How to Leave VMware Without Sacrificing Your Current Service Level, and Start a Relationship With a New Technology Partner
Introduction
It’s a well-known concern in the market: Broadcom’s acquisition of VMware had a significant impact on the cost of keeping a company’s current environment running. Policy changes hit the TCO of production environments directly.
For an SMB, keeping that relationship going is close to impossible, and many companies today are stuck on perpetual licenses with no support — something they’re going to need to walk away from sooner or later. The alternatives most commonly discussed for this profile are Proxmox and, why not, Hyper-V, which at the end of the day might be the better fit for this kind of customer.
Now, if we’re talking about an Enterprise environment, the safe path is probably still renewing VMware under the new policies. The landscape is aggressive: Broadcom tends to push a High-End product bundled with everything VMware can do, whether you need it or not (for example, Tanzu is part of the license even if you have no need for Kubernetes). It also requires a minimum of 72 licensed cores to get started, and pushes customers into three-year commitments to access a reasonable price — corrupting, so to speak, the “on-demand” nature that cloud consumption models are supposed to have. That’s the deal if you want that service level… until now.
The news: Nutanix opened things up together with Dell, and now Acropolis (AHV) can run on top of PowerStore. (see the news)
Why a Change Is Worth Considering, Beyond the Price of the Solution
Not every technology company acquisition is chasing the same outcome. When EMC bought VMware in 2004 for around $625 million, it let VMware operate autonomously and kept investing in its growth — three and a half years later, VMware went public at a $19.1 billion valuation, with EMC retaining most of the equity. That was a growth bet: the buyer made money by growing the value of the product.
There’s another pattern, documented across several acquisitions of mature, market-leading software: buy a product with a captive customer base — customers for whom switching platforms is expensive and slow — and use that captivity to raise revenue immediately, instead of investing in improving the product. This is what’s known as exploiting a “captive customer.” The market is slow to react precisely because the cost of switching platforms is high — and that delay is part of the calculation, not a side effect.
This is a documented practice, not a suspicion: it’s on record in Broadcom’s acquisitions of CA Technologies (2018) and Symantec’s enterprise division (2019), and a good part of the market has been pointing out the same pattern in Broadcom’s acquisition of VMware. You don’t need to share that reading to get the underlying point: once the owner of a platform no longer needs to win the customer over, the customer loses the one real guarantee they had — that it’s in someone else’s interest for the platform to keep getting better.
Three Solutions, Each Solving Its Own Technical Front, Now Coming Together
Nutanix virtualizes and runs the infrastructure. PowerStore secures the operational data. Veeam manages data continuity.
Nutanix AHV, Dell PowerStore and Veeam are each, on their own, mature products proven in enterprise environments. What changed just a few weeks ago is that they can now be thought of as one stack: Nutanix added external storage support for AHV, and Dell was the first to integrate PowerStore on top of it.
This combination is worth exploring — but honestly: this is a July 2026 development, and not every fine-grained integration between the three parties is fully documented or tested together yet. What follows is the architecture and the reasoning; further down we flag specifically what’s confirmed and what still needs validation.
The Typical Scenario
Many organizations today keep their business continuity running on a familiar combination: VMware vSphere as the hypervisor, a high-end external storage array, VMware Site Recovery Manager (SRM) to orchestrate recovery between sites, and Veeam Backup & Replication for daily protection. It works, but it comes with a growing cost — in licensing, in operational complexity, and in dependency on a single virtualization stack.
The question starting to come up at IT tables is a direct one: can you change hypervisors without giving up the same level of protection and orchestration a high-end external storage array provides today?
What’s New: AHV Can Now Run on External Storage
Until very recently, Nutanix AHV only worked with Nutanix’s own distributed storage (Nutanix Cloud Infrastructure). That changed: with Nutanix Cloud Infrastructure (NCI) 7.6, Nutanix added official external storage support for AHV. Dell was the first partner to integrate this capability, first with PowerFlex (April 2026, including synchronous replication) and now also with Dell PowerStore, validated with a full lifecycle (Day 0 through Day 2), non-disruptive upgrades, and independent scaling of compute and storage.
