One signed image. The whole stack.
Compute, storage, networking, data protection, operations, and security all ship in the same firmware image. No add-on bundles, no separate licenses, no suite to assemble and reconcile.
Every layer, one image.
Hardware-accelerated virtualization
A hardware-accelerated type-1 hypervisor, built into the signed image.
Full VM lifecycle
Create, start, stop, clone, and manage VMs from CLI, API, or web.
Live migration
Move running VMs between nodes with zero downtime. Storage is already shared cluster-wide, so only memory state moves.
HA with fencing
On node failure, VMs restart automatically on a healthy node, with quorum-confirmed fencing to prevent split-brain.
Maintenance mode
Evacuate a node via live migration, patch it with a signed update, and bring it back, non-disruptively.
GPU passthrough
Assign physical GPUs to VMs for AI/ML and graphics, with cluster-aware placement. No per-GPU software tax.
Firmware, not a suite.
There is no external management server to babysit. The control plane lives in the nodes; one OTA update rolls the entire stack atomically, with automatic rollback.
One image, A/B updated
Hypervisor, storage, networking, and control plane in a single signed artifact. Update flips partitions; failure rolls back.
Nodes dial out, nothing dials in
Every node connects outbound to the control plane's message bus and console relay. No SSH daemon, no inbound management ports, no agent to patch on its own schedule.
Self-healing
Health in plain language. Automatic reactivation, rebalancing, and drain-on-disk-failure with no operator on call.
Not features they lack. Things their model prevents.
These aren't gaps in VMware's roadmap. They're choices Broadcom's revenue model can't unmake without cannibalizing the acquisition.
See it scale, then price it.
Watch a cluster grow from one node to five hundred, and fail over from edge to datacenter.