VMware turns its back on small businesses: 72-core license policy backlash | Hacker News Reader

VMware turns its back on small businesses: 72-core license policy backlash

systemadministration.net

47 pointswalterbell38 comments

unpopular opinion but there are so many mature well established open source alternatives to vmware in 2025 that not only outperform it but are easier to set up.

even the humble cockpit project supports vms using libvirtd and kvm, and provides a workable interface for virtualization that supports migration and snapshots. do give it a try.

Broadcom has been so abrasively hostile to the idea of the vmware business it acquired that its intentional rejection of profitable markets should be enough to start purging members of the board of directors.

This looks pretty anemic compared to what I've seen from XCP-ng or proxmox:

https://cockpit-project.org/guide/195/feature-virtualmachine...

Thanks for the tip, though.

If you don't mind the terminal.

vAdmins are historically very click-ops. Not a slight against them; VMware grew up in a time when this was en vogue. Also, VMware has made their API very opaque and difficult to use aside from happy path scenarios. (Check out govc to see what I mean. It's great until you start hitting edge cases that are easy in the UI but very hard or darn near impossible with the SDK, like creating a distributed vSwitch and adding VMs to different port groups).

vSphere is very, very nice. Super mature interface, and that's before you pair it with vSAN and NSX.

AFAIK no-one's caught up

100% agree there is really nothing like VMware. Incredibly robust. Just works. It’s a juggernaut.

I am sure it is a "Call for pricing" situation, but just curious what is the magnitude of the jump. Is a 16 core seat $10k, and a 72 core is $400k?

Not Broadcom, but we just had a vendor pull a similar stunt at the start of the year. With a month before our renewal, we were told that the new license was going to jump 2.5x owing to the new minimum purchase requirements. Which is a head scratcher to me, because I have never thought the product was that good or lacking in competition. There is now an internal initiative for everyone to fully migrate off by end of year.

At the low end it's remarkable to see the influx of new Proxmox users since VMWare removed the free tier of ESXi licensing. Proxmox is great, FWIW.

Are the organizations building on VMWare going to learn any lessons from this?

What did you expect?

There exist and have existed for decades robust full featured open source solutions.

Use them.

> What did you expect?

I can't speak for all orgs, but I'm guessing it was something along the lines of "There's plenty of competition out there, we're paying them good money and they're profitable - why would they screw up a good thing?"

Almost every organisations will choose current cost over potential future issues to a greater or lesser degree. Backups most will spend a reasonable amount on ensuring they have them working. A cold-standby DR site, fewer.

I doubt many are defensively engineering their organisational IT strategy to run two competing platforms at every level just incase one gets bought out by the mob who turns up the pricing by several multiples.

I'm no VMware fan, but this:

> There exist and have existed for decades robust full featured open source solutions.

is reaching.

vSphere + NSX for advanced network virtualization + vSAN - I'm not sure what "robust open source solution" has covered that ecosystem, let alone for decades.

I've gone from XenServer to vSphere to ProxMox. I played a lot with lxc/lxd in its earlier days.

Openvswitch and ceph probably, although ceph is a real stretch to match vsan

Not surprising. Hock made it extremely, abundantly clear that he only cares about the biggest handful of customers (that purchase hundreds of thousands of cores). Killing free vSphere was one of his first moves.

That said, no SMB is running a vDC on 72 cores unless they're okay with insane overcommit, which they would've already moved over to KVM if so.

I've been using Proxmox on a few production servers for nearly 5-years (part of my COVID19 extra-curricular activities was migrating off the over-bloated VPSs I had been using). It's primarily hosting web apps, web servers, and dev infrastructure, and it's honestly rock solid from my (admittedly limited) perspective.

Aaaaaand there it is. That guarantees I’ll never, ever approve any design plan that involves VMware in any way.

What shortsighted jackassery. Know where big customers come from? Little customers. There’s now zero incentive to ever consider using it unless you’re at a giant company, and if you’re a giant company, there are better options.

This is the death knell. It’s dead.

It's just throwing money away too. Small customers aren't using all the advanced features, probably barely touching the support tickets. It's a software license, selling it costs them nothing.

I guess you can cut the size of your sales team. You no longer have to go hunting for customers, because you can just look at market caps.

BRCM generally has an account manager (or several) for specific companies in the Enterprise. This is the target customer they're after, not volume. There's no profit in volume support here so BRCM doesn't waste their time and resources.

You might be surprised, but for many businesses that serve both enterprise and small customers, the smaller accounts often end up being a poor use of time. While it can vary, in many cases, small customers contribute only a small fraction of total revenue yet require a disproportionate amount of support and resources.

I'm not surprised, but I also don't think that's likely to be a thing for VMware: Most small customers have one to three servers running a basic vSphere install and nothing complicated to speak of.

