Your team
Manage workloads, networks, capacity and access through the Console, API or Terraform. Choose the servers and the maintenance plan.
Your servers, run like a public cloud.
Manage Virtual Machines, Containers, Kubernetes, storage and networks from one Console.
Your team runs the infrastructure.
Neverinstall
Our engineers provide 24×7 support under the agreed contract. We maintain KVM and KubeVirt patches and deliver signed updates.
Platform architecture
Supported x86 and arm64 servers
One ConsoleWeb Console, CLI, Go SDK and Terraform
Virtual Machines
Containers
Kubernetes
Storageblock, object and NFS
Networksone per tenant
IdentityIAM and secrets
The Console your team works in.
Compute, storage, networks and operations.
Compute
Run Windows and Linux Virtual Machines, plus Containers and Kubernetes clusters.
Storage
Use Block Volumes, S3-compatible Object Storage and NFS. Back up Virtual Machines to an off-cluster vault.
Networking
Give tenants separate networks, with routing, Load Balancers and VPN access.
Identity
Connect your identity provider. Assign roles, quotas and access to resources.
Operations
Monitor workloads, review the Audit Trail and automate provisioning through the API and Terraform.
Full platform specifications
Compute
Virtual Machines and Containers
Windows and Linux Virtual Machines from images or cloud-init, beside Containers that autoscale and run scheduled jobs.
Managed Kubernetes
Kubernetes clusters for your teams, with an Image Registry that replicates between sites.
Day-two operations
Clone, hibernate, resize and live-migrate Virtual Machines, with secure boot and Windows sysprep.
Storage
Block, object and NFS
Block Volumes, S3-compatible Object Storage with versioning and lifecycle rules, and NFS shares, on internal disks or your SAN.
Backups
Scheduled backups copy running Virtual Machines to an off-cluster vault, and restore them as new machines. Veeam is supported.
Encryption and keys
Encrypted disks and buckets, with keys held in each tenant’s own vault. Bring your own keys and set rotation schedules.
Networking
Tenant Networks
Each tenant gets a private network with Subnets, Route Tables, internet and NAT Gateways, and Security Groups.
Load Balancers
Network (TCP and UDP) and application (HTTP) Load Balancers, internet-facing or internal, with TLS certificates.
VPN, DNS and SSH
Site-to-site IPsec VPN, DNS inside each network, and audited SSH to private machines through a managed bastion.
Identity
IAM, SSO and secrets
Roles, API keys, single sign-on through your identity provider, and secrets management. Identity is not licensed separately.
Tenants and quotas
Isolated tenants with nested compartments, quotas, and resources shared between tenants when you allow it.
Audit Trail
Review administrative actions with actor and source details. Forward audit records to your SIEM.
Operations
Console, CLI, Go SDK and Terraform
Automate provisioning through the API, CLI, Go SDK and Terraform. Review coverage for your workflows during evaluation.
Monitoring and metering
CPU, memory, disk and network graphs for every machine and tenant, with no guest agent, and usage records you can export.
Rolling upgrades
Upgrade the cluster node by node from signed bundles, while workloads keep running.
Public addresses and network requirements
Give Virtual Machines and Load Balancers public addresses on the Layer 2 network you already run.
The platform answers for each address itself, so you need no BGP and no changes to your routers.
- You provide
A block of addresses routed to your network segment
- BGP
Not required
- Ownership
An address stays with its machine or Load Balancer until you delete it
- Alerts
A warning when an address has nothing behind it
Check your existing hardware first.
Send us your server, disk and network models. We review the configuration before the pilot, including any changes it needs.
Add supported nodes as capacity grows. Review failure scenarios, spare capacity and recovery requirements before choosing a highly available setup.
- 1 node
Runs your workloads, without high availability
- 2 nodes
1+1 high availability: either node carries the workloads if the other fails
- 3 or more
Full high availability, sized for the failures you need to cover
- Cluster size
Up to 1,000 nodes
- Install
About 30 minutes, with a first cluster running in a day
Hardware and cluster requirements
- Processors
x86 and arm64
- Mixed nodes
Different makers, models and sizes in one cluster
- Generations
Servers back to 2012
- Compatibility
Confirmed against your server configuration before the pilot
- Requirement
CPUs with hardware virtualization, on hardware that runs Debian
- Storage
Internal disks, your existing SAN, or both
- Network
1 Gbps minimum, 10 Gbps recommended
Back up and restore before moving workloads.
