Server Virtualization Platforms

As recently as last year, Gartner's position on this market was that it did not need a quadrant of its own. The hypervisor had become a feature of a larger infrastructure decision, and the scorecard that governed the estate carried a wider name.

That position has been reversed. On 14 September 2026, Gartner published a Magic Quadrant for Server Virtualization Platforms, a standalone evaluation of a market the firm had spent several years covering inside its distributed hybrid infrastructure chart.

The reversal is the story. Gartner does not restore a retired category because vendors asked. It restores one when the buying decision underneath it becomes contested again, and there is no serious argument that server virtualization has become one of the least contested layers of enterprise infrastructure. Broadcom's acquisition of VMware turned a renewal that most organizations treated as administrative into a re-evaluation of the whole estate. The quadrant came back because the question came back.

The name came back after a decade

The history matters here, because the gap is what gives the new edition its shape.

Gartner's earlier coverage of this market sat under narrower names and treated the hypervisor as a market in its own right. As the technology matured, the firm shifted its attention outward. The estate became hybrid by default, spanning on-premises hardware, public cloud, and edge locations, and Gartner built the distributed hybrid infrastructure evaluation to score the vendors selling that whole posture rather than the hypervisor inside it.

Read from the outside, that looked like a demotion. Virtualization had become a capability rather than a category, and the vendors who sold it were being scored on how well they could sell everything around it. That reading was reasonable, and for a period it was correct.

What changed is licensing. Broadcom restructured VMware's pricing around per-core subscriptions and narrowed the licensing options available to customers, and the practical effect was that a large population of organizations that had not run a hypervisor evaluation in years suddenly had a reason to run one. Gartner's own framing, as reported alongside the new edition, is that the market now faces its most significant disruption in more than a decade.

A disruption of that kind reopens a decision that had closed. That is the condition under which a category earns its own quadrant back.

Inside the September 2026 Magic Quadrant for Server Virtualization Platforms

The edition is dated 14 September 2026, with the quadrant plotted as of August 2026, and it is authored by Tony Harvey, Daniel Bowers, Paul Delory, Tony Iams, and Owen Marino.

Seventeen vendor solutions met the inclusion criteria out of more than forty that Gartner tracks in the market. The bar was set in three parts: mandatory product features had to be generally available by 1 May 2026, a vendor needed a minimum of ten production customers per region in at least four of seven defined geographic regions, and the product had to cover at least four of the six use cases defined in the companion Critical Capabilities research.

The field breaks down across all four quadrants.

Leaders. Nutanix, Broadcom (VMware), Microsoft, and Red Hat. Nutanix is placed highest on Ability to Execute. Broadcom, which enters under the VMware name it acquired, sits furthest right on Completeness of Vision.

Challengers. HPE, Canonical, and Oracle.

Visionaries. SUSE, Platform9, H3C Technologies, IDT SYSTEM, and Sangfor Technologies.

Niche Players. Proxmox, Vates, VergeIO, Citrix, and Scale Computing.

Two features of that roster are worth pausing on. The first is that the Leader set is small relative to the field, which is normal in a mature infrastructure category and unusual in a market Gartner has just described as disrupted. The second is that four of the five Visionaries are not primarily North American or European vendors, and one of them, H3C Technologies, appears alongside Sangfor Technologies as part of a Chinese infrastructure cohort that has been largely invisible in Western hypervisor evaluations.

What Broadcom's licensing did to the placement table

The quadrant's most interesting movement is not a vendor changing tier. It is that pricing models became a scoring surface at all.

Gartner cites per-core subscription cost complaints and reduced licensing choice as the drivers of renewed competition, and the alternatives it names are instructive because they are not all the same kind of answer. Canonical, HPE, Oracle, Vates, VergeIO, H3C Technologies, and Sangfor Technologies are identified with alternative pricing models. Proxmox is identified with a different structure entirely: a downloadable open-source stack with paid support subscriptions.

Those are two distinct buyer arguments, and conflating them is a common mistake. A vendor offering a different commercial model still sells a supported enterprise product with a contract, a roadmap, and a renewal conversation. A vendor offering an open-source stack with paid support sells a support relationship and leaves the product itself in the customer's hands. The first is a pricing answer. The second is an ownership answer.

The placement table reflects that both are now live in the market. It does not say which one wins, and in practice the two tend to serve different estates. Organizations with a small number of high-criticality workloads and a compliance surface to defend tend to want the first. Organizations running large, relatively homogeneous fleets, particularly in lab, education, and hosting environments, have been quietly demonstrating demand for the second for years.

Seventeen vendors, and the three pricing answers

Read the whole field against the licensing question and it sorts into three groups rather than four quadrants.

The first group defends the incumbency position: the hypervisor is enterprise infrastructure, it should be licensed like infrastructure, and the value is in the operational maturity of the platform. That is the Broadcom position, and to varying degrees the Microsoft, Nutanix, and Red Hat position as well. All four are Leaders, which means Gartner's evaluation currently rewards it.

