Free tools Windows power users keep installed
One-click scans. No signup required.
The NVIDIA Vera Rubin NVL72 is a rack-scale AI system built around 72 Rubin GPUs and 36 Vera CPUs. It also includes NVLink 6 switches for communication among the rack’s GPUs, plus ConnectX-9 SuperNICs and BlueField-4 DPUs for networking and infrastructure tasks. NVLink handles communication within the rack; Quantum-X800 InfiniBand or Spectrum-X Ethernet can connect it to systems beyond the rack.
What does the Vera Rubin NVL72 rack include?
NVIDIA describes the NVL72 as an integrated rack-scale system, not simply a collection of GPUs. Its listed components combine accelerator and CPU compute, a high-speed internal GPU fabric, and networking and infrastructure hardware.
| # | Preview | Product | Price | |
|---|---|---|---|---|
| 1 |
|
1U Rack Mount Pro Enclosed 3X Jetson Front Removable | Buy on Amazon | |
| 2 |
|
Alphacool ES H200 141GB NVL PCI-E 1-Slot-Design GPU Water Block | $409.98 | Buy on Amazon |
| Component | Role in the rack |
|---|---|
| 72 Rubin GPUs | The system’s AI accelerators. NVIDIA presents them as able to operate together as a rack-scale accelerator. |
| 36 Vera CPUs | The rack’s CPU compute, paired with the GPUs. The specifications establish the count, but do not prescribe how every workload divides work between CPUs and GPUs. |
| NVLink 6 switches | The internal scale-up fabric for GPU-to-GPU communication. NVIDIA’s DGX specifications list nine L1 NVLink switches. |
| ConnectX-9 SuperNICs | Network interfaces for connections beyond the rack. NVIDIA’s DGX specifications list InfiniBand and Ethernet interface options at 800 Gb/s. |
| BlueField-4 DPUs | Data-processing units included among the rack’s networking and infrastructure components. The cited specifications do not establish a particular deployment-specific offload configuration. |
| Quantum-X800 InfiniBand or Spectrum-X Ethernet | Named scale-out networking options for connecting the rack to a larger AI-factory network. |
These are NVIDIA-published product specifications and descriptions, not independent test results. For deployment planning, interface choices and configuration details should be checked against current product documentation.
How do Rubin GPUs, Vera CPUs, and NVLink work together?
Rubin GPUs provide accelerator capacity
The 72 Rubin GPUs are the system’s main AI accelerators. NVIDIA says that within a DGX Vera Rubin NVL72, the GPUs can operate as one rack-scale accelerator through NVLink 6. That describes the system’s scale-up design; it does not mean every application will use all 72 GPUs in the same way or achieve the same performance.
Recommended Free Tools
#1 Best Overall
- 1U rack mount Pro Enclosed 3x Jetson - FRONT REMOVABLE
- This rack mount comes complete with front-removable trays for 3x Jetson. Compatible with: NVIDIA Jetson Nano 2GB Developer Kit, NVIDIA Jetson Nano Developer Kit, NVIDIA Jetson Orin Nano Developer Kit, NVIDIA Jetson Orin Nano Super Developer Kit, NVIDIA Jetson Xavier NX Developer Kit
- Each Jetson can be individually removed from the rack mount from the front, while the rest remains in place. Quickly and easily with two handy fast-click buttons. No tools needed!
- The screws to secure the Jetson to the tray are also included. Extra (spare) Fast-Click buttons are also included.
- High grade Aluminum. Matt black powder coated finish. NOT made in China.
Vera CPUs provide the rack’s CPU compute
The 36 Vera CPUs are paired with the GPU resources in the rack. At the system level, they provide CPU compute alongside GPU acceleration. The listed component counts do not establish a universal workload split, and CPU memory should not be treated as GPU memory.
NVLink connects GPUs inside the rack
NVLink 6 is the internal GPU interconnect, or scale-up fabric. NVIDIA describes it as providing all-to-all connectivity among the rack’s GPUs, so data can move between accelerators without relying on an external Ethernet or InfiniBand network for that within-rack role.
Is NVLink the same as Ethernet or InfiniBand?
No. They serve different parts of the system:
- NVLink 6: scale-up communication among GPUs within the NVL72 rack.
- Ethernet or InfiniBand: scale-out networking that links the rack with other systems or the wider AI-factory network. NVIDIA names Spectrum-X Ethernet and Quantum-X800 InfiniBand as options.
These technologies complement one another: NVLink provides the rack’s internal GPU fabric, while Ethernet or InfiniBand provides connectivity beyond the rack.
Rank #2
- Water block for the Nvidia H200 NVL, 141GB HBM3 graphics card
- The cooler top is made of carbon
- Cooler only needs 1 slot in the server rack instead of the previous 1.5 slots
- Dimensions (L x W x H): 268.59 x 94.45 x 20.67 mm
What NVLink bandwidth does NVIDIA specify?
NVIDIA’s figures differ by source, and the available descriptions do not reconcile their definitions. The NVL72 product overview lists 216 TB/s of rack NVLink bandwidth, while NVIDIA’s March 16, 2026 newsroom announcement cites 260 TB/s. Separately, NVIDIA’s NVLink reference lists 3,600 GB/s per GPU for sixth-generation NVLink and describes full all-to-all communication. These are different kinds of figures and should not be combined as though they measure the same thing.
All three are vendor-published specifications or claims, not independent measurements. The product overview’s rack figure and the newsroom figure may reflect different system-level definitions or specification updates; the available information does not establish which explanation applies.
What is known about production and availability?
On March 16, 2026, NVIDIA said the platform’s chips were “now in full production.” That is a dated company statement, not confirmation that every NVL72 configuration is available in every region. Regional availability, pricing, and delivery schedules are not established here; organizations should ask NVIDIA or an authorized systems integrator about current configuration and local availability.
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.
