# VMware alternatives: the credible options, the trade-offs and a safe migration plan > Why VMware customers are rethinking their renewal, how the main alternatives compare on licensing, skills and storage, and a staged plan for moving off vSphere. - URL: https://computese.com/vmware-alternatives/ - Author: Duong Quan Nguyen, CEO, Computese - Published: 2026-10-01 - Updated: 2026-10-09 - Topics: Cloud, Hosting ## In short - VMware under Broadcom is subscription-only: the flagship offers are the VMware Cloud Foundation and vSphere Foundation bundles, licensed per core with a 16-core minimum per physical CPU. - The credible alternatives are OpenShift Virtualization, Proxmox VE, Microsoft Hyper-V or Azure Local, Nutanix AHV and XCP-ng. Each trades licence cost against skills, storage design and support. - Every serious target has a VMware import path, but since late August 2026 Broadcom no longer offers public downloads of VDDK, a library several tools use to read VMware disks. Check whether your tool needs it. - The migration risks that bite are Windows drivers, backup coverage, monitoring and Microsoft's per-core minimums, not disk formats. Pilot one workload and keep the old ESXi hosts until sign-off. - Running VMware on Azure, Google Cloud or AWS hardware buys time without retraining, but since November 2025 you bring your own VCF subscription, so you stay in the same licensing conversation. The credible VMware alternatives in 2026 are Red Hat OpenShift Virtualization, Proxmox VE, Microsoft Hyper-V with Azure Local, Nutanix AHV and XCP-ng, plus the option of keeping VMware on rented cloud hardware. The real differences are licensing model, skills, storage and migration tooling, not hypervisor features. If a Broadcom renewal quote is sitting in your inbox, this guide is the working document: what changed at VMware, how the alternatives compare, how to choose, and a migration plan with commands taken from the vendors' own documentation. The facts below come from vendor documentation read in October 2026 and are linked where they are used, with press reports attributed as such. ## Why so many teams are re-evaluating VMware in 2026 Broadcom [completed its acquisition of VMware on 22 November 2023](https://www.prnewswire.com/news-releases/broadcom-completes-acquisition-of-vmware-301995900.html). On 11 December 2023 it announced what its headline called a [dramatic simplification of the offer lineup and licensing model](https://www.broadcom.com/company/news/articles/cloud/vmware-by-broadcom-business-transformation): the end of perpetual licence sales, a move to subscriptions, and two flagship bundles, VMware Cloud Foundation (VCF) and vSphere Foundation (VVF), the latter aimed at mid-sized and smaller customers. In January 2024 VMware [confirmed the end of availability of perpetual licensing and SaaS services](https://blogs.vmware.com/cloud-foundation/2024/01/22/vmware-end-of-availability-of-perpetual-licensing-and-saas-services/), withdrawing the free vSphere Hypervisor and pointing customers with lighter needs to subscription editions of vSphere Standard and vSphere Essentials Plus Kit. The bundles are licensed per core, with a [minimum of 16 cores per physical CPU](https://knowledge.broadcom.com/external/article/313548/counting-cores-for-vmware-cloud-foundati.html), and the bundled vSAN capacity derives from the same cores: 1 TiB per core with VCF, 0.25 TiB per core with VVF. That KB states only the per-CPU floor, so confirm any order-level minimum with your reseller in writing. Lifecycle rules changed too. [vSphere 7.0 reached end of general support on 2 October 2025](https://knowledge.broadcom.com/external/article/415405/end-of-general-support-for-vsphere-70.html). Broadcom's vSphere 9.0 compatibility KB adds that [vSphere Standard and Enterprise Plus are supported only up to version 8 Update 3](https://knowledge.broadcom.com/external/article/409233/vsphere-essentials-plus-license-compatib.html) and that, at present, vSphere 9.0 features come only with VVF 9.0 and VCF 9.0. If a perpetual licence's support contract lapses, another KB says [ESXi hosts and vCenter Server continue to operate normally](https://knowledge.broadcom.com/external/article/429208/perpetual-license-functionality-after-su.html), but patch downloads and support requests end. That article covers perpetual licences only, so for a subscription read your contract's expiry terms. Two smaller points. The free hypervisor returned in April 2025 with the [ESXi 8.0 Update 3e release notes](https://techdocs.broadcom.com/us/en/vmware-cis/vsphere/vsphere/8-0/release-notes/esxi-update-and-patch-release-notes/vsphere-esxi-80u3e-release-notes.html); the [accompanying KB](https://knowledge.broadcom.com/external/article/399823/vmware-esxi-80-update-3e-now-available-a.html) lists its limits: 8 vCPUs per VM, no vCenter management, no vMotion, HA, DRS or API-based backup, and no Broadcom support. And in April 2026 Broadcom [wrote that some cloud service providers were affected by its partner program reorganisation](https://www.broadcom.com/company/news/articles/partners/vmware-cloud-service-provider-ecosystem-private-cloud-vcsp); if you buy VMware through a hosting provider, ask where yours stands. The newest change touches the exit itself. In late August 2026 the public download pages for VDDK, VMware's Virtual Disk Development Kit, stopped working. Red Hat's support article for its migration toolkit, [dated 27 August](https://access.redhat.com/solutions/7146995), says the old links are no longer active and tells customers to request the image from Broadcom Support. [Broadcom told TechTarget](https://www.techtarget.com/it-infrastructure/news/366648768/Broadcom-confirms-change-in-VMware-migration-tool-access) that