For many CIOs, the VMware question is no longer whether to act but how fast, and how much risk to take on while doing it. New research from Unisphere Research, sponsored by Rimini Street, gives a data-backed picture of how enterprise leaders are answering. The survey of 269 current VMware users was fielded between December 2025 and February 2026. Because 31% of respondents were the primary decision-makers for IT virtualisation, the report focuses on that group’s intentions, which means the headline figures come from a subset of roughly 80 senior leaders and should be read as directional rather than definitive.
Cost is pushing people out, and the clock is a second driver
The most striking number is the share of decision-makers considering leaving VMware/Broadcom because of higher licensing costs: 90%. The respondents tie that to forced product bundles and subscription packaging. A majority, 54%, also cite the end of general support for VMware perpetual licences, which the report places at October 2027. Support for vSphere 7.x and related products already ended in October 2025, so for some organisations that deadline has already passed.
Yet the survey describes a market that is hesitating rather than stampeding. Among current VMware perpetual licence users, 48% have no plans to convert any assets to VMware Cloud Foundation (VCF) today. That share falls to 38% over three years, while the proportion planning to put all assets on VCF rises from 11% to 27%. In other words, a meaningful slice of the base will end up on Broadcom’s subscription platform, but a large group is still undecided or resisting. CIOs who assume a single, uniform outcome across the industry are probably planning against the wrong baseline.
The destination is a mixed estate, not a single replacement
The report’s strongest finding for architecture planning is that organisations are not making wholesale exits. They are diversifying. Today, 60% of VMware users are shifting some assets to a non-VMware hypervisor, and 10% have moved everything. Within three years, the share that have moved all assets to another hypervisor is expected to reach 34%, and 82% expect at least some of their estate to run outside VMware.
The shift in where workloads live is just as notable. The share of respondents running a majority of workloads on-premises is expected to fall from 42% to 37% over three years. Those running a majority in public cloud nearly double from 13% to 23%, and the share running a majority in private cloud rises from 12% to 18%. Containers follow the same curve: respondents running most of their workloads on containerised approaches grow from 10% to 18%.
Notably, containers are not replacing virtual machines so much as joining them. Close to half of decision-makers, 47%, favour a best-of-both-worlds approach that pairs container agility with the isolation and management tooling of established virtualisation platforms. The reasons given for moving to a hybrid hypervisor-plus-container model are operational efficiency (65%) and lower costs for cloud-native applications (53%). For CIOs, the practical implication is that platform decisions should be made workload by workload, not as a single enterprise-wide bet.
Cost leads the roadmap, but flexibility is close behind
When asked about the goals of their virtualisation roadmaps, 73% of top decision-makers named cost savings across capital and operating expenditure. Scalability and flexible capacity came second at 64%, and modernisation third at 43%. Performance gains and faster server provisioning each landed at 37%. The ordering matters: cost may be what gets the project funded, but flexibility is what most leaders expect the new architecture to deliver.
The market backdrop supports that ambition. The report cites Mordor Intelligence in sizing the global IT virtualisation market at about $110 billion this year, growing to $225 billion by 2031 at a compound annual growth rate above 15%. Capital is flowing towards alternatives, which means more options, but also more vendors to evaluate.
The barriers are operational, not technical
The obstacles respondents cite should temper any optimism about an easy migration. Operational complexity tops the list at 40%, ironically for a technology meant to simplify IT. Multi-vendor management follows at 38%, then an expanded attack surface and skills requirements at 37% each. Regulatory and compliance concerns and vendor end-of-support both sit at 34%, hypervisor sprawl at 31%, and vendor lock-in at 29%. When asked in open-ended questions where they need help, respondents ranked migration planning and implementation first, followed by security updates and patching, closing skills gaps, real-time visibility, and integrating AI.
That ranking is a useful planning checklist. The hardest problems are not choosing a hypervisor. They are executing a migration, staffing a more heterogeneous environment, and keeping it secure while the change is under way.
Security is the pressure point
The security findings may be the most actionable for CTOs. Close to two-thirds of respondents are dissatisfied with Broadcom’s patching programme, and 77% said they would value a solution that protects them proactively against vulnerabilities rather than depending on that programme. The report frames the risk plainly: in a virtualised environment, a single unpatched hypervisor can expose hundreds of guest machines, and leaders admit manual patching is too slow against modern zero-day exploits.
The operational cost of the current model also shows up in the open-ended responses. Many describe responding to critical vulnerabilities as highly disruptive, with some reporting that operations or employee productivity can stop for up to a day, particularly for platforms such as ERP. Integration with existing security tools was the most commonly cited security challenge, ahead of the need to hire consultants and compliance burdens. AI-powered attacks also made the top five.
The report’s recommended direction is a shift from reactive patching to runtime protection: monitoring hypervisor behaviour while it runs, using techniques such as memory introspection, micro-segmentation and virtual patching, so that a critical vulnerability becomes a background task rather than a business-halting event. Whether or not a given organisation adopts that approach, the underlying question is worth asking now: how long does it take your team to move from a disclosed hypervisor vulnerability to a deployed mitigation, and what does that cost the business?
Third-party support is going mainstream
Reflecting these pressures, 66% of respondents either use third-party support (18%) or are considering it (48%). The reasons are not primarily about price. Security is the top function at 60%, followed by installation and implementation at 57% and service and support issues at 54%. The report also notes that 21% of organisations hit service support issues weekly or more often, and that few rate their OEM support teams as very skilled.
One word of caution. This research was sponsored by Rimini Street, which sells third-party VMware support and hypervisor security, and the report’s closing pages describe those products. The data on how many leaders are considering alternatives is useful in its own right, but CIOs should evaluate any provider on its own evidence. The report itself advises looking for demonstrated expertise across heterogeneous environments, proactive vulnerability management, strong service-level commitments, and strategic guidance for modernisation.
AI ambition is outrunning AI skills
The final data point concerns readiness. Among respondents, 61% see AI’s role in their virtualisation roadmap as supporting AI and machine learning workloads, 47% see predictive analytics and proactive maintenance, and 40% see enhanced security and compliance. Only 22% are looking at autonomous AI agents today. But 88% say their teams lack the full skills to implement AI within their virtualisation framework, with 60% saying they have them only partially and just 12% saying they have them mainly. The report warns this partial knowledge can lead to ‘pilot purgatory’, where experiments never scale.
What this means for your next 12 months
Taken together, the data suggests a sequence rather than a single decision. Start by mapping workloads against your support deadlines, since the October 2027 date for perpetual licences is the one hard constraint in the research. Decide workload by workload what stays on VMware, what moves to another hypervisor, and what belongs in containers or cloud, because the survey shows most peers expect a mixed estate. Treat security as a parallel track rather than an afterthought, given how many respondents flagged patching disruption and integration gaps. Budget for people as well as platforms: the skills shortfall, cited for both virtualisation and AI, is the barrier most likely to slow a migration that looks sound on paper. And if you evaluate third-party support, test it against the criteria your own risk profile demands, not just the cost comparison.
The report’s own conclusion is that the future of virtualisation will be defined by flexibility, cost efficiency and reduced dependence on any single vendor. The organisations that get there will be the ones that plan the transition as carefully as they plan the destination.