The second group competes on commercial terms. Canonical, HPE, Oracle, Vates, VergeIO, H3C Technologies, and Sangfor Technologies are offering a structurally different price, and their argument is that the technical capability gap has narrowed enough that price can decide the purchase. Their placement across Challengers, Visionaries, and Niche Players reflects that the argument is credible but not yet sufficient on its own.

The third group competes on ownership. Proxmox, and to a degree Scale Computing, sell a product the customer controls rather than a subscription the customer rents. That position has historically been dismissed in enterprise evaluations as a niche preference. Its persistence in a quadrant that required ten production customers per region in four of seven regions suggests it is not as niche as the dismissal implied.

The scope now reaches Kubernetes and confidential computing

The most consequential detail in the new edition is not the vendor list. It is what Gartner has decided the category now contains.

The scope covers Kubernetes integration, confidential computing, GPU support, and non-x86 architectures including ARM, IBM Power, and RISC-V. Every one of those is a departure from the classic definition of a hypervisor, which was software that abstracted x86 hardware so virtual machines could run on it.

Kubernetes integration means the quadrant is now scoring how a virtualization platform coexists with container workloads on the same estate, which is the question most infrastructure teams are actually arguing about internally. Confidential computing means the hypervisor is being evaluated as a security boundary rather than an efficiency layer. GPU support means the platform is being scored on how it handles the workload that is currently consuming most of the incremental infrastructure budget. And non-x86 support means the assumption that the estate is Intel or AMD is no longer safe.

A category defined this broadly is not the same market its name describes. Buyers comparing the September 2026 placement table to their memory of an earlier quadrant are comparing two different evaluations.

Where the distributed hybrid chart fits now

The distributed hybrid infrastructure quadrant has not gone away, and the relationship between the two charts is the part most likely to confuse a buyer.

Distributed hybrid infrastructure scores the vendor on the whole posture: on-premises, cloud, and edge, and the control plane that ties them together. Its September 2026 edition, published a week earlier on 7 September 2026, named AWS, Microsoft, Broadcom, Oracle, and Nutanix as Leaders for the third consecutive year, with Google, Red Hat, and SUSE as Visionaries, IBM and Alibaba Cloud as Challengers, and Huawei and Tencent Cloud as Niche Players.

Server virtualization platforms scores the vendor on the layer beneath that, which the distributed hybrid chart treats as one component among several.

The practical consequence is that the same vendor can appear in both charts with different placements and a different set of competitors, because the two evaluations ask different questions. A buyer deciding how to run a fleet of virtual machines should read the new quadrant. A buyer deciding which vendor should own the estate's control plane should read the older one. Reading either as a verdict on the other will mislead.

What the ranking cannot tell a buyer

Two limitations are worth stating plainly.

The first is that a quadrant scores a vendor's position, not a specific product's fitness. The companion Critical Capabilities research does the use-case scoring, and the four-of-six use case coverage requirement means most vendors on the chart have been scored on use cases that are not yours. The quadrant is a field map. It is not a shortlist.

The second is that the licensing disruption driving the new edition is still in progress. Several vendors on this chart are running pricing models that are themselves new, and some of the alternative pricing identified by Gartner is recent enough that the commercial terms have not been tested through a full renewal cycle. A placement earned during a pricing disruption is a snapshot of a market in motion, not a settled ranking.

There is a third point that Gartner's own inclusion criteria make for you. Requiring ten production customers per region in at least four of seven regions is a meaningful bar, and it is also a bar that measures breadth rather than depth. A vendor can clear it while the specific estate you run looks nothing like the customers that got them there.

Four questions before you renew the hypervisor

Ask what the last renewal actually cost, per core and in total, including the support tier that made the platform workable. Organizations that switched after the Broadcom repricing frequently discover the comparison they ran understated the incumbent's real number, which means the alternative looked better than it was.

Ask whether the alternative pricing is a contract or a stack. The two carry different exit costs, different support relationships, and different upgrade obligations, and vendors on this chart are offering both.

Ask which of the six Critical Capabilities use cases your estate actually matches, and read the scoring for that one rather than the quadrant placement for the vendor. A Niche Player with a strong score in your use case is a legitimate shortlist entry.

Ask what the migration costs, in engineer time, before comparing license numbers at all. The licensing disruption made the license fee salient. It did not make it the largest line in the comparison.

Gartner does not endorse any vendor, product or service depicted in its research publications, and does not advise technology users to select only those vendors with the highest ratings or other designation. Gartner research publications consist of the opinions of Gartner's research organization and should not be construed as statements of fact. Gartner disclaims all warranties, expressed or implied, with respect to this research, including any warranties of merchantability or fitness for a particular purpose.

The companion piece on this site is Strategic Cloud Platform Services. The 2026 quadrant published on 1 September with five Leaders instead of three: Alibaba Cloud enters the top band for the first time as the only Asia-Pacific vendor in it, and AWS's share slips to 39.8 percent.

The related category on this site is Infrastructure Automation Platforms. Forrester's 2019 evaluation included HashiCorp. Its 2020 evaluation did not, right as Terraform became the default way organizations provisioned cloud infrastructure. Analyst vendor lists describe who competes for procurement, not always what engineers actually use.