In practice: VM virtual disks on AHV map to volumes on PowerStore over NVMe/TCP, with Nutanix’s Stargate service mediating I/O. Snapshot management, protection policies, and replication are handled from Prism Central, with three modes: Metro (zero RPO with automatic failover), synchronous (zero RPO), and asynchronous (on a scheduled interval) — the same options Nutanix already offers with its own native storage.
This opens a concrete door: the hypervisor and the storage stop being tied to each other. You can migrate from vSphere to AHV while adding PowerStore as external storage — starting from a PowerStore that’s already in production, or bringing it in as part of the move.
The Three Pillars of the Solution
Nutanix AHV provides the hypervisor — simpler to operate and license than vSphere, with Prism as a single management plane for compute, networking (Flow), and now external storage as well.
Dell PowerStore remains the storage layer: a unified NVMe all-flash array (QLC models like the 5200Q scale up to 25PB effective per cluster), with inline deduplication and compression backed by a 6:1 data reduction guarantee, and snapshots that are immutable by design. Since PowerStore 3.5 there are also “secure snapshots,” which cannot be deleted manually — not by an administrator, and not by an attacker with compromised credentials — until their retention period expires.
Veeam covers protection and recovery orchestration:
- Native, agentless backup of VMs on AHV through the Veeam Plug-in for Nutanix AHV, with application-consistent snapshot support and granular restore of individual files without needing to spin up the full VM.
- Veeam Recovery Orchestrator (the evolution of Veeam Availability Orchestrator) as the functional equivalent of SRM: documented recovery plans, non-disruptive testing, and automated failover/failback, with no tie to the VMware stack.
Ransomware: The Risk Domain Changes
No known attack has managed to encrypt the contents of a vDisk when it lives as its own volume on external storage, isolated from the shared datastore that’s the typical target of campaigns like ESXiArgs or DarkBit — documented cases of mass encryption of .vmdk files on VMFS. They could still delete it, but deletion isn’t ransomable: there’s nothing to negotiate when there’s no key left to sell.
There’s no such thing as absolute protection — no security discipline promises that, because someone always finds a way around it. What does change here is the blast radius (how much can be touched from a single compromise) and the chance to simplify isolation (one volume per VM, not a shared datastore), combined with fast recovery tools: deduplicated snapshots, retained for weeks, invisible to whoever already broke into the virtual environment.
In a large enterprise, once this is properly understood, a good Technology Partner — even an expensive one — gets paid gladly: any solution that ends up cheaper than the problem justifies its own cost. After that, it’s a matter of choosing the most suitable one — or the cheapest one — for each situation.
Deduplicated Snapshots as an Anti-Ransomware Mechanic
Storage with real deduplication, like PowerStore, changes the math on how many recovery copies make sense to keep. Since each snapshot only takes up space for the blocks that changed, keeping daily (or more frequent) versions for weeks stops being an expensive decision. Combined with the immutability of “secure snapshots,” this gives you a ransomware recovery layer that doesn’t depend on restoring from backup: you roll back to an earlier point directly from the storage, in minutes, with a copy the attacker never could touch or delete.
An Underused Capability, and a Powerful One: Exploring Inside the Snapshot
Veeam has a native feature — Veeam Explorer for Storage Snapshots — that connects directly to the storage array and lets you browse what’s inside an array snapshot, without that snapshot needing to be a Veeam backup at all. You can reach into a VM living inside that snapshot and pull out a single file, without restoring the whole machine. It’s a tool that’s barely used in our regional market, and one that’s genuinely valuable in enterprise environments for fast, surgical recoveries.
