Skip to main content
KONST

SE-H200

Sweden
NVIDIA H200 facility for mixed workloads

A Nordic facility with 37 nodes and 296 H200 GPUs, each with 141 GB of memory, for training, inference, rendering, and CAD workloads.

TrainingInferenceRenderingCAD

296

Total GPUs

37

Nodes (8x H200 per node)

141GB

GPU memory per card

2.88 and up

USD/hr/GPU (12, 24, or 36 month term)

Compute nodes

37 nodes with 8 NVIDIA H200 GPUs per node.

GPU
296 units · 8x H200/node
GPU memory
141 GB HBM3e per GPU, approximately 41.7 TB total
CPU
2x AMD EPYC 9374F per node
Memory
1.5 TiB per node
Storage
OS: 2x 2 TiB NVMe; data: 2x 3 TiB NVMe (10 TiB raw per node)

Network and cooling

Air cooling with 3x ConnectX-7 per node.

Network (GPU Fabric)
10 Gbit/sec Ethernet plus 3x ConnectX-7 per node
Network (External)
Same as GPU Fabric
Cooling
Air cooling

Note

Prices are indicative hourly rates per GPU and vary by contract length, memory option, and payment structure. Some facility cities, data center names, Tier ratings, and contracting entities are available on request and will be disclosed during contracting. The formal quote sets the final terms.

One facility for both training and graphics workloads

Share your workload mix and preferred lease term. We will respond with a node configuration and tiered pricing.

Book a consultation