Kubernetes Instant GPU clusters
with InfiniBand™
Self-service, multi-node clusters provisioned in 20 minutes. Managed Slurm or Kubernetes — no setup, no sales calls
- Rapid provisioning
- Flexible storage
- PAYG
AI Cloud
From rapid prototyping to foundation training to scalable inference
Compute
GB300 NVL72
New1x tray to 2+ racks · NVLink v5
Self-service, multi-node clusters provisioned in 20 minutes. Managed Slurm or Kubernetes — no setup, no sales calls
Multi-node GPU clusters with InfiniBand interconnect. Self-service availability via console, API, or Terraform
4x GB300 SXM6 1152GB VRAM GPU nodes
Price per node
$34.480/h 8x B300 SXM6 2100GB VRAM GPU nodes
Price per node
$60.000/h 8x B200 SXM6 1440GB VRAM GPU nodes
Price per node
$48.880/h 8x H200 SXM5 1128GB VRAM GPU nodes
Price per node
$32.000/h Kubernetes 
SLURM Deploy NVIDIA GPU clusters with InfiniBand interconnect from console, API, or Terraform. No sales calls, no waitlists.
Scale cluster capacity with maximum flexibility. No upfront commitments, no minimum contract duration.
Provision GPU clusters in as little as 20 minutes — up to 50% faster than most AI neoclouds.
Managed Slurm or Kubernetes, integrated observability, pre-configured NVIDIA drivers, and optimized storage — all with expert support.
Leverage bare-metal-level performance with continuous health checks.
We ship clusters with everything a distributed training job requires, pre-configured
Managed Slurm and Kubernetes. Terraform and SkyPilot integrations out of the box.
Logs and metrics for cluster monitoring — integrated with your Grafana, Prometheus, or Loki stack.
pyxis, enroot, hpcx, nccl, nvcc, topology.conf, and dcgmi health checks plugged into Slurm's HealthCheckProgram.
Proactive support from infrastructure engineers and optional co-engineering with Verda's internal AI Lab.
SOC 2 Type II, C5, ISO 27001/27017/27018/27701 certified. GDPR compliant. Hosted in the Nordics on 100% renewable energy.
“Verda is the perfect mix of being nimble and having production-grade reliability for low-latency service like ours. Our startup times and compute costs both dropped significantly.”
Lars Vågnes
“Having direct contact between our engineering teams enables us to move incredibly fast. Being able to deploy any model at scale is exactly what we need in this fast moving industry.”
Iván de Prado
“Our entire language model journey is powered by Verda's clusters, from deployment to training. We can focus on achieving exceptional results without worrying about hardware issues.”
José Pombal
Fast access to the latest GPU instances from NVIDIA, with 1×–8× configurations and NVLink models available.
Learn moreAuto-scaling GPU containers for inference and batch jobs, scaling from zero to hundreds of GPUs, billed per use.
Learn moreHigh-speed NVMe virtual disks, attachable across GPU instances for low-latency reads, writes, and checkpoints.
Learn morePOSIX-compliant storage for concurrent multi-node access, ideal for shared datasets, scratch space, and distributed training.
Learn moreOCI-compliant image storage co-located with Verda compute, for fast pulls into Serverless containers, Batch jobs, and other services.
Learn more
Built in Europe, trusted globally
From rapid prototyping to foundation training and scalable inference — on a single full-stack AI cloud