the SDK licence has always limited VDDK to backup and recovery products from its technology-alliance partners. The migration plan below shows which routes avoid it. None of this makes vSphere worse software; it makes "what else could run these VMs" a normal procurement question. ## The VMware alternatives at a glance The table compares the shape of each option, not its price: list prices change, most real discounts are negotiated, and the interesting difference is the licensing model itself. Backup support matters more than most teams expect, so the table names what Veeam, a common incumbent, supports on each platform. | Platform | Licensing model | Live migration and HA | Storage model | Veeam support | Import from VMware | | ------------------------- | ------------------------------------------------------------------------------------------ | ---------------------------------------------------------------- | -------------------------------------------------- | ----------------------------------------- | ---------------------------------------------- | | OpenShift Virtualization | Red Hat subscription; Virtualization Engine edition per node, unlimited VMs, no containers | Live migration needs RWX storage; HA needs remediation operators | Persistent volumes; RWX and Block mode recommended | Native from v13.1 (KubeVirt plug-in) | Migration Toolkit for Virtualization | | Proxmox VE | Open source (AGPLv3); support subscriptions per CPU socket | Live migration; HA needs three nodes | Integrated Ceph and ZFS | Supported; Instant Recovery experimental | ESXi import wizard (tech preview) | | Microsoft Hyper-V | Included with Windows Server (per core, 16 minimum per server) | Live migration, failover clustering, Replica | Storage Spaces Direct or SAN | Supported | Preview importer removed September 2026; SCVMM | | Microsoft Azure Local | Flat rate per physical core, billed through Azure | Hyper-V clusters managed through Azure Arc | Hyperconverged; SAN billed at the L2 tier | [Supported](https://www.veeam.com/kb4047) | Azure Migrate (needs VDDK) | | Nutanix AHV | Included with Nutanix Cloud Infrastructure, per physical core | HA and live migration built in | Nutanix distributed storage fabric | Supported, with Instant Recovery | Nutanix Move | | XCP-ng with Xen Orchestra | Open source; Vates sells support | Live migration; HA needs shared storage | Storage repositories; XOSTOR for HCI | Fully supported in v13 | Xen Orchestra V2V | If the vocabulary in the table is new (hypervisor, live migration, quorum), our walkthrough on [setting up a virtual machine on your PC](https://computese.com/how-to-set-up-a-virtual-machine-on-your-pc/) shows the same moving parts at desktop scale. HPE VM Essentials, Harvester and plain KVM follow the main five. ## Red Hat OpenShift Virtualization: VMs beside containers OpenShift Virtualization [runs and manages traditional VMs alongside containers and serverless workloads](https://www.redhat.com/en/technologies/cloud-computing/openshift/virtualization) on one platform, as an included feature of OpenShift. It is built on [KubeVirt](https://www.redhat.com/en/topics/virtualization/what-is-kubevirt), so a VM becomes a Kubernetes object, managed with the same tools as everything else on the cluster. For estates that want the VMs without the rest of OpenShift, Red Hat sells the [OpenShift Virtualization Engine](https://www.redhat.com/en/technologies/cloud-computing/openshift/virtualization-engine): per-node licensing for bare metal, unlimited VMs, with containers requiring additional subscriptions. Red Hat's [4.22 documentation](https://docs.redhat.com/en/documentation/openshift_container_platform/4.22/html/virtualization/about) maps the vSphere concepts honestly. vMotion becomes live migration and requires the VM's volume to use the ReadWriteMany (RWX) access mode, which means shared storage. DRS becomes pod eviction policy plus the descheduler. NSX has no direct equivalent; the doc points to OVN-Kubernetes or certified third-party CNI plug-ins. HA needs remediation: without machine health checks, VMs set to `runStrategy: Always` are not relocated automatically from a failed node, so you add the Self Node Remediation or Fence Agents Remediation Operator. Tested maximums are 255 vCPUs and 6 TB of memory per VM. ![One server node drawn as a wide shelf: three virtual machine boxes on the left, a cluster of small container cubes on the right, joined by one thin orange line.](https://computese.com/images/blog/vmware-alternatives/openshift-vms-containers.4c85164e58-1536.webp) *On OpenShift, a VM is one more workload on the same node as your containers, managed with the same tools.* Migration tooling is the Migration Toolkit for Virtualization (MTV), which takes vSphere as a source provider; [version 2.12 (July 2026)](https://www.redhat.com/en/blog/storage-processing-accelerates-vm-migrations-migration-toolkit-virtualization-212) is the current release. [Version 2.11 (February 2026)](https://www.redhat.com/en/blog/migrate-your-vms-faster-migration-toolkit-virtualization-211) made storage offload generally available: instead of streaming disk data over the network, the toolkit asks the storage array to copy it, with arrays from Hitachi, Pure, NetApp and Dell. Version 2.12 extended offload to warm migrations, which copy data while the VM runs and end with a brief shutdown. Offload also matters for a second reason: Red Hat's VDDK article names it as the workaround while VDDK images are unavailable, if your array supports it. On backup, [Veeam protects OpenShift Virtualization natively from Data Platform v13.1](https://www.veeam.com/blog/openshift-vm-protection.html) through a KubeVirt plug-in, with