Today this capability is documented and proven on VMware vSphere datastores integrated via the Universal Storage API (the same one the Dell PowerStore plug-in uses). For AHV, we found a signal pointing the other way: the technical documentation for the Veeam plug-in for Nutanix AHV describes a “worker” mechanism that accesses VM disks over iSCSI, using Nutanix’s native snapshot — not the Universal Storage API. This isn’t an official confirmation from Veeam, but it’s a detailed and consistent technical source, so it’s likely that this specific feature isn’t available today on AHV+PowerStore. This is very new ground — worth confirming with Veeam before promising it to anyone.
What’s Confirmed and What Still Needs Validation
Confirmed today:
- AHV supports PowerStore as external storage (NCI 7.6), managed from Prism, with Metro/synchronous/asynchronous replication.
- Veeam does native, agentless backup of VMs on AHV, with individual file restore.
- Veeam Recovery Orchestrator orchestrates recovery plans independently of VMware.
- PowerStore offers immutable snapshots, “secure snapshots,” and high-level deduplication.
To confirm with Dell/Veeam before committing to it with a customer:
- Whether Veeam backup for AHV can offload/restore directly against PowerStore’s own snapshots. There are signs today that it can’t: the plug-in’s technical documentation describes a “worker” mechanism that accesses disks over iSCSI using Nutanix’s native snapshot, not the Universal Storage API used for vSphere+PowerStore. Not an official confirmation from Veeam, but a strong signal.
- Whether Veeam Explorer for Storage Snapshots can browse PowerStore snapshots when the hypervisor is AHV — same uncertainty as the point above.
Enterprise Databases: Oracle and SAP HANA
Two typical enterprise workloads deserve a separate mention, because they don’t behave the same way in this migration.
Oracle Database remains a well-known pain point in the market, and moving to AHV doesn’t solve it. Oracle’s partitioning policy explicitly names Nutanix AHV — alongside VMware vSphere and Hyper-V — as “soft partitioning”: it doesn’t qualify as hard partitioning, no matter the pinning, affinity rules, or sub-cluster isolation involved. In practice, Oracle expects to license every physical core in the cluster — exactly the same problem it has with vSphere today. The only real way out remains Oracle Linux KVM (Oracle’s own distribution, with hard partitioning officially recognized) — and that’s where an interesting architecture comes in: a small, isolated Oracle Linux KVM farm, connected to the same PowerStore over iSCSI or Fibre Channel, while the rest of the environment runs on AHV. One storage array, two compute islands, each solving its own licensing problem.
SAP HANA is a different case: its historical restriction was always about certification, not licensing, and that certification already exists. The HANA-HyCI program (HANA on hyperconverged infrastructure) formally certifies combinations like the Dell XC Family with Nutanix AOS and AHV as the hypervisor — valid through October 2028. An SAP customer running HANA on supported VMware hardware today has already invested in certified hardware; moving that certification to AHV isn’t a conceptual leap, it’s a real path.
Why This Matters Beyond the Technical Details
This integration isn’t exclusive to Dell: Nutanix already offers the same thing with Pure Storage FlashArray, and it wouldn’t be surprising if, over time, NetApp, Nimble, or any storage vendor with the block-level control capabilities — Change Block Tracking, native snapshots — this model needs joins in too. It’s a path similar to the one VMware walked with vVols: NetApp was one of the launch reference partners back in 2015, and the rest of the storage market gradually joined in over the following years until it became compatible with nearly every serious vendor.
Back when VMware was a genuine Technology Partner, we didn’t mind the implicit lock-in that came with each proprietary API — VASA, VAAI, VADP, VAIO, PSA, and so many others VMware kept creating because, as the market leader, it was the first to run into every new problem. The whole ecosystem was comfortable with that because VMware showed up with a tailored solution for every type of customer.
Nutanix could, over time, become a new point of dependency. But it isn’t the same kind: AHV is built on KVM, the same open core used by Proxmox, Red Hat/oVirt, and OpenShift Virtualization, among others. If Nutanix ever stops behaving like a partner, the distance to the next migration is a lot shorter than the one that separated VMware from any alternative — it’s not starting from zero, it’s switching wrappers on the same foundation.
