NVIDIA HGX B200
Eight Blackwell GPUs in a node. The workhorse Blackwell platform.
- 8 GPUs per node
- 1.4 TB GPU memory
- Air or liquid
- Available Q1 2027
Configurations
Pick the footprint that fits. Every one is sold on the same contract.
- 8 GPUsOne node
- The smallest unit. Air cooled, in an existing Santiago data centre, live from December 2026.
- 16 to 248 GPUsTwo to thirty-one nodes
- One fabric, one cluster.
- 256 GPUs and moreThirty-two nodes and up
- Rows of racks, air or liquid cooled.
Specifications
- GPUs
- 8 Blackwell per node
- GPU memory
- 1.4 TB, 180 GB HBM3e per GPU
- NVLink
- 1.8 TB/s per GPU, 14.4 TB/s in the node
- CPUs
- 2 x86 CPUs
- System memory
- 2 TB, up to 4 TB
- Networking
- ConnectX-7, 400 Gb/s per GPU; BlueField-3
- Power
- About 14.3 kW per node
- Cooling
- Air or liquid
- Fabric at Fulmis
- InfiniBand or Spectrum-X Ethernet, sized to the workload
- Software
- Slurm and managed Kubernetes
- Location
- A data centre in the Santiago Metropolitan Region
- Available
- From December 2026
Performance
Per node, as NVIDIA publishes it.
- FP4 Tensor, dense
- 72 PFLOPS
- FP8 / FP6 Tensor, dense
- 36 PFLOPS
- FP16 / BF16 Tensor
- 36 PFLOPS
- TF32 Tensor
- 18 PFLOPS
- FP64
- 296 TFLOPS
- GPU memory
- 1.4 TB HBM3e, 64 TB/s
- NVLink bandwidth
- 14.4 TB/s
- Scale-out network
- 0.8 TB/s, ConnectX-7
- Node power
- About 14.3 kW (NVIDIA DGX B200)
NVIDIA's published figures. Tensor Core numbers carry sparsity unless the row says dense; H100, H200 and RTX PRO figures are per GPU because NVIDIA publishes no node total. Power figures are the ones we reserve and protect to, with NVIDIA's own where it publishes one.
How we deploy it
HGX B200 is the eight-GPU Blackwell node, the platform most of the market quotes by the node. It trains and serves, it is sold from one node, and it is the platform to name when a workload does not need the 72-GPU NVLink domain.
A node is the smallest unit. Fabric, storage and software are sized to the cluster you describe.
Fulmis places the purchase order, or takes delivery of the fleet you already own. Either way we bring it into Chile, commission it against the vendor baseline and operate it for the term. You reserve the capacity, or your own hardware; the workload and the data stay yours.
Terms
- Term
- Three to five years
- Price
- Fixed for the term, in the term sheet
- Commitment
- Take-or-pay
- Prepayment
- None to 25% at signature
- Smallest block
- One node, 8 GPUs
- Start
- Q1 2027, first slot
Other platforms
| Platform | GPUs | GPU memory | NVLink | Power | Cooling | Smallest block | Available |
|---|---|---|---|---|---|---|---|
| GB300 NVL72Rack scale | 72per rack | 20 TB288 GB per GPU | 130 TB/s1.8 TB/s per GPU | 132 to 140 kW155 kW peak | Liquid | 1 rack, 72 GPUs | Q1 2027 |
| GB200 NVL72Rack scale | 72per rack | 13.4 TB186 GB per GPU | 130 TB/s1.8 TB/s per GPU | 120 to 132 kWper rack | Liquid | 1 rack, 72 GPUs | Q1 2027 |
| HGX B3008-GPU node | 8per node | 2.3 TB288 GB per GPU | 14.4 TB/s1.8 TB/s per GPU | About 14 kWper node | Air or liquid | 1 node, 8 GPUs | December 2026 |
| HGX H2008-GPU node | 8per node | 1.1 TB141 GB per GPU | 7.2 TB/s900 GB/s per GPU | About 10.2 kWper node | Air | 1 node, 8 GPUs | Q1 2027 |
| HGX H1008-GPU node | 8per node | 640 GB80 GB per GPU | 7.2 TB/s900 GB/s per GPU | About 10.2 kWper node | Air | 1 node, 8 GPUs | Q1 2027 |
| RTX PRO 6000 Blackwell8-GPU node | 8per node | 768 GB96 GB per GPU | PCIe Gen 5no NVLink | 5.4 to 6.7 kWper node | Air | 1 node, 8 GPUs | Q1 2027 |
| Vera RubinRack scale | 72per rack | HBM4not yet published | Not published | Not published | Liquid | 1 rack | On request |