Skip to main content
KONST

JP-CHIBA-H100

Chiba, Japan
NVIDIA H100 plan with no down payment

A Chiba facility with 3 nodes and 24 H100 GPUs, 400G Ethernet, and air cooling.

AI TrainingInferenceHPC

24

Total GPUs

3

Nodes (8x H100 per node)

400G

Ethernet

2.20 and up

USD/hr/GPU (36 month term)

Compute nodes

3 nodes with 8 NVIDIA H100 GPUs per node.

GPU
24 units · 8x H100/node
GPU memory
80 GB per GPU, 640 GB per node
CPU
Dual AMD EPYC 4th Gen 96C or dual Intel Xeon 4th Gen 24C (upgradeable)
Memory
1.5 TB DDR5 ECC DIMM per node
Storage
OS: 2x 3.84 TB NVMe (RAID 1); data: up to 15 TB enterprise NVMe SSD

Network and cooling

Air cooling with 400G Ethernet.

Network (GPU Fabric)
Ethernet 400G · NVIDIA ConnectX-7
Network (External)
10 GbE uplink per node
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.

Need H100 capacity without a large upfront payment?

The no-down-payment plan has limited availability. Share your lease term and required GPU count, and we will confirm delivery timing.

Book a consultation