- Backups
Scheduled Virtual Machine backups to an off-cluster vault
- Restore
Restore a backup as a new Virtual Machine
- Recovery targets
Planned per workload and tested in the pilot
- Two-site disaster recovery
On the roadmap
Move one workload group first.
Run beside your current platform.
- Review the environment
Check your server list, storage, network and workload dependencies. Agree on the pilot scope.
- Run a workload
Import representative Virtual Machines or deploy an application. Test access, networking and performance.
- Test daily operations
Run maintenance, apply an update, back up a Virtual Machine and restore it. Test the recovery targets you agreed.
- Review the results
Keep the sizing and pricing report. Decide which workloads to move next and agree on their cutover and rollback plans.
A subscription for each node.
Identity and the Console are included. The quote records support and contract terms.
- Unit
Per node, not per core
- Modular
Virtualization, storage and Containers, separately or together
- Commitment
No 3-year lock-in
- Trial
A pilot on your servers before you buy
Where teams start.
Evaluate in depth.
Compare private cloud platforms
| VMware Cloud Foundation | Neverinstall | |
|---|---|---|
| Licensing | Subscription only, per physical core, with at least 16 cores licensed per CPU. Perpetual licenses ended in January 2024. | Per node, quoted for your deployment. Storage, Containers and Managed Kubernetes are included, and extra Virtual Machines add no platform charge. |
| Storage | Each VCF core includes 1 TiB of vSAN. More capacity is licensed separately. | Block, Object and File Storage on internal disks or an existing SAN, with no charge per TiB. |
| Administration | vSphere, vSAN, NSX, VCF Operations and VCF Automation, with licenses registered through VCF Operations and the VCF Business Services console. | Compute, storage, networking and identity from one Console. Console, CLI, Go SDK, first-party Terraform provider and an OpenAPI 3 API. |
| Disconnected sites | Disconnected sites upload a license usage file at least every 180 days. | Offline installation and signed update bundles, with no license check-in. |
| Hardware | x86 for production. ESXi on Arm is a tech preview in 9.1, without vSAN or NSX. | x86 and arm64 servers back to 2012, mixed in one cluster. We check your models before the pilot. |
GPU hardware and pooling
Add GPU servers to the cluster. Schedule jobs across available GPUs.
- Modes
vGPU, MIG, time-slicing and passthrough
- Pooling
GPUs across nodes, where a GPU interconnect links them
- Cards
Each mode needs a card that supports it; we check yours first
On the roadmap.
Ask us for dates and early access.
- Live migration from vSphere
Move running Virtual Machines from VMware to Neverinstall Private Cloud without stopping them.
- Two-site disaster recovery
Replicate workloads to a second site and fail over to it. Planned for October 2026.
- Network peering
Connect Tenant Networks to each other, with consent on both sides.
- Serverless Containers
Containers that scale to zero, with revisions and traffic splitting between them.
- Image factory and cloud shell
Build golden images from the Console, and work from a terminal in the browser.
- Compliance and recommendations
Compliance reports for your environment, and suggestions to right-size what you run.
Engineering research
- Building overlapping tenant networks with Cilium
We extended Cilium so tenants can reuse private address ranges across Virtual Machines and Containers. The article covers the design and a three-node test run.
- Model checking exposed failures in cloud recovery
A June 2026 review with TLA+ models produced 14 correctness findings. The models exposed recovery failures and changed how retries and completion checks work.
- Keeping failed cloud deletions visible and retryable
A failed delete keeps its record, so cleanup can resume. The article covers retained records, ordered cleanup and the checks made before a delete reports complete.
Private cloud questions.
How long does installation take?
About 30 minutes. Most teams have a first cluster running in a day, after a hardware review of your server, disk and network models.
Will migration require downtime?
Plan a cutover window for imported workloads and test it in the pilot. Live migration of running Virtual Machines from vSphere is on the roadmap.
Can x86 workloads move to ARM servers?
Hardware support does not make a workload portable between processor architectures. Review guest operating systems, images and applications against the servers they will run on.
What recovery targets can we agree?
Define recovery time and acceptable data loss for each workload. Test backup and restore against those targets. Two-site disaster recovery is on the roadmap.
How do we review support and attestations?
Support is 24×7, from the engineers who build the platform, with response targets and SLAs in your contract. Neverinstall is SOC 2 Type II attested and ISO 27001 certified; the report and certificate are available under NDA.
Plan your private cloud.
Send us your server list and the workloads you want to move first.