Free tools Windows power users keep installed
One-click scans. No signup required.
Gartner’s 35% figure is a forecast, not a measured count of VMware customers that have already left. Gartner research vice president Julia Palmer reportedly made the prediction at the September 2025 Gartner IT Symposium in Gold Coast, Australia. Public reporting does not disclose the forecast’s denominator, methodology, sample, confidence range, or a precise definition of “workload.” The context appears especially relevant to VMware workloads delivered through hyperscalers, where Broadcom’s licensing and channel changes may give cloud providers stronger incentives to steer customers toward native services.
That makes the number a warning about possible large-scale movement by 2028—not proof that 35% of every VMware virtual machine will be replaced by another hypervisor. Each organization still has to decide which systems to retain, move, modernize, retire, or place in a different cloud.
What the 35% prediction actually says
Ars Technica reported that Julia Palmer predicted 35% of VMware workloads would migrate elsewhere by 2028 after her Gartner IT Symposium comments. The report does not establish whether “workloads” means virtual machines, applications, customers, hosts, clusters, licensed capacity, or another unit. It also does not specify whether “elsewhere” includes a different hypervisor, public-cloud infrastructure, cloud-native services, retired systems, or a move between VMware-operated environments.
The available public material does not establish a global scope, a customer segment, a regional limit, or whether 2028 means the end of that calendar year or an approximate forecast horizon. It is therefore inaccurate to write that Gartner found 35% of VMware customers had already defected. The defensible reading is an analyst prediction about a substantial potential shift, particularly among workloads obtained through hyperscaler channels. Ars Technica’s report is the public source for the attribution.
The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →#1 Best Overall
- MODEL P74439-005: Compact and affordable HPE ProLiant MicroServer Gen11 powered by Intel Pentium Gold G7400 3.7GHz processor, ideal for file sharing, NAS, and basic business workloads
- READY OUT OF THE BOX: Includes 16GB DDR5 UDIMM memory (expandable to 128GB), one 1TB SATA 6G Business Critical HDD, embedded Intel VROC SATA, dedicated iLO-M.2 port kit, 180w external power adapter and 1/1/1 warranty for dependable plug-and-play server operation
- WHISPER-QUIET & SPACE-SAVING: Ultra-compact mini tower design fits easily in small office spaces; supports wall, flat, or vertical placement for deployment flexibility
- INTEGRATED REMOTE MANAGEMENT: Comes with HPE iLO 6 and embedded TPM 2.0 for secure, license-free remote server administration through shared port access
- EXPANDABLE DESIGN: Two PCIe slots (including PCIe 5.0) and four LFF-NHP drive bays provide robust options for storage and component scalability. Features new MR408i-p controller support for enhanced storage performance
A separate CloudBolt survey reported by Ars Technica in February 2026 indicates active footprint reduction among respondents, but it is not an independent validation of Gartner’s forecast or a census of the VMware market. Respondents reported that 36% had migrated 1–24% of their environment, 32% had migrated 25–49%, 10% had migrated 50–74%, and 2% had migrated at least 75%. The survey also reported public-cloud IaaS as a destination for 72% of migrated workloads and Microsoft’s Hyper-V/Azure stack for 43%. Those percentages are respondent-reported and may allow multiple destinations. See the survey coverage.
Why Broadcom’s VMware changes are prompting reassessment
Broadcom completed its VMware acquisition in November 2023. Post-acquisition changes reported in the public coverage include a move away from perpetual licensing toward subscriptions, consolidation into fewer and more expensive bundles, and a narrower reseller and channel model. Customers—especially smaller organizations—may face higher costs, fewer purchasing options, and greater dependence on Broadcom-controlled sales channels.
Those commercial changes explain why buyers are reviewing alternatives; they do not prove that a particular workload can be moved at a lower total cost. A VMware renewal can still be less risky than a migration when an application is certified only on vSphere, when a team has deep VMware expertise, or when compliance and downtime requirements make conversion expensive.
Why hyperscaler-delivered VMware is especially exposed
Hyperscalers want to supply infrastructure, managed services, and the commercial relationship with the customer. The reported VMware changes can reduce that control by requiring customers to obtain VMware licensing directly from Broadcom while the cloud provider supplies the underlying infrastructure. A March 2026 migration analysis described this as a Broadcom-controlled licensing relationship and attributed November 1, 2025 as the effective date for a major cloud-provider resale restriction. That date and the applicable eligibility and portability rules should be checked against current Broadcom and cloud-provider documentation before a contract decision.
This creates a strategic tension: AWS, Microsoft, Google Cloud, and other providers can continue supporting VMware workloads while also encouraging customers to use native virtual machines, databases, containers, and other services. A customer may therefore leave a VMware-based service because the economics and vendor incentives changed, not because vSphere technically failed. The analysis is described in Technologent’s migration discussion.
