Mainsail Industries
Night sky over a layer of clouds and distant mountains
Starlight

Secure virtualization

Every workload.
Under one sky.

Secure virtualization for the VMs you run today, with Kubernetes and private AI on the same platform.

Request a deployment review
A fine lattice of small glass cells with one cell lit from within

Secure virtualization

Secure virtualization for the mission.

Starlight runs your existing VMs, Kubernetes, and private AI on one secure platform, from the data center to the edge.

  • Virtual machines
  • microVMs
  • Kubernetes
  • Containers

The Starlight console

Your entire estate.
One console.

Virtual machines, Kubernetes, containers, and private AI, managed from your browser with security built into every layer.

OverviewInfrastructureVirtual machinesContainersStorageFilesAISecretsSecurity policiesSearch or command...+ Newadmin@starlight.localA

Home

6 servers · Full refresh 11:44:53 AM

CustomizeRefresh

Servers

6/6 up

All servers responding

core-0141d 0h
core-0238d 5h
core-0335d 10h
core-0432d 15h
core-0529d 20h
edge-pool-west26d 1h

Alarms

No triggered alarms right now.

142 checks passing across 6 servers

Performance

Average load (1m)

Memory used · 1.21 TB / 1.92 TB

Jump back in

prod-db-01VM · core-02
llama-3.3-70b-instructModel · core-04
k8s-prod-eastKubernetes · 3 microVM nodes

Instances

38

Networks

9

Images

21

Containers

64

Storage

41.2 TB

Starlight console

A glass Starlight star lying on its side, lit softly along its edges

Security

Defense in depth, from the ground up.

Security in Starlight isn't a layer added on top. It runs through every part of the platform, so each workload is protected by multiple independent safeguards from the start.

  • Confidential compute
  • Immutable host
  • Runtime protection
  • Post-quantum protection
  • Defense-grade hardening
  • STIG
  • FIPS

All your infrastructure.
One view.

Virtual machines

Windows & Linux VMs.
Full lifecycle automation.

OverviewInfrastructureVirtual machinesContainersStorageFilesAISecretsSecurity policiesSearch or command...+ Newadmin@starlight.localA

prod-cluster-east

healthy+ New VM

5 hosts · shared storage · live migration enabled

Virtual machines

38

36 running

Hosts reporting

5/5

validated connections

Allocated vCPUs

212

of 320 vCPU

Allocated memory

884 GiB

of 1.5 TiB

NameHostGuest OSvCPUMemoryState
prod-web-01core-01Windows Server 2022816 GiBrunning
prod-web-02core-02Windows Server 2022816 GiBrunning
prod-api-01core-01Linux48 GiBrunning
prod-api-02core-03Linux48 GiBrunning
prod-db-01core-02Linux1664 GiBrunning
prod-db-02core-04Linux1664 GiBrunning
sql-ag-01core-03Windows Server 20221248 GiBrunning
sql-ag-02core-05Windows Server 20221248 GiBlive migrating
ad-ds-01core-01Windows Server 201928 GiBrunning
ad-ds-02core-04Windows Server 201928 GiBrunning
file-srv-01core-05Windows Server 2022416 GiBrunning
rds-gw-01core-02Windows Server 2022416 GiBrunning
siem-collector-01core-03Linux832 GiBrunning
ci-runner-01core-04Linux816 GiBrunning
ci-runner-02core-05Linux816 GiBstopped

Starlight MCP

AI-assisted infrastructure.

Infrastructure

Clusters, shared storage, and live migration in one view.

OverviewInfrastructureVirtual machinesContainersStorageFilesAISecretsSecurity policiesSearch or command...+ Newadmin@starlight.localA

prod-cluster-east

clusteredCluster settings

5-host cluster · shared storage · placement and failover policy enforced

Hosts

5

All hosts healthy

Network

Cluster networking

9 networks · 4 vNet switches

VM storage

Shared storage

41.2 TB used of 96 TB

VM placement

Automatic

Live migration ready

Hosts

HostRolesVMsCPUMemoryState
core-01compute, control-plane962%58%active
core-02compute, control-plane871%66%active
core-03compute, control-plane755%61%active
core-04compute, GPU868%74%active
core-05compute, GPU647%52%active
sql-ag-02live migrating core-05 → core-0364% · no downtime

Private AI

Serve models next to the workloads they support.

