The Five Stages of VMware Grief (and Why It’s Time for Acceptance)

Where I’m coming from

My first VMware install was ESX 3.0, and I’ve been writing about it on this blog since 2010 (My First blog post was about VMware SMB Designs!). Along the way I picked up a VCP and a VCAP-DV, made the top 100 virtualization bloggers list, and went to VMworld more than once on a blogger pass.

Quick disclaimer before I go further: I work for HPE on the Zerto team, so I have skin in the game. This is my opinion, not HPE’s. Nothing here comes from inside HPE. Everything I cite is public and linked at the bottom, and I’m not pitching you a product, just ranting a bit.

If you were there, you remember. The community area at VMworld was packed with bloggers writing side by side, trading notes and arguing about storage. The vExpert briefings gave us an early look at what was coming. John and Corey built and ran something very magical. Partners also poured money into the show floor because everyone wanted to be there. I still smile thinking about Red Wednesday at the Zerto booth. Hundreds of people (dare I say a thousand) showed up in red Zerto shirts for the Wednesday prize drawing. That was the ecosystem: huge, a little chaotic, and everybody invested in it.

So I’ll be honest: this isn’t a post I ever wanted to write. But here we are…

My denial didn’t last long. It ended about the time I read about the new licensing. Everyone was forced onto subscriptions at much higher prices, and then partners and customers were cut loose just because they weren’t giant accounts. That’s when I knew the VMware my IT career started on wasn’t coming back.

The pattern

No single one of these moves is the story, but when you put them in order, they read like a plan.

  • November 2023: Broadcom closes the $69 billion VMware deal.
  • February 2024: Partner Connect shuts down, replaced by an invite-only program. Perpetual licenses give way to subscriptions.
  • June 2025: The Registered reseller tier is eliminated. Only Pinnacle, Premier and Select survive.
  • November 2025: The cloud provider program goes invite-only. The white label program that let small providers buy through bigger ones ends with it.
  • January to March 2026: Europe gets the same treatment. CISPE, a European cloud provider trade group, files an antitrust complaint with the European Commission.
  • April 2026: evoila (a German IT consulting company) counts 19 authorized cloud providers left in the US. The global program once had thousands.
  • August 2026: The public VDDK download pages go dark. No announcement.

That last one is the tell. VDDK, the Virtual Disk Development Kit, is the library software uses to read a VMware virtual disk. Backup vendors use it. So do the migration tools: Red Hat’s Migration Toolkit for Virtualization, Nutanix Move, Microsoft’s Azure Migrate, Platform9’s vJailbreak. Broadcom support told customers who asked that it is “no longer available for use or download.” Broadcom told TechTarget the licensed use “has always been backup and recovery,” and that VDDK “is not a customer entitlement.”

Maybe. But the practical effect is that the easiest way to copy your own VMs off your own hypervisor now runs through a Broadcom gate.

It’s not just resellers and cloud providers, either. I talk to partners and other vendors all the time, the companies whose products plug into vSphere. I keep hearing the same thing from a lot of them: the program terms keep getting tighter, especially for anyone who also sells an alternative.

VMware is becoming the mainframe

I don’t mean that as an insult to the mainframe. Plenty of big banks still run on IBM Z, and IBM makes good money on it. That’s the model.

Here’s what a mainframe business looks like, and how closely VMware now matches it:

  1. A captive base. Mainframe customers stay because the workloads are too embedded to move quickly. VMware’s are too. CloudBolt surveyed 302 enterprise IT decision-makers in January. 86% are reducing their VMware use, but only 2% have moved 75% or more of their environment off it. Wanting to leave and being able to leave are two different things.
  2. Fewer, bigger customers. You don’t need the long tail when the top of the base pays enough. Info-Tech’s Scott Bickley put it plainly: keeping just the top 10,000 VMware customers would be enough for Broadcom to hit its goals.
  3. A controlled ecosystem. Authorized partners only, and fewer of them every year. Now the toolkit to read your own disks is behind a gate too.
  4. Pricing power. In that same survey, 59% saw prices rise by more than 25%. CISPE claims prices rose tenfold for European cloud providers. Either way, the rent goes up because it can.

And nobody builds anything new on a mainframe. It’s where existing workloads go to be maintained, very profitably, for a very long time. That’s the future I see for VMware: still running, still expensive, and no longer where the industry builds what’s next.

If you’re one of those top 10,000 accounts, you may get treated very well. If you’re not, you’re the long tail, and the long tail is the part the model is designed to shed.