Native NVMe integration with external storage is a real differentiator today, and like any differentiator, it comes at a price. But being first doesn’t make it permanent: the same openness that let Nutanix build this integration is what will let other players reach the same place over time.
On the compute side, that openness has already existed for over ten years: Nutanix runs equally well on Dell, Lenovo, HPE, Cisco, or Fujitsu hardware, in addition to its own NX brand. If storage fully joins that same logic, the result is a genuinely agnostic software layer end to end: you choose the hardware that suits you best, the vendor doesn’t choose for you.
One clarification is worth making here, so as not to fall into the same trap we’ve been pointing out with other vendors: the certified hardware Nutanix requires isn’t lock-in in the same sense. It’s a service-level commitment — Nutanix’s lifecycle management (LCM) needs to know the hardware in order to automate updates and keep continuity without surprises. You pay for that, but you’re paying for something real: fewer operational headaches, not less freedom to choose your next move.
And that freedom to choose your next move is, at bottom, the real point of this Solution. I’ve watched up close what happens when a business relationship stops being treated as a long-term partnership and starts being managed as a financial asset to be squeezed: perpetual licenses left without support, local partners who stopped being necessary to sell the product, R&D investment redirected to wherever the capital markets are looking today, not to wherever the customer will need it tomorrow. Even if VMware is still the best Enterprise virtualization solution today, the market should be looking at an alternative — not necessarily because of cost, but to keep a Technology Partner relationship going over time.
Who This Solution Is For
Organizations that currently run vSphere + external storage + SRM + Veeam, and are evaluating how to reduce their dependency on VMware without giving up the level of DR orchestration they have today.
Transformation: The Road Toward Containers, Down the Line
Everything above describes what we used to call Virtualization, and what most of the market now prefers to call “Private Cloud” — the name is a bit confusing when the infrastructure is on-premises, but it’s the term that stuck. This architecture handles that traditional world well. The real transformation — the one that changes how applications get built — is a different one: containers, CNCF, DevOps, Kubernetes. It isn’t part of this Solution today, but it’s worth mapping out the terrain, because all of this can be added over time on the same AHV + PowerStore foundation.
On the VMware side, the bet for that is called Tanzu — which still exists as Tanzu Platform after Broadcom’s reorganization, though the Kubernetes-on-vSphere piece was renamed vSphere Kubernetes Service (VKS). Nutanix, for its part, has its own offering that’s less well known in our market: Nutanix Kubernetes Platform (NKP), which inherited the D2iQ acquisition combined with Nutanix Kubernetes Engine, positioned as a full alternative to Red Hat/Rancher. It doesn’t have the name recognition of Tanzu or OpenShift locally, but it exists and competes in the same category.
Beyond Nutanix’s own option, AHV can run almost every relevant Kubernetes distribution: Red Hat OpenShift has an official deployment guide for AHV (Nutanix is a Red Hat-certified hypervisor), and SUSE Rancher and Canonical Charmed Kubernetes are viable too, following the same general principle as any Kubernetes distribution on any hypervisor: as long as the workloads are stateless, installation is pretty straightforward.
Where it starts requiring real design work is with persistent (stateful) data. That’s where the Container Storage Interface (CSI) comes in: Nutanix has its own CSI driver to expose Volumes/Files as persistent storage to Kubernetes. On top of a CSI, you can run Kasten — now part of Veeam, rebranded Veeam Kasten for Kubernetes — which does native backup and mobility for Kubernetes applications, and explicitly supports Nutanix’s CSI, Portworx, and Dell PowerStore’s CSI. It’s a second way Veeam shows up in this architecture, this time on the container side instead of the VM side.
We only know this part from the lab so far, not from production: we’ve tested Nutanix CSI + Kasten on AHV, and separately Proxmox + vanilla Kubernetes + Portworx, on small, non-production setups. It’s enough to know it works and how each piece fits together, but it’s presented here as functionality to explore and add over time, not as a deliverable part of the Solution yet.
Marcelo Soria: https://www.linkedin.com/in/marcelo-soria-475aab1