no new licence. Choose it when containers are part of your plan anyway, or when you want a per-node rather than per-core licence model. Skip it if nobody on the team wants to learn Kubernetes: the [OpenShift vs Kubernetes](https://computese.com/openshift-vs-kubernetes/) comparison explains what that learning curve looks like. We have run this class of platform move in production: the [banking platform migration](https://computese.com/work/banking-platform-migration/) case study moved a bank's digital channels from OpenShift 3 to OpenShift 4 inside one short change window, with the runbook discipline that a VM estate deserves too. ## Proxmox VE: the open-source workhorse Proxmox VE is the open-source all-rounder: a [KVM and LXC platform](https://pve.proxmox.com/pve-docs/chapter-pve-intro.html) with its source under the AGPLv3 licence, plus clustering, live migration, HA, software-defined networking and integrated storage. Version 9.0 ([5 August 2025](https://www.proxmox.com/en/about/company-details/press-releases/proxmox-virtual-environment-9-0)) moved to Debian 13 with QEMU 10.0.2, LXC 6.0.4, Ceph Squid 19.2.3 and ZFS 2.3.3, and the project reported more than 1.6 million hosts at that launch. The current release, [9.2 (21 May 2026)](https://proxmox.com/en/about/company-details/press-releases/proxmox-virtual-environment-9-2), added a dynamic load balancer (the closest thing to DRS) and an HA arm/disarm switch for maintenance. The software is free; support is the paid part. [Proxmox's pricing page](https://www.proxmox.com/en/products/proxmox-virtual-environment/pricing) (October 2026) states the model plainly: a subscription per occupied CPU socket, cores do not change the price, and every feature, including HA and live migration, is available at every level. Plans differ by support entitlement, ticket allowance and response time. The migration path is the part to verify in your lab. The [Proxmox VE Administration Guide](https://pve.proxmox.com/pve-docs/pve-admin-guide.html) describes an import wizard that connects to an ESXi host or vCenter as if it were a storage backend and imports guests whole, with most of the config mapped across. Three caveats from the same guide: the wizard arrived in the 8.2 development cycle and is still marked tech preview; it is tested with ESXi 6.5 through 8.0; and guests on vSAN storage cannot be imported directly, while importing through vCenter is 5 to 10 times slower than going straight to the host. The guide also sets the HA floor: at least three nodes for reliable quorum, so budget a third host or a quorum device from day one. Backup is covered with an asterisk: [Veeam protects Proxmox VE](https://helpcenter.veeam.com/docs/vbr/wn/hypervisor_protection.html?ver=13), while its v13.1 notes still mark Instant Recovery to Proxmox as experimental. Choose Proxmox when the team is comfortable on Linux and wants no licence meter running; plan for a smaller enterprise ecosystem around it than VMware or Microsoft offer. ## Microsoft Hyper-V and Azure Local: the Windows-centric path Hyper-V is the quiet incumbent: [included with Windows Server](https://learn.microsoft.com/en-us/windows-server/virtualization/hyper-v/overview) (2025, 2022, 2019, 2016), with live migration, failover clustering and Hyper-V Replica built in, on Storage Spaces Direct or an iSCSI or Fibre Channel SAN. If your estate is mostly Windows and your team already holds Microsoft certifications, the operational distance from vSphere is short. The licensing rhymes with Broadcom's. [Windows Server 2025 is licensed per core](https://www.microsoft.com/licensing/guidance/Windows-Server-2025): every physical core on the server, with a minimum of 8 core licences per processor and 16 per server. Datacenter then allows unlimited VMs on that host; Standard allows two, and stacking more means re-licensing the same cores. Licensing by VM exists, but only with subscription licences or active Software Assurance, and CALs still apply on top. Hyper-V lowers the hypervisor line item; it does not remove per-core arithmetic. The import story changed this autumn. Microsoft's preview [VM Conversion extension for Windows Admin Center](https://learn.microsoft.com/en-us/windows-server/manage/windows-admin-center/use/vm-conversion-extension-overview), which moved VMware VMs to Hyper-V, was removed from the extension feed. The [Windows Admin Center 2610 announcement](https://techcommunity.microsoft.com/blog/windows-admin-center-blog/two-modes-combined-windows-admin-center-version-2610-is-now-in-public-preview/4556863) (September 2026) ties the removal to Broadcom's change to VDDK availability and, for now, points customers to System Center Virtual Machine Manager for Windows Server with Hyper-V, or to Azure Migrate (agent-based) for Azure. Azure Local is the same Hyper-V stack as a Microsoft-managed hybrid product, [controlled through Azure Arc](https://learn.microsoft.com/en-us/azure/azure-local/overview) ([Microsoft renamed Azure Stack HCI to Azure Local](https://learn.microsoft.com/en-us/azure/azure-local/security-book/overview)). Its [billing model is a flat rate per physical processor core](https://learn.microsoft.com/en-us/azure/azure-local/concepts/billing), regardless of how many VMs run, and in the standard model the cluster must connect to Azure at least once every 30 days. Sites that cannot do that have a separate path, [disconnected operations](https://learn.microsoft.com/en-us/azure/azure-local/manage/disconnected-operations-overview) with a local control plane, but it needs an eligible Microsoft agreement, a support plan, a valid business need and extra hardware, and it is billed on its own tier. [Azure