You can also delete most of the problem by limiting the support on given license tiers.

We're speculating, but I think the key point is missed: SMBs likely contribute a negligible share of revenue, small enough that supporting them isn't worth it. Even if support seems minimal, billing issues and tickets inevitably arise, often handled by a different team, which cancels out any marginal gains.

This is a bane of my business. We’re a small business, but we’re a technology business. When we use a product we tend to use advanced features. Things like auditing and log streaming to SIEM, enforced SAML SSO, stuff like that. We tend not to need support, and when we do it’s almost always because something is broken on the vendor end.

In the past six months I’ve filed exactly one support ticket. With Microsoft, due to a persistent internal error when calling a specific Azure API, which only they could fix, and which they did after about a week and a half.

There is virtually nothing we can’t handle internally except for the the issues any customer couldn’t handle.

The irony is that not having an avenue for SMBs makes SMBs sometimes go elsewhere and either spend more money (sometimes with a competitor), or missing out on basically “free” revenue.

I understand the logic but it drives me crazy at times.

This is a classic long tail issue. That segment of technically advanced, low-touch SMBs like yours absolutely exists—but it’s small, and hard to identify upfront. The problem is that the long tail as a whole tends to generate minimal revenue while driving disproportionately high support and product complexity.

Working for a business with clients, I vastly prefer larger clients.

Larger clients usually have better in-house knowledge of our own products, leading to LESS support tickets.

Smaller clients think they are big dogs and are typically very demanding of support for very little pay. Due to these small companies abusing free support, we have had to introduce paid SLA levels.

Totally agree. Small clients need the most hand holding we've found so they're cut loose pretty quickly. Seems BRCM does the same with VMWare.

unfortunately you can't stop broadcom from vendors you have agreements with

do people make new deployments with vmware these days? I haven't really dealt with that side of things in a good while but i stopped using vmware at home (migrated everything to virtualbox) about a decade ago and it's not something that's crossed my mind since then. my impression is that vmware is just used by big companies that have legacy entrenchments and there's no reason to pick it as something to build atop in 2025

I would say VMware is very comfortable infrastructure and despite the fact Hyper-V is adequate and effectively free for a Windows shop, I would prefer VMware.

That being said, when the alternative is free, VMware really doesn't have the power it thinks it has to dictate terms.

Free doesn't allow for selling support contracts which is what BRCM is after. Good luck with Hyper-V if that's the direction you go in. Free doesn't generally include any support.

To be clear, Hyper-V is not "free" as in "no cost". It's "free" as in "no additional cost": Most who have Windows VMs in a virtual datacenter have Windows Server Datacenter licensing which allows infinite VMs per processor core. So if you have Windows Server VMs, you can already use Hyper-V as your hypervisor, but many still prefer VMware because it's seen as more stable and secure than using Windows hypervisors underneath your Windows VMs. VMware certainly adds value, but not infinite value, when most of its customers already are licensed for Hyper-V anyways.

Be careful with VirtualBox outside tiny home labs. If you thought Broadcom was a pain in the neck, wait until you talk to Oracle.

I thought oracle licensing shenanigans only applies if you used their extension pack? If you only used the community/GPL version you should be fine.

> I thought oracle licensing shenanigans only applies if you used their extension pack?

I believe that's correct.

The thing one must be cautious regarding is installing the Extension Pack without careful consideration.

That’s probably true, but the idea of using only the GPL version throughout an organization, then being asked to demonstrate to Oracle’s satisfaction that you don’t have one single extension pack anywhere, is enough to make me leery of it.

My impression -- and this is entirely based on superficial observations and a massive amount of extrapolation, so I may be entirely off-base -- was always that VirtualBox's priority was always speed (security was nominally there, but felt retroactive), and VMware's was always security (speed was nominally there, but felt retroactive). Would be nice if someone could clarify whether this was/is accurate.

>(migrated everything to virtualbox)

Virtualbox is not a replacement...

When businesses say "vmware" they mean esxi/vsphere to manage large numbers of VMs, not desktop VMs, not vmware workstation... The replacement would be hyper-v, proxmox, xen or moving workloads to Microsoft 365, Azure, AWS or GCP.

Besides, vmware workstation is also vastly better and also given out for free now.

This has nothing to do with virtualbox. VMware workstation is the desktop product and is given away for free since it's practically insignificant to the portfolio.

This is all about managing and deploying applications across multiple servers (vSphere) which is a fleet of computers running ESXi as their operating system.

Enterprise is the focus for VMWare ESXi nowadays. Expensive support but fewer support accounts. Broadcom isn't interested in volume.

KVM is better anyway.

Except it isn’t for the target customer base of VMware.

Reply on news.ycombinator.com