ONLINEUPTIME:99.99% Network Availability SLAONLINEPOWER:100 kW/Rack Direct-to-Chip Liquid CoolingONLINEUS-East:128 x H100 SXM AvailableONLINEUS-West:64 x H200 AvailableLIMITEDEU-Central:GB200 NVL72 Reserve OnlyFABRIC:GPUDirect RDMA Enabled · 3.2 Tbps InfiniBand · Data Egress: $0.00/GB
ONLINEUPTIME:99.99% Network Availability SLAONLINEPOWER:100 kW/Rack Direct-to-Chip Liquid CoolingONLINEUS-East:128 x H100 SXM AvailableONLINEUS-West:64 x H200 AvailableLIMITEDEU-Central:GB200 NVL72 Reserve OnlyFABRIC:GPUDirect RDMA Enabled · 3.2 Tbps InfiniBand · Data Egress: $0.00/GB

Hardware & specs

Non-virtualized nodes, liquid-cooled at full clock.

Every node is dedicated bare metal on a 3.2 Tbps InfiniBand fabric. No hypervisor, no noisy neighbors, no thermal throttle.

NVIDIA B200

FP8 / FP4 Support

$6.90

per GPU / hour

Memory

180GB HBM3e

Bandwidth

8.0 TB/s Bandwidth

Fabric

3.2 Tbps InfiniBand

Node specification
Architecture
Blackwell
Memory
180GB HBM3e
Memory Bandwidth
8.0 TB/s
Precision
FP8 / FP4 / BF16
Interconnect
NVLink 5 · 1.8 TB/s
Cooling
Direct-to-chip liquid
GPUMemoryBandwidthBest for$/GPU/hr
NVIDIA B200180GB HBM3e8.0 TB/s BandwidthFP8 / FP4 Support$6.90
NVIDIA H200141GB HBM3e4.8 TB/s BandwidthUltra-large LLM Inference$3.60
NVIDIA H100 SXM80GB HBM33.35 TB/s Bandwidth3.2 Tbps InfiniBand$2.45
NVIDIA L40S48GB GDDR6864 GB/s BandwidthFine-tuning & Vision$1.10