Migrate](https://learn.microsoft.com/en-us/azure/azure-local/migrate/migration-azure-migrate-vmware-overview) moves VMware VMs to Azure Local without agents inside them, but its [replication step](https://learn.microsoft.com/en-us/azure/azure-local/migrate/migrate-vmware-replicate) needs you to supply the VDDK package and warns that access to it might be restricted. Choose this path when you are Windows-heavy, already pay for Windows Server, and want Azure's management plane; skip it if per-core licensing is precisely what you are trying to escape. ## The other serious contenders ### Nutanix AHV Nutanix sells hyperconverged infrastructure, and its hypervisor is not the product: [AHV is not sold or licensed standalone](https://www.nutanix.com/products/ahv), it is included with a Nutanix Cloud Infrastructure (NCI) licence, which Nutanix [applies per physical CPU core](https://www.nutanix.com/products/cloud-platform/software-options). HA, live migration and the distributed storage fabric are built in, and [Nutanix Move](https://www.nutanix.com/products/move) migrates VMs from vSphere, Hyper-V and AWS with a pre-staged, low-downtime cutover. Veeam supports AHV, including as an Instant Recovery target. Choose Nutanix when you want an appliance-like platform with one vendor accountable for the whole stack; the trade is that you replace one proprietary per-core relationship with another. ### XCP-ng with Xen Orchestra XCP-ng is the open-source Xen platform stewarded by Vates: [XCP-ng 8.3, released on 7 October 2024, became the long-term support release on 16 June 2025](https://docs.xcp-ng.org/releases/release-8-3/), managed at scale through [Xen Orchestra](https://xcp-ng.org/), with commercial support from Vates. [High availability](https://docs.xcp-ng.org/management/ha/) restarts VMs from a failed host on the others, but it needs shared storage (iSCSI, Fibre Channel or NFS) for its heartbeat, and Vates recommends at least three hosts. The migration story improved recently: [Xen Orchestra 5.110 introduced a new VMware importer](https://docs.xen-orchestra.com/guides/v2v-migration-guide) that reads VM disks straight off ESXi over VMware's NFC protocol, without any VMware-provided library, so the VDDK change does not affect it. It does a warm migration with the source VM running for most of the transfer, and today it lives in the older XO 5 interface rather than the new XO 6. On backup, [XCP-ng is a fully supported hypervisor in Veeam Backup & Replication v13](https://helpcenter.veeam.com/docs/vbr/wn/hypervisor_protection.html?ver=13). Choose it for cost-sensitive virtualization with a Xen heritage; expect a smaller vendor and ecosystem than Red Hat, Microsoft or Nutanix. ### HPE VM Essentials, Harvester and plain KVM Three more options are worth a lab bench. [HPE Morpheus VM Essentials](https://www.hpe.com/us/en/products/software/morpheus-software/virtualization.html) pairs a KVM-based hypervisor (HVM) with a console that manages VMware and HVM clusters side by side, includes a built-in ESXi-to-HVM migration feature, and is licensed per socket with support included. [Harvester](https://harvesterhci.io/) is SUSE's open-source HCI, built on Linux, KVM, Kubernetes, KubeVirt and Longhorn, for teams that want VMs and cloud-native workloads on one stack. And plain [KVM](https://www.linux-kvm.org/page/Main_Page), in mainline Linux since 2.6.20, is the substrate under OpenShift Virtualization, Proxmox, HVM and Harvester; running it bare with libvirt suits a small, Linux-fluent team that wants to assemble its own tooling. ## Keep VMware, leave the hardware Sometimes the right answer to a hardware refresh and a renewal at the same time is to move VMware itself. [Azure VMware Solution](https://learn.microsoft.com/en-us/azure/azure-vmware/introduction) gives you vSphere clusters on dedicated bare-metal Azure infrastructure, 3 to 16 hosts per cluster, with vCenter, vSAN and NSX provisioned for you. [Google Cloud VMware Engine](https://docs.cloud.google.com/vmware-engine/docs/overview) is a fully managed VMware platform on Google Cloud bare metal. And [Amazon Elastic VMware Service went generally available on 5 August 2025](https://aws.amazon.com/about-aws/whats-new/2025/08/aws-general-availability-amazon-elastic-evs/): VCF inside your VPC, with license portability so you can bring VCF licences you already own. This path keeps skills and tooling intact and turns the migration into an evacuation: same platform, new landlord. What it does not change is the licensing conversation. Microsoft's documentation says that in November 2025 Broadcom changed its policies [across all hyperscaler platforms](https://learn.microsoft.com/en-us/azure/azure-vmware/license-included-service-retirement) to require customers to bring their own portable VCF licence; Azure's pay-as-you-go license-included SKUs retire on 31 October 2026, and all license-included services end on 30 August 2027. Google Cloud [describes the same change](https://cloud.google.com/blog/products/compute/broadcom-vcf-licensing-changes-for-vmware-engine) for VMware Engine from 1 November 2025. Either way you buy VCF from Broadcom. In the language of our [cloud migration strategy](https://computese.com/cloud-migration-strategy/) guide, this is relocate, not rehost, and it pairs well with a slower plan to replatform the workloads that should not stay VMs at all. ## How to choose between them Ask five questions, in this order: 1. **What does the team already run?** Windows Server skills point at Hyper-V; Linux skills point at Proxmox, XCP-ng or KVM; Kubernetes skills (or appetite) point at OpenShift or Harvester. The cheapest platform is the one your team can operate at 3 a.m. 2. **What is the workload mix?