HN Debrief

Tesco moving 40k server workloads off VMware amid Broadcom's abusive conduct

  • Infrastructure
  • Enterprise Software
  • Open Source
  • Business Strategy

The story says Tesco, the UK's largest supermarket chain, is trying to move roughly 40,000 workloads off VMware while fighting Broadcom in court over what it calls abusive licensing and sales conduct after the VMware acquisition. Ars notes that Tesco has already picked a replacement platform but is running into compatibility problems with backup and disaster recovery tools including Veeam and Zerto. That detail drove most of the useful discussion. People did not treat the move itself as surprising. They treated it as another large enterprise finally deciding the renewal math is worse than migration pain.

If your infrastructure still depends heavily on VMware, treat this as a budgeting and execution problem now, not later. The lock-in is deepest in backups, storage, networking, and old workloads, so that is where migration planning needs to start before renewal pricing forces the timeline.

Discussion mood

Strongly negative toward Broadcom. The mood mixes anger at the pricing and sales tactics with a pragmatic acceptance that Broadcom may be winning financially because many large customers are too locked in to leave quickly.

Key insights

  1. 01

    Large VMware exits are now routine

    At enterprise scale, moving tens of thousands of guests is presented as an operational program, not a moonshot. One commenter working on Red Hat migrations said 40,000 servers would be large but still ordinary, with 500 to 1,000 guests per day possible once scoping is done and conversion tooling like virt-v2v and Migration Toolkit for Virt is in place. That reframes Tesco's plan from headline drama into a mature services market created by Broadcom's pricing shock.

    Do not assume size alone makes migration impossible. Benchmark your estate by migration throughput, storage constraints, and exception handling so you can estimate a real exit timeline instead of treating VMware renewal as unavoidable.

      Attribution:
    • rwmj #1
  2. 02

    The schedule is dominated by bureaucracy

    The long pole is not usually copying VMs. It is procurement, compliance, GDPR review, legal signoff, coordination between old and new platform teams, and user training around changed workflows like virtual desktop access. That explains why an 18 month plan can be rational even when the technical cutover per machine is measured in hours or minutes.

    If you are planning a platform exit, start with governance and stakeholder mapping. Security, legal, procurement, and training delays can sink the schedule long before the migration tools do.

      Attribution:
    • mjfisher #1
    • to11mtm #1
  3. 03

    Price shocks can justify migration risk

    One firsthand account said Broadcom's renewal quote made the decision easy because even outages during migration cost far less than staying put. The commenter described obscure driver issues and workload quirks after the move, but said the total damage was still well under the proposed renewal cost. That is the clearest practical signal in the discussion. Migration pain is now quantifiable and often cheaper than compliance with the new price sheet.

    Run the comparison explicitly. Include likely outage costs and remediation work in the model, because the answer may still favor moving faster than your organization is comfortable with.

      Attribution:
    • sokoloff #1
  4. 04

    The replacement platform is probably not the hard part

    People named several realistic targets, including OpenShift Virtualization built on KubeVirt, Nutanix, HPE Morpheus, and Proxmox. The argument that mattered was not which logo wins. It was that the blocker is the surrounding stack, especially backup and disaster recovery integration. Because Veeam and Zerto incompatibility showed up in the article, commenters used that to rule out some candidates and to highlight that hypervisor swaps become ecosystem projects fast.

    Map every VMware-adjacent dependency before choosing a target. Backup, disaster recovery, storage, monitoring, and network tooling will narrow your options more than hypervisor features will.

      Attribution:
    • Fordec #1
    • d3Xt3r #1
    • cloudie78 #1
    • p_l #1
    • Flere-Imsaho #1
    • nick__m #1
    • proxysna #1
    • digitalsin #1

Against the grain

  1. 01

    Many customers still cannot leave

    The clean migration narrative breaks down in environments where deployments, monitoring, backups, and operational runbooks are deeply VMware-specific. One commenter said plenty of enterprises and US government organizations are renewing at painful rates because they cannot unwind that coupling quickly, and predicted some workloads will still be on VMware years into exit programs. That pushes against the idea that Broadcom has overplayed its hand already.

    Audit for VMware-specific operational assumptions, not just guest compatibility. The more your tooling and process depend on VMware, the more likely you are to end up paying for a stranded footprint during a multi-year transition.

      Attribution:
    • stackskipton #1
  2. 02

    VMs are not disappearing anytime soon

    The claim that Broadcom is simply harvesting a dying category got challenged hard. Commenters pointed out that containers still usually run on servers that are themselves virtualized, and that a huge share of enterprise migrations involve old Windows systems and abandoned or vendor-locked software that will never be neatly containerized. The future may use different control planes, but the underlying need for virtual machines remains stubbornly durable.

    Do not build strategy around the assumption that containerization removes your virtualization problem. Plan for a long tail of VM-based workloads and choose platforms that can support that reality for years.

      Attribution:
    • twoodfin #1
    • rwmj #1
    • bigstrat2003 #1

In plain english

Broadcom
A large technology company that acquired VMware and then changed its pricing and product packaging.
GDPR
General Data Protection Regulation, the European Union privacy law that gives people rights over how organizations collect, use, and delete personal data.
HPE
Hewlett Packard Enterprise, a company that sells enterprise servers, storage, and infrastructure software.
hyperscalers
Very large cloud providers such as Amazon Web Services, Microsoft Azure, and Google Cloud.
KubeVirt
An open source project that lets Kubernetes manage virtual machines.
Migration Toolkit for Virt
A Red Hat toolkit that helps plan and execute migrations of virtual machines into OpenShift Virtualization.
Morpheus
An infrastructure and cloud management platform owned by HPE that can be used in VMware replacement strategies.
Nutanix
An enterprise infrastructure platform that combines virtualization, storage, and management software.
OpenShift Virtualization
A feature of OpenShift for running virtual machines alongside containers, largely based on KubeVirt.
Proxmox
An open source virtualization platform used to manage virtual machines and containers.
Veeam
A backup and recovery product commonly used to protect virtual machines and other enterprise systems.
virt-v2v
An open source tool for converting virtual machines from one virtualization platform to another.
virtual NIC
A virtual network interface card presented to a virtual machine for networking.
VMware
A widely used enterprise virtualization platform that lets companies run many virtual machines on shared physical servers.
workloads
Applications, services, or computing tasks running on servers or virtual machines.
Zerto
A disaster recovery and replication product used to keep copies of systems available after failures.

Reference links

Migration tooling and talks

Broadcom and VMware business impact

Open source database forks

  • WarehousePG
    Mentioned by a commenter promoting an open source fork of Greenplum for teams trying to move off Broadcom-related products.