Sovereign Cloud Infrastructure
from Germany
Easy to start. Easy to scale.
Prices
Cloud services
- Only pay for what you use
- Exclusively for B2B
- Freely scalable
- Start now
| name | vcpu | ram | disk | hour | month |
|---|---|---|---|---|---|
| s1.small | 2 | 2.048 | 25 | 0.01997 | 14.38 |
| s1.medium | 4 | 4.096 | 50 | 0.03994 | 28.76 |
| s1.large | 6 | 8.192 | 75 | 0.06727 | 48.43 |
| s1.xlarge | 8 | 16.384 | 100 | 0.10932 | 78.71 |
| s1.xxlarge | 12 | 32.768 | 150 | 0.19342 | 139.26 |
| s1.xxxlarge | 16 | 65.536 | 200 | 0.3364 | 242.21 |
| name | vcpu | ram | disk | gpu | hour | month |
|---|---|---|---|---|---|---|
| g1.large | 4 | 16.384 | 200 | A10 | 0.9482 | 682.7 |
| g1.xlarge | 8 | 32.768 | 350 | A10 | 1.0487 | 755.06 |
| g1.xxlarge | 16 | 65.536 | 500 | A10 | 1.2251 | 882.07 |
| g2.large | 4 | 16.384 | 200 | A40 | 1.1782 | 848.3 |
| g2.xlarge | 8 | 32.768 | 350 | A40 | 1.2787 | 920.66 |
| g2.xxlarge | 16 | 65.536 | 500 | A40 | 1.4551 | 1047.67 |
| g3.xxlarge | 32 | 32.768 | 400 | RTX 6000 Blackwell Server Edition | 2.380227 | 1713.76 |
Features
Our services for your success
Cloud server
clients. You can choose between General Purpose (s1), Performance Optimized (p1) and Dedicated (d1) instance families - depending on your CPU-to-RAM ratio and I/O requirements. All flavors can be billed hourly, with no minimum term.
- High Performance NVMe
- Dedicated vCPU
- Multiple OSes and Flavors
- Backup and snapshots
- Nvidia RTX 6000 Blackwell Server Edition
Networking
database, or connect multiple projects via a shared private network. Floating IPs allow you to flexibly move public IP addresses between instances.
- Virtual Private Cloud
- IPV4+IPv6
- Security Groups
- Floating IPs
S3 Object Storage
- Versioning & lifecycle management
- Object Locking & Immutable Storage
- Amazon S3 API compatible
- OpenStack Swift API compatible
- Storage classes: S3, S3-Eco
- Find out more about our S3 storage
Administration
You manage your resources the way that suits your team: via the MyNWS dashboard for visual management, or completely programmatically via the native OpenStack API. Infrastructure as Code works
out-of-the-box with Terraform, Ansible and Pulumi - no proprietary CLI, no vendor-specific abstractions. What you configure here today can be ported to any other OpenStack provider tomorrow.
- Native API
- Easy and Expert-Mode UI
- SDKs
- Third party tooling (Ansible, Terraform, Pulumi)
Storage
- Volumes
- NVMe
- Object Storage
- Volume Types
- Backups and snapshots
- Air Gapped Backups
GPUs
Our GPU instances are for teams that need powerful compute - and whose data needs to stay in Germany. Whether ML training, inference or scientific simulations: you can choose between NVIDIA A10 (24 GB VRAM), NVIDIA A40 (48
GB VRAM) and the RTX 6000 Blackwell Server Edition (96 GB VRAM). PCI passthrough enables direct hardware access for maximum performance. All GPU instances run GDPR-compliant in Nuremberg - without a detour via US infrastructure.
- NVIDIA A40
- NVIDIA A10
- Nvidia RTX 6000 Blackwell Server Edition
- PCI passthrough
- Learn more
Load Balancer
sessions keep user sessions consistent on a backend - important for stateful applications. Blue-green deployments and canary releases can be mapped using weighted traffic routing.
- TCP/UDP
- Health Checks
- Sticky Sessions
Backup and snapshots
Protect your infrastructure on multiple levels: VM snapshots for fast point-in-time recovery, automated volume backups for daily protection, and air gapped backups for full disaster recovery. The dedicated
Air Gapped Backup is physically separated - no network access means maximum protection against ransomware and accidental deletion. All backups are stored in Nuremberg, not in an external cloud service.
- Volumes
- S3 Buckets
- NVMes
- Dedicated storage cluster for backups
- Air Gapped Backups
Our customers
Our solutions are in use – and with measurable success. Take a look at how other companies benefit from our services.
Awards
Powered by OpenStack
Why OpenStack?
OpenStack is the most widely used open source cloud platform in the world – used by CERN, Deutsche Telekom, BMW and hundreds of other companies that have deliberately avoided proprietary hyperscalers. Unlike AWS, Azure or GCP, OpenStack has no vendor that can increase licenses, adjust roadmaps or change access policies.
For you, this means that your infrastructure can be controlled via the standardized OpenStack API – with Terraform, Ansible, the OpenStack CLI or the Python SDK. What you set up today will also run tomorrow in another OpenStack data center, with another provider or on-premises. No vendor lock-in due to proprietary abstractions.
NWS has been running OpenStack for over a decade and is a member of the OpenInfra Foundation.
VMware alternative: OpenStack
Since the Broadcom takeover, many companies have reported VMware license increases of between 300% and 1,000%. If you need to recalculate your infrastructure anyway, make the decision for OpenStack now – not in two years’ time.
NWS Cloud Services is the direct operating environment for this change: OpenStack Compute (Nova) replaces vSphere/ESXi, Ceph block storage (Cinder) replaces vSAN, and OVN-SDN replaces VMware NSX – with full API compatibility for Terraform and Ansible, which you are already using. Existing VM images can be migrated via `virt-v2v`; our NWS MyEngineer accompanies the process.
The decisive difference to switching to AWS or Azure: No new proprietary dependencies. No new lock-in. Open standards from start to finish.