** In a Windows-heavy estate, Windows Server licensing may remain the largest line whichever hypervisor you pick, so model it before you count savings. 3. **Are containers coming?** If yes, OpenShift Virtualization (or Harvester) lets VMs and containers share one platform. If no, a Kubernetes-based hypervisor adds a control plane you did not need. 4. **What storage exists?** A good SAN suits Proxmox, XCP-ng and Hyper-V, while OpenShift needs RWX-capable shared storage for live migration. No shared storage pushes you toward HCI designs (Nutanix, Azure Local, XOSTOR, Harvester); on the others, check how live migration behaves without it. 5. **What support do you need?** Open source with optional support subscriptions (Proxmox, XCP-ng) suits self-sufficient teams. A single accountable vendor (Nutanix, Microsoft, HPE, Red Hat) suits lean teams and strict compliance regimes. And if some workloads should not be VMs at all, say so in the plan. Our piece on [where serverless computing is heading](https://computese.com/the-future-of-serverless-computing-streamlining/) is the case for retiring a workload instead of moving it. ## A migration plan that survives production Every target above has an import path, so the mechanics are solved; the discipline is what you bring. Start with one check that did not exist a few weeks ago. > [!WARNING] > Ask every migration tool vendor whether the tool needs VDDK. Azure Migrate's Azure Local steps require the VDDK package, Red Hat calls the VDDK image the recommended route for its migration toolkit, and Microsoft pulled its preview Hyper-V importer over it. If you already hold a VDDK package, keep it somewhere safe; if not, request one from Broadcom Support before the pilot, or choose a route that does not need it. The sequence that works: 1. **Inventory.** Export the estate with PowerCLI, which Broadcom itself ships as the counting tool for its per-core KB. The [`Get-VM` cmdlet](https://developer.broadcom.com/powercli/latest/vmware.vimautomation.core/commands/get-vm/) retrieves every VM on a vCenter system, and PowerShell's [`Export-Csv`](https://learn.microsoft.com/en-us/powershell/module/microsoft.powershell.utility/export-csv) turns the result into a sheet: ```console Connect-VIServer vcenter.example.com Get-VM | Export-Csv .\vms-inventory.csv ``` [RVTools](https://www.dell.com/en-us/shop/vmware/sl/rvtools) is the common free alternative; Dell Technologies says it acquired the tool in 2023 and has always made it available at no cost. Add hosts, datastores, port groups and VLANs, tag each VM with an owner and a criticality, and count physical cores per host, because that number prices what you pay Broadcom today and what per-core alternatives would charge. 2. **Classify.** Sort VMs into retire, retain and move, then group the movers by dependency so that VMs which talk to each other travel in the same wave. Pick a low-risk pilot that is still real enough to exercise backup, monitoring and a Windows guest. 3. **Build the target beside the source.** Stand up the new cluster with production-grade settings (HA, backup jobs, monitoring) before the first VM moves. A pilot on a hand-built snowflake host proves nothing. 4. **Convert.** Prefer the platform's own importer: MTV for OpenShift, the import wizard for Proxmox, V2V for XCP-ng, Move for Nutanix, the built-in feature for HPE. The generic route for KVM targets is [virt-v2v](https://libguestfs.org/virt-v2v.1.html), which converts a whole guest and installs the virtio drivers for you. Its [VMware input page](https://libguestfs.org/virt-v2v-input-vmware.1.html) lists several ways to read the disks, and the NFC transport needs no VDDK: ```console virt-v2v -ic 'vpx://root@vcenter.example.com/Datacenter/esxi' \ -it nfc web-01 -o local -os /var/tmp ``` The documentation's own example adds `?no_verify=1` to the URI, which bypasses certificate verification; leave it out when your vCenter certificate is trusted. The manual route is [qemu-img](https://www.qemu.org/docs/master/tools/qemu-img.html) for a disk you have copied from the datastore or exported, then an import into the target. Both commands below are from the tools' official documentation: ```console qemu-img convert -f vmdk -O qcow2 web-01.vmdk web-01.qcow2 qm disk import 104 web-01.qcow2 local-lvm ``` 5. **Fix the guest.** On KVM-based platforms, Windows guests need the [virtio-win drivers](https://github.com/virtio-win/kvm-guest-drivers-windows) to use paravirtual disk and network devices, and VMware Tools comes out. [Proxmox's migration guide](https://pve.proxmox.com/wiki/Migrate_to_Proxmox_VE) advises installing the VirtIO drivers before migrating, falling back to an IDE or SATA boot disk if a migrated VM will not boot, and removing the old hypervisor's guest tools beforehand. Map networks and VLANs explicitly, set the MAC address by hand where a DHCP reservation or firewall rule depends on it, and check whether Windows and licensed applications need reactivation on the new virtual hardware. ![A virtual machine box crosses a bridge from one hypervisor platform to another while a gear-shaped driver part is lifted out and a new one, drawn in orange, is lowered in.](https://computese.com/images/blog/vmware-alternatives/windows-driver-swap.b653b5f6ac-1536.webp) *Converting the disk format is the mechanical step; the driver swap is the part that decides whether Windows boots on the new platform.* 6. **Cut over in waves, with a warm rollback.