Platform migration shortlists reliably contain the same two or three names, and reliably omit the hypervisor that ships with an operating system most of these organisations already license heavily.
The omission is usually inherited opinion from around 2012, when the gap was real. It has not been re-examined since.
What the case actually is
Licensing overlap. If you run Windows Server workloads, you are already paying for Windows Server licensing. Datacenter edition entitles unlimited virtual instances on the licensed host. The marginal cost of the hypervisor in that scenario is small, and for organisations with an enterprise agreement the arithmetic can be striking.
This is the whole argument and it is a strong one. It is also the argument that requires somebody to actually model it rather than assume.
Windows workloads are first-class. If your estate is predominantly Windows, running it on the vendor’s own hypervisor removes a support boundary. Integration with Active Directory, Windows-native management, and familiar tooling are real operational advantages for a Windows-centric team.
The features are there. Live migration, clustering, replication, storage integration. The checklist that mattered in 2015 is filled in.
Azure adjacency. If part of your strategy involves Microsoft’s cloud, the management and hybrid story is more coherent than it is from a third-party hypervisor.
Where it is genuinely weaker
Linux at scale. It runs Linux and it runs it fine. The ecosystem, the tooling, and the operational familiarity are weaker, and a predominantly Linux estate is not the natural fit.
Third-party integration breadth. The ecosystem of backup, monitoring and automation tooling is narrower than the incumbent’s, though the major products all support it. Check your specific tools.
Management at large scale. Managing a very large estate needs the full management stack, which is a product with its own complexity and its own learning curve. Organisations expecting a single-pane experience out of the box find the assembly required.
Institutional knowledge. Your team probably has ten years of the other platform in their hands. That is a real switching cost regardless of the technology.
The thing that has changed
Licensing economics shifted sharply, and a hypervisor whose incremental licence cost is near zero for a Windows estate now looks different from how it looked when the incumbent’s pricing was merely annoying.
That is not a technical argument and it is the one that should put it on the shortlist. The technical question — is it good enough to run our workloads — has a fairly clear answer for Windows-dominated estates, and the answer is yes.
How I would evaluate it
Model it as a first-class option with the same rigour as the others. Same growth assumptions, same headroom, same migration effort estimate, same operating model analysis.
Then be honest about the Linux proportion of your estate and about your team’s skills. If you are seventy percent Windows with a Windows-centric team, the case is strong. If you are seventy percent Linux with a Linux team, it is weak, and that is a legitimate conclusion reached properly rather than by reflex.
The failure I want to prevent is the one where an organisation pays a great deal of money to migrate to a third platform without ever having priced the one included in software they already buy. That happens, and the reason is always that nobody put it on the list.