s DGX-Ready Data Center program matched DGX-1 and DGX-2 customers with facilities equipped for power and heat loads of 30–50kW at that time, as well as cooling and airflow needs. In the same period, ScaleMatrix said its DDC cabinets supported more than 52kW of thermal load and described a target PUE of 1.15–1.20. It also promoted temperature and airflow management, environmental controls, and biometric cabinet security. These are ScaleMatrix’s 2019 claims, not current audited measurements or independent test results. Read the company’s January 2019 explanation of DGX-Ready facilities.
The roughly 40kW figure for the DGXPod and the cabinet’s stated thermal-load support describe different things: one is the announced configuration’s reported power requirement, while the other is a cabinet capability claim. Neither figure alone establishes what a particular site can support. A deployment still needs site-specific electrical and thermal assessment, along with airflow, security, and installation planning.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Colocation or deployment on your premises?
ScaleMatrix’s current-facing DGX page presents two broad options: hosting in its DGX-Ready data centers or deploying on customer premises using NVIDIA GPUs and DDC enclosures. The page lists locations in Southern California, Washington, Texas, North Carolina, and Florida. These are vendor descriptions on an undated page, not independently audited statements about current availability or capacity. See ScaleMatrix’s DGX deployment options.
| Consideration | Colocation | On-premises enclosure |
|---|---|---|
| Location | Equipment is hosted in a provider facility; confirm the specific site and service terms. | Equipment is installed at the customer’s site, subject to site suitability and installation planning. |
| Facility infrastructure | The provider’s facility supplies the environment, but capacity and service details need confirmation. | The customer must arrange an appropriate location and ensure power, cooling, airflow, and physical safeguards. |
| What to verify | Available power and cooling, security, connectivity, costs, and contract terms. | Electrical capacity, heat removal, enclosure requirements, security, installation, and ongoing operations. |
The original announcement and promotional materials do not provide enough information to compare total costs or determine which option is preferable for a specific organization. The practical choice depends on site readiness, operating responsibilities, and the provider’s current terms.
Is this the same as NVIDIA DGX Spark or DGX Station?
No. ScaleMatrix’s story concerned a 2019 rack-scale DGX-2, storage, and networking configuration. NVIDIA’s May 18, 2025 announcement describes DGX Spark and DGX Station as later desktop AI systems based on Grace Blackwell. NVIDIA lists Spark at up to 1 petaflop with 128GB of unified memory, and Station at up to 20 petaflops with 784GB of unified system memory; these are NVIDIA’s stated maximum specifications and may change. The newer products are distinct systems, not updated names or specifications for the 2019 DGXPod. NVIDIA’s announcement of DGX Spark and DGX Station says they are intended to support work ranging from desktop prototyping toward cloud or data-center deployment.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
