Euro-cloud Anexia moves 12,000 VMs off VMware to homebrew KVM platform Broadcom has lost another sizable customer for its VMware platform: Austrian cloud provider Anexia has moved 12,000 VMs, some of them rented by major European businesses, to an open-source system based on the KVM hypervisor. Anexia was founded in 2006, is based in Austria, and provides cloud services from over 100 locations …
Based on what I've see planning our own migration, the cost of migration is anything from 2x to 10x the cost of VM License increase and that isn't including disruption costs which I assume the customer is picking up in this case. I am sure this company is probably paying more per VM for licenses due to their licensing model so migration cost is probably closer to VM license costs but I would still expect it to be 2x until they have spare hardware, hosting and licenses for new platform lying around.
I am also guessing the company is absorbing the migration cost ?
From the article, it sounds like price was a concern, but a bigger factor was a change in the terms for their billing arrangement. The CEO said they had been paying month-by-month for each previous month's usage. Broadcom was expecting them to pay for the next 2 years up front.
For a subscription-based organization like a hosting service I could see how this would be a huge thumb on the scale...
This post has been deleted by its author
We have, our VM licensing increase is going up by around $40m when we renew in 2027. The migration project includes alternative hypervisor licensing, dual hardware, app changes, testing, etc comes in between $75m and $100m. I've spoken to companies where migration figures vs VM licensing are even worse and some of that comes down to the type of VM licensing. Its a bit of a fudge as modernisation which would have happened anyway gets included in that migration cost but suffice to say that migration isn't free. Maybe it is for other organisation ?
Yes, we will accept your figures, but the +$40m is on top of the previous contract's cost. Plus, it is then reoccurring with that increased rate, either yearly or bi-yearly.
The $75-$100m has a portion of hardware costs included and then yearly/bi-annual reoccurring costs should get significantly reduced.
Pay now and acquire long-term reduced costs and increased productivity (due to hardware upgrade); pay later and stay with current levels of performance plus high subscription costs.
Long-term goal or short-term stagnant lock-in (IMHO) - it's your call.
We are definitely getting off and not cost but the shoddy way broadcom operates. vmware is just one of several of their products we are or have ditched.
Hardware wise, yes we would do rolling upgrades anyway but pulling ahead 5 years spending into 24 months is big hit on capital spending. So potentially 3 years where we don't have to do any replacements. But then introduces a new.issue that instead of doing 300 hosts a year over 5 years, its now 750 a year for 2 years.
Also the licensing alternate hypervisor is cheaper but just as much of a pain. Nobody idoes perpetual licenses but that would have happened with vmware as well. And there is little doubt that as a product vmware had it issues but was pretty decent and well supported by 3rd party.
Thing is at the end of the day I suspect whatever you do you will have pain, just be pain from a different vendor.
Then I would strongly question your ability to manage such a project.
We migrated from VMWare to XCP and I can not even begin to imagine how YOU came up with a migration cost estimate of 2-10X the 3x increase in VMWare cost.
Could you provide me your companies name? I would like to bid on that migration!
Nah. This is more of the 'We own the market so let's make hay'.
Customers are wising up to moves like this (and the MS price hikes in Asia) and rebelling.
One has to hope that large stockholders also see what is going on and tell the BOD to stop pissing off their customers.
The BOD will ignore this advice as they want that new G-600 ASAP. Status symbols and bragging rights baby.
Yup. I completely agree.
I'll be very interested to read Broadcom's end-of-year 2025 report on how their cunning scheme has lost them money because all those big accounts left them.
The cynical in me says that Broadcom would have better off gouging the little accounts for twice as much, because little accounts do not have the resources (time, knowledge and/or money) to migrate to something else.
The big accounts take one look at the new bill and say : nope, not paying that, let's go somewhere else.
And they demonstrably can, because this isn't the first time it has happened.
We're considering the same, albeit to Hyper-V. Once a single Broadcom license is in a cluster, it poisons the whole set so VMWare long term licenses can't be used. We have servers going end of life soon and can't replace vSAN/vCenter licenses for them as easily as the machines themselves.
"I'll be very interested to read Broadcom's end-of-year 2025 report on how their cunning scheme has lost them money because all those big accounts left them."
The feeling I have from the various news I've read on this topic is that Broadcom's price hikes are so large that they can afford to lose a lot of large accounts and still come out on top. The rationale will be that...
