For more than two decades, IT organizations have benefited enormously from virtualization. We took expensive, often seriously underutilized physical servers and made them do the work of multiple systems. Fewer servers. Less power. Less rack space. Less hardware to maintain. More flexibility.
And VMware was the 800-pound gorilla that made it happen.
VMware didn’t actually invent virtualization. IBM was experimenting with virtual machines back in the '60s. IBM’s CP-40 project began in '64, followed by CP-67, and eventually VM/370 in '72. z/VM History: Timeline - IBM [ibm.com], [ibm.com]
That is 1964, for you younglings out there.
The problem IBM was trying to solve sounds surprisingly familiar today: computers were incredibly expensive, and organizations weren’t using those expensive resources efficiently.
Then came the distributed-computing boom of the 1980s and 1990s. Servers got smaller and cheaper, and pretty soon everything had its own box.
Web server? Buy a server.
Database? Buy a server.
Domain controller? Server.
Application needs 4 percent CPU utilization on alternate Tuesdays? Congratulations, here’s your server.
We filled racks with hardware.
VMware didn’t create virtualization, but it helped make the concept practical on commodity x86 hardware. Instead of maintaining 20 physical servers running 20 operating systems for 20 different purposes, we could deploy three or four larger servers and run 20, 30, or more virtual machines on them.
Smaller footprint. Lower power consumption. Less hardware maintenance. Same services.
It was brilliant.
And companies embraced it.
I certainly did.
At my previous company, Virtual Instruments, later Virtana, our entire business revolved around monitoring the health and performance of storage and virtualized infrastructure. I spent 13 years there supporting a 50-rack data center filled with VMware infrastructure and our own technology for engineering, development, testing, and corporate IT.
We had an enterprise licensing agreement with VMware, multiple vCenter clusters, and enough fiber optics to make you question every career decision that led you into a data center at 2:00 a.m.
VMware wasn't just another piece of software.
It was part of our infrastructure strategy.
And we weren't alone.
VMware built an ecosystem. Centralized management, storage virtualization, networking, vCenter, vSAN, NSX, and a growing catalog of technologies made VMware extremely compelling.
The interesting thing was that eventually nobody really talked about virtualization anymore.
When someone said, "Spin up a VM," everyone knew what that meant.
VMware had accomplished something few technology companies manage to do: its product category had become almost synonymous with its name.
Then Came Broadcom
Hello, Broadcom.
We need to talk.
Broadcom completed its acquisition of VMware in November 2023, and shortly afterward VMware by Broadcom announced major changes to its product and licensing strategy, including completing the transition away from perpetual licensing toward subscription licensing and simplifying the portfolio. VMware by Broadcom Dramatically Simplifies Offer Lineup and Licensing Model [oplee.com]
Broadcom called this"simplification."
Customers may have chosen other words.
The VMware relationship we had grown accustomed to changed remarkably quickly.
Perpetual licenses gave way to subscriptions.
Licensing increasingly focused on physical CPU cores.
The product portfolio was consolidated into fewer offerings.
And today, VMware Cloud Foundation and VMware vSphere Foundation licensing require customers to license the physical cores in their CPUs, subject to a minimum of 16 licensed cores per physical processor. Counting Cores for VMware Cloud Foundation and vSphere Foundation and TiBs for vSAN [knowledge....oadcom.com]
Apparently Broadcom even wants to know how many cores we're seeing.
That's when you know the relationship has gotten serious.
Broadcom explained the transition with familiar phrases such as continuous innovation, faster time to value, predictable investments, and simpler offerings. VMware by Broadcom Dramatically Simplifies Offer Lineup and Licensing Model [oplee.com]
And I understand the argument.
Software companies want predictable recurring revenue. Supporting enormous product catalogs is complicated. Subscription licensing has become common throughout enterprise software.
I understand all of that.
But there is a larger problem.
The economics that made VMware compelling have started working in reverse.
The original VMware conversation was:
"How can we reduce the amount of physical infrastructure we're paying for?"
The Broadcom-era conversation increasingly feels like:
"How can we reduce the amount of VMware we're paying for?"
That is quite a transformation.
We went from optimizing hardware because software made it possible to optimizing hardware because of how the software is licensed.
And that's when I started looking around.
Broadcom, We Need to See Other Hypervisors
I currently lead IT for a mid-sized medical technology company, and thankfully our VMware footprint is relatively small.
Our annual VMware costs have gone from roughly $12,000 to $30,000 and now to approximately $62,000.
And eventually someone has to ask the uncomfortable question:
For what?
Not because VMware suddenly became bad technology.
It didn't.
VMware remains an extraordinarily capable virtualization platform.
But this is where Broadcom seems to misunderstand the relationship.
You can possess excellent technology and still price a customer into questioning whether they need it.
At some point, the discussion stops being about whether VMware is technically better and starts becoming:
Is it $62,000-a-year better for what we actually need it to do?
That's a very different question.
And, Broadcom, you accidentally encouraged me to ask it.
Because here's the awkward part.
We have options.
We're already licensing Windows Server Datacenter on these systems, so Hyper-V deserves a serious look. There are other alternatives as well, including Proxmox and additional virtualization platforms.
Could moving mean giving up some VMware features?
Absolutely.
Could it mean more manual administration?
Possibly.
But somewhere between $12,000 and $62,000, I rediscovered my enthusiasm for manual administration.
Funny how that works.
Perhaps we don't need every bell.
Maybe we don't need every whistle.
Maybe we just need the workloads to boot on Monday morning.
So, Broadcom, It’s Over
I don't say this with anger.
Okay, maybe a little anger.
But mostly this is economics.
VMware spent decades building customer loyalty, technical expertise, operational dependency, and an extraordinary ecosystem. IT organizations invested countless hours learning the products, building infrastructure around them, and trusting VMware with mission-critical workloads.
That relationship had tremendous value.
Broadcom inherited it.
And from where I sit, Broadcom appears to have underestimated just how quickly that loyalty can become a spreadsheet exercise when the economics change enough.
I'm not leaving virtualization.
I'm leaving you.
There is an important difference.
So, as we prepare ourselves for life after VMware, I thought it appropriate to place a small classified ad:
VIRTUALIZATION PARTNER WANTED
Long-term IT department seeking a stable, trustworthy virtualization partner after ending a once-happy relationship that became expensive, complicated, and increasingly one-sided. We enjoy reasonable licensing, predictable renewals, useful features, and partners who don't measure affection by the number of CPU cores we own. Must play well with servers, respect our budget, and understand that commitment is earned, not bundled into a three-year subscription. Hyper-V and other emotionally available platforms encouraged to apply.
Broadcom, I genuinely wish you the best.
We'll always have vMotion.
But it's time for us to see other hypervisors.
It isn't me.
It's definitely you.
