Available now
SYS-A22GA-NBRT B200 Node
- GPUs
- 8× NVIDIA HGX B200
- GPU memory
- 180 GB HBM3e per GPU
- Interconnect
- 5th-gen NVLink (intra-node)
- NICs
- NVIDIA ConnectX-8 SuperNICs
- Power draw
- ~14.5 kW per node
- Form factor
- Air-cooled rack node
The lineup
Two nodes. Eight GPUs each.
Flagship
SYS-822GS-NB3RT B300 Node
- GPUs
- 8× NVIDIA B300 (Blackwell Ultra)
- Interconnect
- 5th-gen NVLink (intra-node)
- NICs
- NVIDIA ConnectX-8 SuperNICs
- Power draw
- ~19 kW peak / ~14.5 kW sustained
- Density
- 4 nodes per 80 kW cabinet
- Power feed
- 415 V 3-phase PDUs
Networking
One node or a hundred.
Single nodes run entirely on internal 5th-gen NVLink — no external fabric, no extra cost. When a job needs to span racks, nodes are stitched together with non-blocking InfiniBand (copper or optical) so a large cluster behaves like one machine.
InfiniBand is provisioned only for multi-node clusters — you don't pay for fabric a single node never uses.
- Intra-node5th-gen NVLink, all 8 GPUs
- Inter-nodeNon-blocking InfiniBand fabric
- On-boardConnectX-8 SuperNICs included
- TopologySized to your training job
Power & cooling
Built for sustained GPU draw.
GPU servers don't idle. Raw Metal deploys into power-dense colocation engineered to feed and cool them at full tilt.
In every contract
What you get by default.
ConnectX-8 SuperNICs
Every node ships with NVIDIA ConnectX-8 SuperNICs on board — no add-on networking card to buy or wait for.
Full root access
Bare metal, end to end. Bring your own OS, drivers, CUDA, and orchestration. Nothing sits between your stack and the silicon.
Burn-in before handoff
Nodes are racked, cabled, and stress-tested in power-dense colocation. Fabric is validated before you ever log in.
Tell us what you're training.
We'll size the nodes, fabric, and cabinets around your workload.
Request capacity