(a) many customer directors won't have a clue what VMware is or does and will thus be fearful to meddle with the Jenga-like stack of software most large companies have, or
(b) there's plenty of companies who'll think "this is outrageous, but we've got bigger fish to fry, let's pay up and worry about this in future", and then there's
(c) up-their-own-arse companies with deep pockets who follow the L'Oreal reasoning.
Sadly I think Broadcom will see big returns from this for a good few years, albeit at the expense of slowly killing off their own product. But when did any director earn a bonus or ramp the share price through long term thinking?
Anexia admit themselves they were in a fortunate position to make the migration before committing to a 2 year contract. Most won't so the 25/26 accounts should show a huge revenue increase. Of course a 2 year window may be good to design and execute their more difficult migrations.
You probably need to look 4/5 years out to see Broadcomm made a packet out of the aquisition, execs and stockholders rewarded. The operation was successful but the patient died.
VMWare RIP.
Might have been the plan all the time. Good news is this will move minds from dependence on proprietary products to open source. Not a perfect solution as we have seen in the IBM/Red Hat aquisition
Anexia were in the fortunate position ..... of having the big swinging balls required to both pull finger and actually pull it off. Other companies may have more timorous chittering weenies easily herded up the arroyo by Broadcom fear wolves.
You should not wait for details. Despite the loud and public exits of a few reasonable-sized customers (like here), many other large customers are quietly signing new contracts:
"The latter, which measures the value of future revenue from subscriptions, saw $2.7 billion worth of deals done in the quarter – up $200 million from Q3. Tan revealed VMware sold subs for 21 million processor cores in the quarter – up from 19 million in Q3.
The CEO also told investors that 17 million of those newly-sold cores will be used to run the flagship private cloud suite VMware Cloud Foundation (VCF), and that 4,500 of Broadcom's top 10,000 VMware customers have signed up for VCF since the acquisition.
Full-year revenue for Broadcom’s software division hit $21.5 billion, up from $7.6 billion for FY 2023 – an increase of $13.8 billion. VMware’s last full year of revenue as an independent company was $13.4 billion, and Broadcom did not own the virty giant for a few weeks of its FY 2024 and therefore can't count a few hundred million dollars of revenue."
It therefore looks a lot like VMware revenue is growing and Broadcom’s strategy is working.
https://www.theregister.com/2024/12/13/broadcom_q4_fy_2024_vmware/
Ticket price up 500%, sales revenue up about 10% or so. That tells me most people aren’t signing new contracts. The question is, how many of the people who did plan to migrate before the next contract renewal? I suspect it is >0%, and therefore revenue in future will go down.
And this is why you should look beyond just quarterly earnings reports with number go up.
> I'll be very interested to read Broadcom's end-of-year 2025 report on how their cunning scheme has lost them money because all those big accounts left them.
2025 will be fine because they know most customers won't be as agile as Anexia and so will have to renew while they plan to leave.
And Broadcom know this full well - it's not a coincidence that they are asking for two year contracts because they know plenty could be ready within a year so they're going to force them to pay for two.
A better way to make trouble for Broadcom is to look at their SEC 10K filings and make a complaint to the SEC regarding an omission of material fact if they don't include words to the effect: "our 2024 VMware licensing decision will significantly reduce our customer base by end 2026, with commensurate loss in revenues, profits and opportunities to cross-sell services."
The cynical in me says that Broadcom would have better off gouging the little accounts for twice as much, because little accounts do not have the resources (time, knowledge and/or money) to migrate to something else.
The "little accounts" might do something extreme like continue to use their pre-Broadcom perpetual licenses without ongoing support/updates.
You also wouldn't see any new VMWare customers, as future employees won't get experience with VMWare, and med-large companies grown from little startups wouldn't ever have gotten a start with VMWare.
Making a packet out of the VMWare base is Broadcom's plan.
Maybe an opportunity for Anexia to do likewise and transition from being a cloud company to a migration specialist. Why waste that skill hard won under pressure as a 'once in a lifetime experience'.
There must be thousands of Broadcom prisoners looking for help in cutting their chains.
Surely much more fun and lucrative? Only issue is if Broadcom kill their base too quickly. What to do next?
There will potentially be some shift in pricing. Right now the cost of migration is probably similar to the cost of increase in VM license. Plus add in the cost of hypervisor is a significant but small part of the overall hosting cost. So don't expect much change in the difference between hosting companies.
When I started pricing out ESXI 8 I noticed that I was having problems locating the perpetual license. Then I began digging and found all of the changes Broadcom was implementing. I had my team start planning alternatives in November 2023. We completed our migration last summer.
So...Broadcom...FYOU!
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
