AI scale, hybrid cloud drive virtualisation reset: HPE survey

As artificial intelligence (AI)-driven workloads increase operational complexity and push up cloud expenditure, enterprises are reassessing whether their existing infrastructure can sustain competitiveness in a hybrid, data-intensive environment.

That reassessment is no longer theoretical. Across industries, IT leaders are drawing up plans to re-architect virtualised estates that were designed for a different era, one that is defined by predictable workloads, stable licensing models and slower growth in data volumes.

More than two thirds of global enterprises intend to make material changes to their virtualisation strategy within the next two years, yet only 5% consider themselves fully prepared for the transition, according to a survey by Hewlett Packard Enterprise (HPE).

The research, conducted between 15 December 2025 and 4 January 2026, includes insights from over 400 experts, including chief information officers, chief technology officers and senior IT decision-makers.

Beyond Cost the Operating Model Question

While recent licensing changes have added urgency, the survey indicates that cost alone is not the central driver, with only 4% of respondents citing licensing costs as the single biggest catalyst for rethinking virtualisation.

Besides, the increase in the cost of virtual machines (VMs) may have acted as the forcing function, but it exposed deeper structural issues—hybrid complexity outpacing operational models, vendor lock-in and environments that no longer scale efficiently for AI.

Fewer than one in 10 enterprises (4%) cite licensing costs as the primary trigger, underscoring that the urgency is less about price and more about shifting towards a hybrid operating model and preparing infrastructure for AI workloads. Companies are taking a deliberate approach rather than opting for a direct hypervisor swap.

More than half (57%) of enterprises are adopting a phased strategy to futureproof their IT environments. Hybrid cloud is emerging as the preferred path to meet AI performance requirements. Currently, 78% of provisions sit in the public cloud, 61% in virtualised clusters, 48% in private cloud environments and 32% at the edge.

The findings suggest that the reset is not about replacing one hypervisor with another, but about redefining the operating model itself and becoming agnostic to hypervisor, hardware or hyperscaler. It is also about making deliberate decisions about where workloads reside and how confidently they can be run across environments.

AI Readiness Reshapes Priorities

Enterprises overwhelmingly see advanced AIOps or Artificial Intelligence for IT Operations as central to modern virtualisation strategy. Over a quarter of IT decision-makers rank AI readiness as their top priority as they seek to build systems capable of meeting the performance, data and scaling demands of AI.

When shaping future virtualisation and private cloud strategies, 70% of enterprises say unified backup and cyber recovery are very important or business critical. Cross-platform governance is cited by 61%, while 55% identify integrated observability and AIOps as essential capabilities.

“We are seeing enterprise leaders reassess longstanding IT assumptions to balance cost predictability, AI readiness and performance. This isn’t just a rush to rip and replace, but a deliberate shift toward a more flexible and simplified operating model,” said Brian Gruttadauria, CTO of Hybrid Cloud at HPE.

The data indicates that the “Great VM Reset”, as HPE calls it, is less about swapping hypervisors and more about redefining where workloads reside and how consistently they can be governed across environments.

Most respondents indicated that budget constraints (28%), technical complexity (24%), migration risk (21%) and skills gaps (20%) are slowing that transition, underscoring that the reset is as much organisational as it is technological.

The image accompanying this story was created using AI. The article was edited with limited use of AI-based tool.