Direct answer: Neither answer fits every organization. The right path depends on your renewal date, contract size, application portfolio, and team skills. We assess these factors before recommending a direction — and we have no commission on either outcome.

Stay on VMware when:

• Your environment relies on vSphere-specific features without equivalents (vSAN ESA, advanced NSX-T, Tanzu)
• You hold negotiation leverage (large contract, multi-year horizon, EMEA volume)
• Migration cost in time and risk exceeds 3 years of subscription savings
• Mission-critical applications have hardware-level licensing tied to VMware host identifiers

Migrate to XCP-ng when:

• You face renewal increases of 800% or more (common for European mid-market)
• You're paying for bundled products you don't deploy
• Your workloads use standard hypervisor features (HA, vMotion, snapshots, shared storage)
• You want predictable licensing and an exit strategy from vendor lock-in

We provide independent post-Broadcom analysis — including options to stay, options to migrate, and hybrid scenarios. See our complete VMware to XCP-ng migration approach.

Need an honest assessment? Book a free 45-minute expert consultation.

Direct answer: XCP-ng with Vates MSP support typically costs 50–70% less than equivalent VMware bundles post-Broadcom — for comparable production capabilities. Actual savings depend on your workload, support tier, and existing entitlements.

Note: License savings represent only part of the TCO equation. Migration cost (engineering, validation, training), team familiarity, and operational tooling investment must also be considered. We deliver complete TCO analysis including these factors during assessment.

Learn more about XCP-ng Enterprise Virtualization and how Vates MSP support works.

Want a TCO comparison for your environment? Schedule a consultation.

Direct answer: GPU virtualization with passthrough makes sense when you need workload isolation, multi-tenancy, or operational flexibility on shared infrastructure. Bare-metal is preferable for single-tenant maximum-performance training where every percentage of GPU utilization matters.

GPU virtualization (passthrough) fits when:

• Multiple teams or projects share the same GPU infrastructure
• You need to mix AI training, inference, and development workloads on one cluster
• You want unified backup, snapshot, and DR strategies across CPU and GPU workloads
• Hardware budget is constrained and you want to maximize GPU utilization across teams

Bare-metal AI fits when:

• Single dedicated training workload — large language model fine-tuning, foundation model training
• Maximum throughput and minimum latency are non-negotiable
• Multi-GPU NVLink or NVSwitch topologies are required for distributed training

XCP-ng supports full GPU passthrough for NVIDIA H100, A100, and L40S — with no vGPU licensing required. The hypervisor adds negligible overhead for passthrough scenarios. This is the same Xen technology trusted by AWS, Alibaba, and IBM for their cloud GPU instances.

Explore our complete GPU Virtualization for AI Workloads capabilities, or see our broader AI infrastructure solutions.

Designing AI-ready virtualization? Book an AI infrastructure consultation.

Direct answer: Both are mature, production-ready hypervisors built on different foundations — vSphere on VMware's proprietary stack, XCP-ng on the open-source Xen Project. Most enterprise virtualization features have direct equivalents on both platforms.

Where vSphere remains stronger: NSX-T networking depth, Tanzu Kubernetes integration, vSAN ESA performance for very specific workloads, and vCenter ecosystem maturity. Where XCP-ng wins: licensing predictability, no vendor lock-in, GPU passthrough cost, transparent open-source codebase.

Both VMware and XCP-ng are first-class platforms in our practice — we deliver Level 3 expertise on both. Storage compatibility matters too: see our enterprise storage solutions certified for both hypervisors.

Need a platform recommendation for your workloads? Talk to our virtualization team.

Direct answer: We approach unresolved incidents with a structured Level 3 methodology — Diagnose, Isolate, Resolve, Document — applied to performance problems where standard support has run out of escalation paths. Most cases are closed within 24–48 hours with a written root-cause report.

Our approach to unresolved performance issues:

• Workload profiling — actual usage patterns, not theoretical specs
• NUMA topology analysis — vCPU placement against physical socket boundaries
• Storage I/O path investigation — latency at every layer (VM → datastore → array)
• Scheduler contention — CPU ready time, co-stop, ESXi/Xen scheduler logs
• Network stack forensics — physical NIC, virtual switch, MTU, TCP tuning
• Memory pressure analysis — ballooning, swapping, transparent page sharing impact

Real-world example we handle: a cluster underperforming despite correct configuration — no obvious bottleneck, monitoring shows nothing. Root cause turns out to be NUMA imbalance combined with storage contention. Result: 40% performance improvement without hardware changes.

This Level 3 expertise extends beyond VMware. We troubleshoot Exchange, Windows Server, and storage incidents with the same methodology. Explore our complete Level 3 support capabilities.

Active performance incident? Contact our support team directly — no prior contract required.

Direct answer: Yes. XCP-ng is built on Xen — the same hypervisor technology AWS, Alibaba Cloud, and IBM Cloud use to run mission-critical workloads at hyperscale. With Vates MSP support, it delivers enterprise-grade production capability with contractual SLA, not best-effort community support.

