GPU and multi-datacenter

GPU nodes

Worker nodes with NVIDIA A2000, A4000 Ada or A6000 Ada are available as Enterprise configuration for inference, rendering and scientific computing. Open a ticket with the model and count.

Multi-datacenter

Clusters can span FRA1, FRA2, HEL1 and EYL1 for geo-redundancy. Nodes in different sites need the MPLS backhaul between them; the design is done together with you as an Enterprise cluster (architecture, control plane placement, storage topology, knowledge transfer).

Enterprise and custom configurations (GPU nodes, memory-optimised nodes, multi-datacenter) depend on hardware availability. Lead times vary; we agree the schedule with you in the initial call.