** Move the pilot, run it for a full business cycle, then move dependency groups one wave at a time. Keep the old ESXi hosts powered and intact, and the licence valid, until each wave is signed off. Never run the same VM on both platforms at once; Proxmox's guide warns that doing so corrupts the disk. On a perpetual licence with lapsed support the hosts keep operating; on a subscription, align its end date with the final sign-off so it cannot expire mid-program. ![A line of virtual machine boxes moves from a source server rack on the left to a target rack on the right; beneath them, one thick orange arrow curves back to the source as the way back.](https://computese.com/images/blog/vmware-alternatives/rollback-arrow.d4a59a1a5a-1536.webp) *The way back stays warm until the wave is signed off; only then does the old host give up its VMs.* 7. **Decommission on evidence.** After the last wave: archive the exports and logs, end the subscriptions and support renewals, wipe the hosts, and release the hardware. Do it in that order. ## The pitfalls that actually bite | Pitfall | Why it bites | What prevents it | | ----------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | | Drivers and guest tools | A converted Windows VM has no VirtIO driver until you install one, and VMware Tools does nothing on the new platform | Install virtio-win first, or let virt-v2v do it; remove VMware Tools and install the target's guest agent | | VDDK dependency | The tool you chose may need a library Broadcom no longer publishes | Ask the vendor before the pilot; prefer NFC, storage offload or a tool with its own reader | | Backup blind spot | vSphere backup jobs rely on vSphere APIs that the new platform does not offer | Confirm your backup vendor supports the target before wave one, rebuild jobs, and test a restore, as our [backup strategy guide](https://computese.com/how-to-create-a-secure-backup-strategy-for-your-data/) insists | | Monitoring gaps | vCenter alarms and capacity views do not carry over to the new platform | Rebuild alerts and dashboards during the pilot, not after the last wave | | Microsoft per-core maths | Hyper-V does not escape per-core licensing: 16-core minimum per server, two VMs per Standard licence | Model Datacenter versus Standard before the move, using [Microsoft's own guidance](https://www.microsoft.com/licensing/guidance/Windows-Server-2025) | | Network and licence surprises | The network adapter changes, and a port group is not a VLAN is not a Kubernetes network attachment; licences bound to a MAC address or machine identity can stop working | Map networks in the migration tool, test with the pilot, and put licence checks in each wave's runbook | ## Where to start Three moves fit in this quarter. First, run the inventory and the core count, so the renewal quote and every alternative are priced against the same facts. Second, shortlist two platforms from the table and give each one a week in the lab with the same pilot VM, including a check of how its importer reads VMware disks. Third, write the wave plan with its rollback rules before you touch production. If you want a second pair of eyes on the plan, our [cloud transformation](https://computese.com/services/cloud-transformation/) team plans and carries out migration waves with a rehearsed rollback at each stage, and hands the platform over with runbooks and trained operators. Our [hosting and maintenance](https://computese.com/services/hosting-maintenance/) team monitors and patches servers that stay on your own premises, with backups proven by restore drills. Either way, put the renewal date and the end of your support contract on the plan before the first wave. ## Key terms - **Hypervisor**: The software layer that runs virtual machines on physical hardware. ESXi, KVM, Xen and Hyper-V are hypervisors; vCenter and Xen Orchestra are management layers above them. - **VMware Cloud Foundation (VCF)**: Broadcom's flagship hybrid-cloud bundle: vSphere, vSAN and NSX with a management and orchestration suite, sold only as a per-core subscription since the end of perpetual licensing. - **vSphere Foundation (VVF)**: The second Broadcom subscription bundle, aimed at traditional vSphere environments and mid-sized to smaller customers. It includes vSphere, vSAN capacity at 0.25 TiB per core, and operations tooling. - **Per-core licensing**: Paying for software by the number of CPU cores it runs on. Broadcom and Microsoft both publish minimums alongside it, so high core-count servers raise the licence count on either platform. - **KubeVirt**: The open-source project that lets Kubernetes run and manage virtual machines, which then behave like pods. Red Hat OpenShift Virtualization and SUSE Harvester are both built on it. - **Live migration**: Moving a running virtual machine from one host to another without downtime. VMware calls it vMotion; the platforms below have an equivalent, though OpenShift Virtualization needs ReadWriteMany shared storage for it. - **VirtIO**: The standard paravirtual device drivers for KVM guests. Windows VMs moved to a KVM-based platform such as Proxmox VE need the virtio-win drivers to use fast storage and network devices. - **VDDK**: VMware's Virtual Disk Development Kit, a proprietary library that backup and migration tools use to read VMware disks. Broadcom owns it, and its public downloads disappeared in late August 2026. ## Common questions ### Is VMware still selling perpetual licences? No. Broadcom announced the end of perpetual licence sales on 11 December 2023, and VMware confirmed the end of availability in January 2024. New purchases are subscriptions: the VMware Cloud Foundation and vSphere Foundation bundles, plus the standalone vSphere editions that Broadcom's knowledge base still lists. Existing perpetual licences keep running, but without a current support contract you lose patch downloads and support requests. ### Is the free ESXi hypervisor still available? Broadcom withdrew the free vSphere Hypervisor in January 2024, then released a free ESXi 8.0 