Enterprise capabilities supported in XCP-ng:

• High Availability clustering with automatic VM restart on host failure
• Live migration (XenMotion) without service interruption
• Shared storage on iSCSI, NFS, Fibre Channel, and XOSTOR distributed storage
• Resource pools, dynamic memory control, CPU pinning
• Backup and replication via Xen Orchestra and Veeam-compatible workflows
• GPU passthrough for AI workloads (no vGPU licensing required)
• Multi-site disaster recovery with continuous replication

Our production track record across Belgium, France, and the EU includes hundreds of VMs migrated and operated under SLA. We are the official Vates MSP for XCP-ng (Level 1 & Level 2 support) — meaning direct escalation to Vates engineering when needed.

For an enterprise production reference: see our HE2B case study — production virtualization platform deployed for Haute École Bruxelles-Brabant, supporting their education infrastructure.

Storage architecture matters as much as the hypervisor. Our certified enterprise storage solutions integrate seamlessly with XCP-ng for production deployments.

Evaluating XCP-ng for production? See our XCP-ng Enterprise Virtualization page or book a consultation.

Direct answer: Level 3 means resolving incidents that have no documented solution and require deep system-level expertise — beyond what standard MSPs and vendor support deliver. The accountability is tied to resolution, not ticket closure.

Real Level 3 cases we handle:

• VMware support tickets stuck on third escalation with no resolution timeline
• vSAN storage latency collapse where standard L2 cannot identify root cause
• Failed cutover at 2 AM during migration — root cause unknown, next window weeks away
• VM performance issues impossible to reproduce, telemetry inconclusive
• Security incidents on production environments outside contracted SLA windows

We are also the strategic Level 3 partner for Prodata Systems' enterprise virtualization environments — meaning continuous Level 3 escalation under real production SLA constraints.

Most critical cases are resolved within 24–48 hours, even when vendor escalations remained open for weeks — with a written root-cause report.

Need Level 3 expertise on an unresolved case? Contact us directly.

Direct answer: We operate continuous monitoring with defined incident response, SLA built around your business hours (extended or round-the-clock), and proactive alert triage. The single point of contact remains the same regardless of platform — VMware, XCP-ng, Exchange, or Windows Server.

Capabilities for 24/7 production:

• Continuous monitoring with defined incident response procedures
• SLA built together — business hours, extended hours, or 24/7 coverage
• Proactive alert triage and automatic expert escalation
• Emergency response via phone, email, and WhatsApp
• Quarterly infrastructure health reviews
• Engineer assigned directly — no queue, no ticket number

We have intervened overnight on production environments impacting revenue, operations, and customer access — including outside contracted SLA windows. Documented example: full operational recovery for The Source Belgium before opening, on one of their highest-revenue days of the year. The Source had an active contract with us, but the incident occurred at night, outside their contracted SLA coverage — we took ownership anyway.

This 24/7 capability is integrated with our IT infrastructure and storage support — single point of contact, full stack accountability.

Need round-the-clock virtualization support? Explore support options.

Direct answer: Most consultants operate at Level 1 and Level 2 — known issues, documented solutions. Virtualtek operates at Level 3: we resolve incidents that vendor support and standard MSPs have not been able to close, with direct accountability and a written root-cause report.

Our differentiation in practice:

• Official Vates MSP for XCP-ng — Level 1 & Level 2 support
• Strategic L3 partner for Prodata Systems' enterprise virtualization environments
• Official Gigabyte AI and NVIDIA partner — single point of contact for the full infrastructure stack
• Independent of Broadcom — no commission on VMware renewals
• Engineer-led, not commercial-led — we tell you when migration doesn't make sense

This positioning makes us a cross-functional partner for the full IT infrastructure stack, including enterprise storage and AI services — not a single-product reseller.

Want to discuss how this applies to your environment? Book a free expert consultation.

Direct answer: Yes. We have intervened overnight on production environments impacting revenue, operations, and customer access — including outside contracted SLA windows. The Source Belgium is a documented example: full operational recovery before opening, on one of their highest-revenue days of the year. The Source had an active contract with us, but the incident occurred at night, outside their contracted SLA coverage — we took ownership anyway.

This capability is rooted in three pillars: deep platform expertise (VMware, XCP-ng, storage, Exchange, Windows Server), direct engineer engagement (no queue, no ticket number), and willingness to take ownership outside conventional SLA boundaries when stakes justify it.

For ongoing mission-critical environments, we structure contracts with extended or 24/7 coverage, proactive monitoring, and quarterly health reviews. See our complete support capabilities.

Critical incident in progress? Contact our support team directly — no prior contract required.

Direct answer: Two paths depending on the situation. For active incidents, contact us directly — no prior contract required, no onboarding delay. For strategic assessments, book a free 45-minute expert consultation. We assess your situation and tell you honestly what we can do before any commitment is made.

We serve organizations across Belgium, France, and the European Union. For a complete view of related services, explore our migration services, VMware expertise, XCP-ng enterprise, or GPU virtualization.

Ready to talk? Book your free expert consultation.