Where VMware workloads can go
| Destination | Best fit | Main trade-off |
|---|---|---|
| Public-cloud IaaS | Variable demand, global distribution, data-center exit, or applications ready for managed services. | Consumption charges for compute, storage, backup, networking, and egress can exceed on-premises cost for steady, high-utilization workloads. |
| Microsoft Hyper-V and Azure Stack HCI | Organizations standardized on Windows Server, Active Directory, Azure operations, and Microsoft licensing. | Creates or deepens Microsoft dependency and requires validation of non-Microsoft tooling and workloads. Product information: Azure Stack HCI. |
| Nutanix AHV / Nutanix Cloud Infrastructure | Private-cloud or hyperconverged replacement seeking integrated infrastructure management. | Requires a new management model, skills, commercial terms, and hardware or appliance evaluation. Nutanix Cloud Infrastructure. |
| Red Hat OpenShift Virtualization | Organizations combining virtual machines with containers and Kubernetes-oriented modernization. | Not a low-change vSphere substitute; operating practices and skills may change substantially. OpenShift Virtualization. |
| KVM-based platforms, Proxmox VE, or OpenStack | Buyers prioritizing licensing flexibility, control, or open architectures. | License savings can be offset by support, engineering, lifecycle, security, automation, and training costs. |
| Remain on VMware | Environments with deep vSphere dependencies, regulatory constraints, latency needs, or high migration risk. | Subscription and bundle costs may rise, and channel choice may be narrower. |
Public-cloud starting points include Amazon EC2, VMware Cloud on AWS, Azure Virtual Machines, and Google Compute Engine. None is a universal feature-for-feature replacement for every VMware deployment.
Rank #2
- HIGH-EFFICIENCY SERVER FOR BUSINESS-CRITICAL AND VIRTUALIZED WORKLOADS: HPE ProLiant ML350 Gen11 (P69313-005) powered by Intel Xeon Gold 5416S (16 cores, 2.0GHz) with 64GB DDR5 memory and 8 SFF drive bays, delivering improved performance for virtualization, databases, and application consolidation
- PROCESSOR – XEON GOLD FOR HIGHER PERFORMANCE AND EFFICIENCY: Intel Xeon Gold 5416S (16 cores, 2.0GHz) delivers improved performance, cache optimization, and workload efficiency compared to entry-level CPUs, enabling virtualization clusters, database environments, and application consolidation with greater reliability.
- MEMORY – 64GB DDR5 WITH ENTERPRISE-LEVEL SCALABILITY: Includes 64GB DDR5 HPE SmartMemory (2×32GB RDIMM), expandable up to 8TB across 32 DIMM slots, delivering high bandwidth, improved efficiency, and scalability for memory-intensive workloads and long-term infrastructure growth.
- STORAGE – SSD PERFORMANCE WITH FLEXIBLE 8SFF EXPANSION: Configured with 2×480GB SATA SSDs and 8 SFF drive bays, paired with HPE MR408i-o RAID controller (4GB cache) supporting RAID 0/1/10, enabling fast data access, reliable protection, and scalable storage for business-critical applications.
- EXPANSION – PCIe GEN5 PLATFORM FOR I/O AND ACCELERATION: Supports PCIe Gen5 expansion and OCP 3.0 connectivity, enabling upgrades for high-speed networking, storage, and GPU acceleration to support workloads such as VDI, analytics, and compute-intensive applications
Migration is more than copying a virtual disk
Rehost
A rehost moves a VM with little application change. It is often the quickest route, but it preserves technical debt and can reproduce inefficient licensing or infrastructure usage.
Convert
Conversion changes the VM disk and configuration for another hypervisor or platform. The guest may boot while still losing drivers, network behavior, storage integration, backup support, or performance. Boot success is not acceptance testing.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Fix the driver behind crashes, sound loss and screen glitches3Repair Windows errors before they cause bigger problemsReplatform
Replatforming moves an application to a managed runtime, database, or virtualization service. It can reduce infrastructure administration but introduces provider-specific dependencies and requires application testing.
Refactor or modernize
Refactoring redesigns the application for containers, managed databases, serverless functions, or other cloud-native services. It can produce the best long-term result for suitable systems, but it is normally slower and riskier than rehosting.
Retire or replace
Some development systems, obsolete applications, duplicate services, or low-value disaster-recovery copies can be consolidated, replaced with SaaS, or shut down. Retirement requires business-owner approval, data-retention analysis, and confirmation that no recovery obligation remains.
Dependencies that make VMware moves difficult
- vMotion and live-migration assumptions, snapshots, VM hardware versions, VMware Tools, and virtual device drivers.
- Distributed switches, NSX networking and security policies, vSAN, storage replication, and VMware-specific disaster recovery.
- Backup and restore products, monitoring, observability, identity integration, access controls, and operational runbooks written for vSphere.
- Clustered databases, latency-sensitive applications, specialized hardware, GPU or NUMA requirements, and appliances certified only on VMware.
- Operating-system, database, application, backup, and security licenses tied to hosts, sockets, cores, virtual CPUs, or a particular virtualization technology.