OverviewInfrastructureVirtual machinesContainersStorageFilesAISecretsSecurity policiesSearch or command...+ Newadmin@starlight.localA

Models

6 local modelsCatalogCopilotDeploy model

Recent pulls

ready hf://meta-llama/Llama-3.3-70B-Instruct 100%
NameRuntimePlacementStatus
llama-3.3-70b-instructvLLMcore-04 · 2 GPUserving
qwen2.5-coder-32b-instructvLLMcore-05 · 1 GPUserving
gemma-4-12b-it-Q4_K_M.ggufllama.cppedge-pool-westserving
embeddinggemma-300M-Q8_0.ggufllama.cppcore-04serving
whisper-large-v3vLLMcore-05 · 1 GPUserving
mistral-small-3.2-24bvLLMcore-04ready

GPU pool

core-04 · GPU 082%
core-04 · GPU 177%
core-05 · GPU 064%
core-05 · GPU 158%
Serving models5
Active pulls0
Requests / min1,284

Overview

Every server, workload, and alarm at a glance.

OverviewInfrastructureVirtual machinesContainersStorageFilesAISecretsSecurity policiesSearch or command...+ Newadmin@starlight.localA

Home

6 servers · Full refresh 11:44:53 AM

CustomizeRefresh

Servers

6/6 up

All servers responding

core-0141d 0h
core-0238d 5h
core-0335d 10h
core-0432d 15h
core-0529d 20h
edge-pool-west26d 1h

Alarms

No triggered alarms right now.

142 checks passing across 6 servers

Performance

Average load (1m)

Memory used · 1.21 TB / 1.92 TB

Jump back in

prod-db-01VM · core-02
llama-3.3-70b-instructModel · core-04
k8s-prod-eastKubernetes · 3 microVM nodes

Instances

38

Networks

9

Images

21

Containers

64

Storage

41.2 TB

Moving to Starlight

Everything you run today.
Safer and simpler on Starlight.

01

Import

A proven, enterprise-grade import converts your VMs and carries their settings across, so there is nothing to reconfigure.

02

Validate

Run them alongside what you have today, with your backup, monitoring, and automation tooling still in place.

03

Cut over

Move production on your schedule, host by host. Once you are on Starlight, live migration keeps workloads running through maintenance.

Pricing

One per-node subscription.
Batteries included.

Other platforms sell storage, operations tooling, Kubernetes, and security as separate licenses. Starlight includes them, with no core counting and no feature tiers.

  • AI
  • Kubernetes
  • microVMs
  • Containers
  • Backup
  • Security
  • Storage
  • Networking
An operations team at their consoles beneath a wall of status displays

Built for operators

Your tools.
Your workflows.
One platform.

Manage VMs, clusters, and models with the automation you already use, and keep upgrades and backups simple.

  • Ansible
  • Terraform
  • Automated upgrades
  • Backups
  • MCP

Beyond virtualization

One platform.
Every workload.

The same secure foundation that runs your VMs runs containers, Kubernetes, AI agents, and private inference, each with its own isolation boundary.

Wherever your mission runs.

A small team working together in a glass-walled meeting room

Enterprise IT. Bring existing VMs across and add private AI on the same platform.

A clinician preparing a patient on the table of a radiation therapy system

Critical infrastructure. Run operational systems on infrastructure you control, isolated workload by workload.

An operations center team working beneath a wall of maps and displays

Mission operations. Keep running connected, intermittent, or fully disconnected, with the same console.

Who we are

Built by people who have run this infrastructure.

Mainsail Industries was founded by Army and Navy veterans. Our team has built and secured infrastructure at Red Hat, Raytheon, Lockheed Martin, and MITRE.

Meet the team

Partner program

NVIDIA Inception Program

Mainsail Industries is a member of the NVIDIA Inception program.

What's new.

Offer

100 TB of online object storage, on us.

Free with any three- or five-year enterprise commitment.

Talk to us

Blog · September 13, 2026

Kubernetes without the legwork: x0 in five demos

We have written about how x0 works. Now we want to show it. Five short terminal recordings take x0 from creating a Kubernetes cluster to provisioning s0 persistent storage, switching contexts, deploying and scaling an application, and inspecting it all in the x0 TUI.

Read the post

Blog · September 10, 2026

Starlight and air gap

Many products that claim to run at the edge still phone home to establish identity, validate a license, or refresh authorization. The Starlight Offline Authorization Bundle lets a disconnected environment authorize itself and authenticate approved users locally, so connectivity to a vendor never decides whether mission infrastructure stays usable.

Read the post

Blog · September 10, 2026

Why AI needs a new isolation boundary

AI workloads decide what to do while they are already running. Containers and conventional VMs were built for software that did not. MicroVMs give agents and Kubernetes nodes a hardware boundary with a lifecycle that fits modern infrastructure, and confidential computing extends that boundary to the host itself.

Read the post

Your infrastructure. Under one sky.

Make it yours.

Secure virtualization today. Kubernetes and private AI when you are ready. One platform.

Request a deployment reviewDiscover Starlight