Update 3e in April 2025, and its knowledge base still documents that download. The free licence caps VMs at 8 vCPUs, cannot be managed by vCenter, has no vMotion, HA, DRS or API-based backup, and carries no Broadcom support. The release notes describe it as non-production. ### Which VMware alternative is easiest to migrate to? There is no honest easiest, because the answer depends on your team. Proxmox VE and XCP-ng feel closest to the vSphere host-and-cluster model. Nutanix AHV and HPE VM Essentials are commercial platforms with their own migration tooling. OpenShift Virtualization fits when containers are part of the destination. Hyper-V suits Windows-heavy estates that already license Windows Server. ### Can I move VMware VMs to another hypervisor without rebuilding them? Yes. The virtual disks carry over. OpenShift has the Migration Toolkit for Virtualization, Proxmox VE has an ESXi import wizard, Xen Orchestra has a V2V importer, and Nutanix has Move. The generic path is virt-v2v for a whole guest or qemu-img to convert a VMDK to qcow2 or raw. What needs attention is drivers and guest tools, especially on Windows, and how your tool reads the VMware disks. ### What happens to my VMs if I let VMware support lapse? On a perpetual licence they keep running. Broadcom's own knowledge base states that ESXi hosts and vCenter Server continue to operate normally when the support contract expires, but you lose patch and security-update downloads and the ability to open support requests. That is a freeze, not a shutdown. The article covers perpetual licences only, so for a subscription read the term and expiry behaviour in your contract. ### Is OpenShift Virtualization a drop-in replacement for vSphere? No. The concepts map (vMotion to live migration, DRS to the descheduler), but live migration requires ReadWriteMany shared storage, there is no direct NSX equivalent, and day-to-day operations are Kubernetes operations. It is a strong destination when you want VMs and containers on one platform, not a like-for-like vSphere clone. ### Why does my migration tool ask for VDDK, and where do I get it? VDDK is the Virtual Disk Development Kit, which some tools use to read VMware disks. Broadcom's public download pages stopped working in late August 2026, and Red Hat tells Migration Toolkit users to request the image from Broadcom Support. Alternatives are a tool that does not need it, such as Xen Orchestra's importer, array-level storage offload in the Migration Toolkit, or virt-v2v's NFC, SSH and OVA routes. ## Sources 1. [Broadcom Completes Acquisition of VMware](https://www.prnewswire.com/news-releases/broadcom-completes-acquisition-of-vmware-301995900.html), Broadcom, via PRNewswire 2. [VMware by Broadcom Dramatically Simplifies Offer Lineup and Licensing Model](https://www.broadcom.com/company/news/articles/cloud/vmware-by-broadcom-business-transformation), Broadcom 3. [VMware End Of Availability of Perpetual Licensing and SaaS Services](https://blogs.vmware.com/cloud-foundation/2024/01/22/vmware-end-of-availability-of-perpetual-licensing-and-saas-services/), VMware Cloud Foundation Blog 4. [Counting Cores for VMware Cloud Foundation and vSphere Foundation and TiBs for vSAN (KB 313548)](https://knowledge.broadcom.com/external/article/313548/counting-cores-for-vmware-cloud-foundati.html), Broadcom Knowledge Base 5. [End of General Support for vSphere 7.0 (KB 415405)](https://knowledge.broadcom.com/external/article/415405/end-of-general-support-for-vsphere-70.html), Broadcom Knowledge Base 6. [vSphere Essentials Plus license compatibility with vSphere 9.0 (KB 409233)](https://knowledge.broadcom.com/external/article/409233/vsphere-essentials-plus-license-compatib.html), Broadcom Knowledge Base 7. [Perpetual License Functionality After Support and Subscription Expiration (KB 429208)](https://knowledge.broadcom.com/external/article/429208/perpetual-license-functionality-after-su.html), Broadcom Knowledge Base 8. [VMware ESXi 8.0 Update 3e now available as a Free Hypervisor (KB 399823)](https://knowledge.broadcom.com/external/article/399823/vmware-esxi-80-update-3e-now-available-a.html), Broadcom Knowledge Base 9. [VMware ESXi 8.0 Update 3e Release Notes](https://techdocs.broadcom.com/us/en/vmware-cis/vsphere/vsphere/8-0/release-notes/esxi-update-and-patch-release-notes/vsphere-esxi-80u3e-release-notes.html), Broadcom TechDocs 10. [We're Creating an Economically Viable Cloud Service Provider Ecosystem for Our Customers](https://www.broadcom.com/company/news/articles/partners/vmware-cloud-service-provider-ecosystem-private-cloud-vcsp), Broadcom 11. [Broadcom confirms it revoked public VMware migration tool access](https://www.techtarget.com/it-infrastructure/news/366648768/Broadcom-confirms-change-in-VMware-migration-tool-access), TechTarget 12. [Unable to download VMware VDDK images for Migration Toolkit for Virtualization (MTV)](https://access.redhat.com/solutions/7146995), Red Hat Customer Portal 13. [Two modes combined: Windows Admin Center version 2610 is now in public preview](https://techcommunity.microsoft.com/blog/windows-admin-center-blog/two-modes-combined-windows-admin-center-version-2610-is-now-in-public-preview/4556863), Microsoft Tech Community 14. [Red Hat OpenShift Virtualization](https://www.redhat.com/en/technologies/cloud-computing/openshift/virtualization), Red Hat 15. [Red Hat OpenShift Virtualization Engine](https://www.redhat.com/en/technologies/cloud-computing/openshift/virtualization-engine), Red Hat 16. [What is KubeVirt?](https://www.redhat.com/en/topics/virtualization/what-is-kubevirt), Red Hat 17. [OpenShift Container Platform 4.22: About virtualization](https://docs.redhat.com/en/documentation/openshift_container_platform/4.22/html/virtualization/about), Red