- Regulatory controls such as PCI, HIPAA, or FedRAMP that require target-platform approval and evidence after the move.
Before selecting a destination, map every workload to its application owner, dependencies, business criticality, recovery objectives, licensing exposure, and target-platform fit. Include production, test, backup, and disaster-recovery environments in the inventory.
Recommended Free Tools
Rank #3
- HPE ProLiant ML30 G10 Plus Tower Server, perfect for small businesses and remote offices
- Xeon E-2314 4-Core 2.8GHz 8MB CPU, Turbo up to 4.5GHz
- Memory: 32GB (2 x 16GB) DDR4 PC4-25600 3200MHz Unbuffered Memory
- Hard Drive: 4TB (4 x 1TB) SATA III 6Gb/s SSD for Ultra Fast Storage
- Hard drives installation required
How long a migration can take
Ars Technica reported Palmer’s distinction between partial migrations, which could take up to a year, and complete migrations, which could take at least three years. A January 2025 Gartner estimate relayed by Technologent put VMware VM migrations at 18–48 months and said scoping and alternative evaluation could require six or seven full-time employees for up to 10 months. These are attributed planning estimates, not universal schedules.
Duration depends on VM count, application coupling, storage architecture, replication method, downtime tolerance, compliance approvals, target platform, internal staffing, and the quality of rollback testing. A disk-conversion tool may finish in hours while dependency mapping, security validation, backup testing, and operational retraining take months.
Which workloads to move first
Good early candidates
- Development and test systems with documented owners and low recovery impact.
- Low-dependency internal applications and stateless services.
- Workloads with predictable operating-system and application licensing.
- Systems approaching a hardware or software refresh.
- Applications whose owners can support a representative pilot and rollback test.
Poor early candidates
- Latency-sensitive systems and tightly coupled database clusters.
- VMware-dependent appliances or applications without target-platform certification.
- Regulated workloads lacking approval for the proposed destination.
- Systems with unclear ownership, undocumented dependencies, or no tested rollback.
- Applications whose licensing becomes more expensive under a new host, core, or cloud model.
A decision framework for CIOs and infrastructure teams
- Define the objective. Decide whether the goal is lower cost, vendor diversification, a data-center exit, cloud adoption, modernization, or reduced operational complexity.
- Build the baseline. Record VMware subscription and bundle costs, hardware refresh needs, staffing, support, backup, disaster recovery, power, and facility costs.
- Classify every workload. Assign each system to retain, rehost, convert, replatform, refactor, retire, or replace.
- Model three- and five-year total cost. Include migration tooling and consulting, training and hiring, cloud compute, storage, network, egress, backup retention, DR duplication, support contracts, and exit costs.
- Test representative systems. Validate performance, failover, backup and restore, monitoring, security segmentation, identity, licensing, and business workflows—not merely whether a VM boots.
- Design production and recovery together. A production move is incomplete if replication, immutable backup, or disaster recovery still depends on an unsupported VMware environment.
- Pilot before a renewal deadline. Start inventory and low-risk testing at least one renewal cycle early so procurement dates do not force untested cutovers.
- Score strategic risk. Compare portability, exportability, contract flexibility, vendor concentration, roadmap credibility, and future exit costs alongside feature checklists.
Common failure modes
| Failure mode | Safeguard |
|---|---|
| Counting VMs but not application dependencies | Create application-to-VM, network, storage, identity, and backup dependency maps. |
| Treating conversion as validation | Run performance, failover, backup, monitoring, and end-to-end business tests. |
| Omitting cloud egress and retained backups | Use realistic traffic, storage growth, retention, and recovery assumptions in a three- and five-year model. |
| Moving critical or tightly coupled systems first | Begin with low-risk, representative workloads and publish rollback criteria. |
| Losing security posture | Recreate and verify segmentation, least privilege, logging, secrets management, endpoint protection, and recovery controls. |
| Creating a new lock-in | Require exportability, portability, contractual exit terms, and documented migration paths from every finalist. |
How to interpret the market evidence
Customer dissatisfaction, a survey, and rising vendor revenue are not interchangeable measures. The CloudBolt results show that surveyed organizations are reducing VMware footprints, while the Gartner statement is a forward-looking prediction. Broadcom’s reported infrastructure-software revenue growth—25% year over year to $6.6 billion in the cited September 2025 earnings coverage—does not disprove workload migration; subscription pricing and customer mix can increase revenue even as some workloads leave. That historical figure should be checked against the original Broadcom earnings release before being used for a current financial comparison.
The practical question is not whether exactly 35% will move. It is which workloads face the greatest commercial pressure, which have a technically credible destination, and whether migration risk is lower than renewal risk.
Useful planning and protection resources
Organizations evaluating a move can review Azure Migrate, Google Migration Center, AWS Application Migration Service, Veeam, and Zerto. Tool licensing and service costs vary by protected workload, duration, infrastructure, and project scope, so a current, workload-specific quote is necessary.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