The five stages of VMware grief

If you’ve been in this ecosystem as long as I have, you’ve probably been through all five of these. I know I have.

  1. Denial (late 2023). “They paid $69 billion for it. They’re not going to break it.”
  2. Anger (2024). Perpetual licenses gone, bundles you didn’t ask for, renewal quotes that made people say words I won’t print here.
  3. Bargaining (2025). Multi-year deals to lock in a price. Partners scrambling to qualify for a tier that might still exist next year. “If we just commit to VCF, maybe it’s fine.”
  4. Depression (2026). Almost everyone says they’re leaving. Almost nobody is out. The partner you bought from for ten years got a non-renewal letter. And then the VDDK link 404s.
  5. Acceptance. This is the stage where you stop waiting for the old VMware to come back and start planning your next move.

My read: most of the industry is somewhere between four and five right now. And there’s a date that’s going to push a lot of people into five whether they’re ready or not.

The clock

vSphere 8 hits end of general support on October 11, 2027. That’s about 11 months out. ESXi 8, vCenter 8 and vSAN 8 all reach it the same day. If you’re on vSphere 8, you’re making a VCF 9 decision in the next year whether you mean to or not. Renew into the next major version and you’re signing up for another multi-year stretch on Broadcom’s terms.

vSphere 8 and VCF 9 don’t give you the same exits

This is the part I think a lot of people are missing. Right now, on vSphere 8, there are still agentless tools that can move your VMs to Azure, AWS, Hyper-V or another hypervisor with very little downtime. Nothing to install inside the guest. You point the tool at the VM, move it somewhere else, and you’re done.

I wouldn’t assume those agentless options make the jump with you to VCF 9. Ask every vendor in your stack, in writing, what they support on VCF 9 for moving workloads off of it. Don’t be surprised if that answer is a lot shorter than it is on vSphere 8.

Can you still move after that? Sure. But depending on how long you’ve been doing this, you might not remember in-guest agents. Back in my early days in IT, backup software protected VMs with the same agents we put on physical servers, and we walked uphill both ways to patch them. Managing agents at scale is not fun: one on every VM, every OS version, every reboot window.

So if you aren’t ready to be part of VMware’s mainframe era, start moving now. Treat October 11, 2027 as the date you’re done, not the date you start.

A real migration of a production estate isn’t an 11-month project for most shops. So if you’re going to leave, the planning should have already started. You can’t wait any longer.

If you run VMware

Start the conversation now. Get a plan on paper before the renewal quote shows up, and get your CFO in the room early. Then:

  • Pull your vSphere versions and your subscription renewal date, and lay them next to October 11, 2027. Do you even have that long? Or does your contract expire sooner?
  • Inventory every tool that touches vSphere: backup, DR, monitoring, migration. Ask each vendor what happens to their product if Broadcom changes their API access.
  • Find out whether your migration tooling depends on VDDK, and what your vendor’s fallback is if it does.
  • Pilot an exit on a non-critical cluster. You won’t know the real effort until you’ve moved something.
  • Price staying honestly: the full multi-year cost of VCF, not just year one.

If you’re a partner or MSP

  • If more than half your revenue rides on one vendor’s partner program, you’ve already seen how fast that program can change. Fix that.
  • Read the termination clause in your program agreement. Know how much notice you’d get.
  • Build a second-platform practice now, while your VMware customers still need help moving.

What I’m not saying

  • vSphere isn’t bad technology. It’s still the most mature hypervisor out there. This is about the business wrapped around it.
  • Broadcom isn’t being irrational. For its shareholders, this is a coherent strategy. It just doesn’t have your best interests in mind unless you’re in the top tier.
  • Not everyone should rip and replace tomorrow. If you’re a huge shop with a good negotiated deal, staying may be the right call for years. Just make it a decision, not a default.
  • This isn’t settled. CISPE’s complaint went to the European Commission in March, and regulators could still change the picture. I wouldn’t hold my breath on that; we’ve all seen regulators go away once some money moves around.

Acceptance isn’t giving up

The last stage gets a bad rap. Acceptance doesn’t mean you lost. It means you stopped spending energy hoping things go back to how they were (they aren’t going back despite the new vSphere packaging) and started spending it on what’s next.

The VMware party was a great one. I built a career at it. But the host has made it pretty clear who’s still invited, and the exits get a little narrower with every change. October 11, 2027 is coming either way. Go somewhere else on your own schedule, not Broadcom’s.

Sources

Loading

Share This Post

Post Comment