Should you build or buy Hyperconverged Infrastructure (HCI)?
Hyperconverged infrastructure software pools compute, storage and networking on standard servers into a single managed layer, replacing separate SAN storage and hypervisor stacks. IT infrastructure and platform teams use it to run virtual machines with built-in software-defined storage, simplified lifecycle management and disaster-recovery tooling across a data center.
Can we run our own hyperconverged stack instead of paying Nutanix or Broadcom? Yes, and named organizations do it in production on Proxmox VE with Ceph, with reported license savings of 60% to 75%. Be careful before renewing: the vendor is selling convenience and support on a capability your team can replicate, so treat the renewal as negotiable and pilot the alternative on a non-critical cluster before the next contract cycle.
Build it, buy it, or bridge?
When building makes sense
A self-run stack is now a documented production choice, not a hobbyist experiment. Proxmox VE with Ceph runs named clusters at hosting providers, a municipality, a telecom operator and multiple mid-size companies, and the consultancy CodiLime migrated a global cybersecurity company's environment of over 10,000 VMs from VMware ESXi to Proxmox VE with near-zero user-visible disruption. The work is compute, software-defined storage and a hypervisor on standard servers, the same stack at every company, so a team with Linux, storage and on-call depth can replicate what a vendor sells without inventing anything new. Building earns its keep once you can staff Ceph design and 24/7 on-call yourself, and it pays fastest for teams already absorbing a VMware or Nutanix renewal shock. The risk is real, though: this is infrastructure everyone depends on, and a shaky bench turns savings into downtime.
When buying makes sense
Buying still makes sense for teams that lack deep storage and on-call skills, because Nutanix, VCF and HPE SimpliVity bundle validated hardware lists, integrated lifecycle management and single-vendor support that a thin ops team cannot replicate safely. That support has real value when something breaks at 2 a.m. and there is one number to call instead of a Ceph mailing list. But buy knowing the alternative exists: this category is buildable and undifferentiated, and named organizations have reported large licensing reductions after moving off VMware's post-Broadcom pricing, even after courts and regulators documented price increases running into the hundreds of percent for some customers. Treat the next renewal as negotiable rather than automatic, and price out a Proxmox pilot on a non-critical cluster before you sign. The honest reason to keep paying is your team's bench strength, not any capability the vendor holds that you cannot build.
The desk read
Build-versus-buy analysis for Hyperconverged Infrastructure (HCI) is being written. In the meantime, the framework that drives every B4 call is on the B4 Index page.
Frequently asked
Is running our own HCI stack actually realistic in production?
Yes. Named organizations run Proxmox VE with Ceph in production today, including a 10,000-plus-VM migration the consultancy CodiLime ran for a global cybersecurity client, off VMware with near-zero disruption. The technology is proven; the question is whether your team has the storage and on-call depth to run it.
How much can switching off VMware or Nutanix actually save?
Court filings and regulatory reports document VMware price increases running from roughly 85% to over 1,000% for some customers after Broadcom's repricing, and named migrations to Proxmox report large licensing cuts. Treat vendor-published savings figures as directional, and build your own three-year comparison before deciding.
What are we actually paying for if the software is replicable?
Mostly convenience: certified hardware compatibility lists, integrated lifecycle tooling and a single support contract when something fails. None of that is a capability gap your team cannot close, which is why the renewal is worth treating as negotiable rather than a clean buy.
Should every team switch to open-source HCI now?
No. Teams without strong Linux, storage and on-call skills should still expect to buy, since Ceph in production is unforgiving of gaps in that bench. The point is to negotiate the renewal knowing the alternative is real, not to migrate blind.