NVIDIA HGX Rubin NVL8
NVIDIA's eight-GPU Rubin system, on a waitlist for first allocation. The site can be planned before it ships.
- GPUs
- 8 × Rubin
- GPU memory
- 288 GB HBM4 each
- NVLink
- 3.6 TB/s per GPU
- Power
- Not final
- Status
- Waitlist
NVIDIA HGX Rubin NVL8 specifications.
What NVIDIA has published so far.
- Generation
- Rubin
- GPUs
- 8 GPUs per nodeOne NVLink domain per system, as on HGX B300.
- GPU memory
- 288 GB HBM4 per GPU
- NVLink
- 3.6 TB/s per GPU, sixth generationAgainst 1.8 TB/s per GPU on HGX B300.
- Networking
- 1.8 TB/s per systemBi-directional, as NVIDIA states it.
- Host CPUs
- NVIDIA Vera or x86Paired with Vera, NVIDIA calls it HGX Vera Rubin NVL8.
- Coolant inlet
- 45 °C
- System power
- Not finalNVIDIA hasn't published a figure yet.
- Availability
- Waitlist for first allocation
HGX Rubin NVL8 figures are from NVIDIA's HGX platform page and may change before volume shipments.
Planning a site before it ships.
NVIDIA hasn't published a rack layout or a power figure yet. This is what's known and how we plan around it.
- 1
Liquid to every position
NVIDIA lists a 45 °C coolant inlet, and the systems announced so far are liquid-cooled. The hall needs facility water at each rack position.
- 2
Power with headroom
With no published system figure, we size feeds above what HGX B300 needs and confirm the design against the figure in the order.
- 3
Network sized per system
NVIDIA gives 1.8 TB/s per system, against 1.6 TB/s per system for HGX B300. We size the fabric when the order and NVIDIA's reference design are set.
- 4
A Vera or an x86 host
The choice changes the host board, not the GPU side. It's settled in the design, with your software stack in mind.
HGX Rubin NVL8 against HGX B300.
Both are eight-GPU systems. Rubin moves to HBM4 and doubles NVLink bandwidth per GPU; B300 can be built now and runs on air.
| HGX B300 | HGX Rubin NVL8 | |
|---|---|---|
| Generation | Blackwell Ultra | Rubin |
| GPU memory | 288 GB HBM3e per GPU | 288 GB HBM4 per GPU |
| NVLink | 1.8 TB/s per GPU | 3.6 TB/s per GPU |
| Networking | 1.6 TB/s per system | 1.8 TB/s per system |
| Cooling | Air, or liquid where the hall has it | Liquid |
| Power | About 14 kW per system | Not final |
| Availability | About 10 to 14 weeks from signing | Waitlist for first allocation |
Figures from NVIDIA. Rubin figures may change before volume shipments.
Other NVIDIA platforms we build on.
The same four stages on each: design, source, build and operate. All hardware.
- NVIDIA HGX B300Blackwell Ultra8 GPUs per node288 GB HBM3e per GPUAirAvailable
- NVIDIA HGX B200Blackwell8 GPUs per node180 GB HBM3e per GPUAirAvailable
- NVIDIA GB200 NVL72Blackwell72 GPUs per rack186 GB HBM3e per GPULiquidAvailable
- NVIDIA GB300 NVL72Blackwell Ultra72 GPUs per rack288 GB HBM3e per GPULiquidAvailable
- NVIDIA VR200 NVL72Rubin72 GPUs per rack288 GB HBM4 per GPULiquidWaitlist
HGX Rubin NVL8 questions.
We build HGX Rubin NVL8 clusters as a deployment your team runs, a managed cluster you own and we operate, or a reserved cluster we own. Other platforms.
NVIDIA's eight-GPU Rubin system, after HGX B300. It pairs with Vera CPUs, as HGX Vera Rubin NVL8, or with x86 CPU baseboards.
It's on a waitlist for first allocation. We don't publish dates; the proposal sets them once allocation is confirmed.
NVIDIA lists a 45 °C coolant inlet, and the systems announced so far are liquid-cooled. Plan for facility water at the rack.
NVIDIA hasn't published a figure. We plan headroom over HGX B300 and confirm the design against the figure in the order.
NVL8 keeps the eight-GPU NVLink domain of an HGX system. Vera Rubin NVL72 joins 72 GPUs in one rack, for models split across many GPUs, and needs a site built for rack-scale liquid cooling.
Only with facility water at the rack, or a fit-out that adds it. An air-only hall needs a retrofit first.