Hat Documentation 18. [Migrate your VMs faster with the migration toolkit for virtualization 2.11](https://www.redhat.com/en/blog/migrate-your-vms-faster-migration-toolkit-virtualization-211), Red Hat Blog 19. [Storage processing accelerates VM migrations in the migration toolkit for virtualization 2.12](https://www.redhat.com/en/blog/storage-processing-accelerates-vm-migrations-migration-toolkit-virtualization-212), Red Hat Blog 20. [Veeam Data Platform for OpenShift VM Protection](https://www.veeam.com/blog/openshift-vm-protection.html), Veeam Blog 21. [Proxmox Virtual Environment 9.0 with Debian 13 released](https://www.proxmox.com/en/about/company-details/press-releases/proxmox-virtual-environment-9-0), Proxmox 22. [Proxmox Virtual Environment 9.2 with Dynamic Load Balancer released](https://proxmox.com/en/about/company-details/press-releases/proxmox-virtual-environment-9-2), Proxmox 23. [Proxmox VE subscriptions and pricing](https://www.proxmox.com/en/products/proxmox-virtual-environment/pricing), Proxmox 24. [Proxmox VE Administration Guide: Introduction](https://pve.proxmox.com/pve-docs/chapter-pve-intro.html), Proxmox 25. [Proxmox VE Administration Guide](https://pve.proxmox.com/pve-docs/pve-admin-guide.html), Proxmox 26. [Migrate to Proxmox VE](https://pve.proxmox.com/wiki/Migrate_to_Proxmox_VE), Proxmox VE Wiki 27. [Veeam Backup & Replication 13.1 What's New: Hypervisor Protection](https://helpcenter.veeam.com/docs/vbr/wn/hypervisor_protection.html?ver=13), Veeam Help Center 28. [KB4047: Veeam Support for Azure Local](https://www.veeam.com/kb4047), Veeam 29. [Hyper-V virtualization in Windows Server and Windows](https://learn.microsoft.com/en-us/windows-server/virtualization/hyper-v/overview), Microsoft Learn 30. [Windows Server 2025 Licensing Guidance](https://www.microsoft.com/licensing/guidance/Windows-Server-2025), Microsoft 31. [VM Conversion extension in Windows Admin Center (Preview)](https://learn.microsoft.com/en-us/windows-server/manage/windows-admin-center/use/vm-conversion-extension-overview), Microsoft Learn 32. [Azure Local security book overview](https://learn.microsoft.com/en-us/azure/azure-local/security-book/overview), Microsoft Learn 33. [What is Azure Local?](https://learn.microsoft.com/en-us/azure/azure-local/overview), Microsoft Learn 34. [Azure Local billing and payment](https://learn.microsoft.com/en-us/azure/azure-local/concepts/billing), Microsoft Learn 35. [Disconnected operations for Azure Local](https://learn.microsoft.com/en-us/azure/azure-local/manage/disconnected-operations-overview), Microsoft Learn 36. [Use Azure Migrate to move VMware VMs to Azure Local](https://learn.microsoft.com/en-us/azure/azure-local/migrate/migration-azure-migrate-vmware-overview), Microsoft Learn 37. [Replicate VMware VMs to Azure Local using Azure Migrate](https://learn.microsoft.com/en-us/azure/azure-local/migrate/migrate-vmware-replicate), Microsoft Learn 38. [AHV: Virtualization Solution for Enterprise](https://www.nutanix.com/products/ahv), Nutanix 39. [Nutanix Move: application migration](https://www.nutanix.com/products/move), Nutanix 40. [Nutanix Cloud Platform software options](https://www.nutanix.com/products/cloud-platform/software-options), Nutanix 41. [XCP-ng 8.3 LTS](https://docs.xcp-ng.org/releases/release-8-3/), XCP-ng Documentation 42. [VMware migration (V2V)](https://docs.xen-orchestra.com/guides/v2v-migration-guide), Xen Orchestra Documentation 43. [XCP-ng: XenServer based, community powered](https://xcp-ng.org/), XCP-ng (Vates) 44. [High availability](https://docs.xcp-ng.org/management/ha/), XCP-ng Documentation 45. [Introduction to Azure VMware Solution](https://learn.microsoft.com/en-us/azure/azure-vmware/introduction), Microsoft Learn 46. [Azure VMware Solution license-included service retirement, using portable VCF and other options](https://learn.microsoft.com/en-us/azure/azure-vmware/license-included-service-retirement), Microsoft Learn 47. [Google Cloud VMware Engine overview](https://docs.cloud.google.com/vmware-engine/docs/overview), Google Cloud Documentation 48. [Broadcom's VMware license changes as they relate to Google Cloud VMware Engine](https://cloud.google.com/blog/products/compute/broadcom-vcf-licensing-changes-for-vmware-engine), Google Cloud Blog 49. [AWS announces general availability of Amazon Elastic VMware Service](https://aws.amazon.com/about-aws/whats-new/2025/08/aws-general-availability-amazon-elastic-evs/), AWS 50. [QEMU disk image utility](https://www.qemu.org/docs/master/tools/qemu-img.html), QEMU Project 51. [virt-v2v: Convert a guest to run on KVM](https://libguestfs.org/virt-v2v.1.html), libguestfs 52. [virt-v2v-input-vmware: Using virt-v2v to convert guests from VMware](https://libguestfs.org/virt-v2v-input-vmware.1.html), libguestfs 53. [KVM/QEMU Windows guest drivers (virtio-win)](https://github.com/virtio-win/kvm-guest-drivers-windows), virtio-win project on GitHub 54. [Get-VM Command Reference](https://developer.broadcom.com/powercli/latest/vmware.vimautomation.core/commands/get-vm/), Broadcom Developer (VMware PowerCLI) 55. [Export-Csv](https://learn.microsoft.com/en-us/powershell/module/microsoft.powershell.utility/export-csv), Microsoft Learn (PowerShell) 56. [RVTools: free VMware infrastructure management tool](https://www.dell.com/en-us/shop/vmware/sl/rvtools), Dell Technologies 57. [Harvester: open-source hyperconverged infrastructure](https://harvesterhci.io/), Harvester (SUSE) 58. [Kernel Virtual Machine](https://www.linux-kvm.org/page/Main_Page), KVM Project 59. [HPE Morpheus Software: virtualization and VM Essentials](https://www.hpe.com/us/en/products/software/morpheus-software/virtualization.html), HPE