Tech Alternatives Nantes: from dependence to autonomy
On 18 September, at Tech Alternatives Nantes, I took part in the panel "From dependence to autonomy". Here is the case I made for the layer people look at least: infrastructure.
On 18 September 2026, La Cantine x La French Tech Nantes and Clever Cloud held Tech Alternatives Nantes at Halle 6 Est, as part of Nantes Digital Week. The day was devoted to alternatives to the large technology platforms, with one simple rule: no sales pitches, only field experience, panels and concrete cases.
I took part in the 11:50 panel, "From dependence to autonomy: how to break free from the tech giants?", alongside Matthieu Dalvy, co-founder of Déclic.ai, Charles Citroën, head of public affairs at Forum InCyber, and Célia Viejo, CMO of AlphaEdge. Four backgrounds and four angles, from artificial intelligence to cybersecurity, from public policy to marketing, around a single question: how does an organisation take back control of the technologies it depends on?
My angle was the layer people look at least: infrastructure. Here is the case I made.
Dependence starts below the applications
When people talk about dependence on big tech, they first think of the US public cloud, office suites and AI models. For many IT departments, though, the most tangible dependence sits lower down, in the foundation that runs their servers: the hypervisor.
Since Broadcom completed its acquisition of VMware in November 2023, the rules have changed: perpetual licences are gone, subscriptions are the norm, products come in bundles, and renewal bills have often risen sharply. In late August 2026, the public download of VDDK, a VMware library that many backup and migration tools rely on, disappeared.
That last episode sums up the problem. When a vendor controls both the product and the exits, dependence is no longer just commercial: it becomes structural.
You cannot leave a dependence you do not measure
Faced with a licence increase, the first reflex is often to look for an alternative. The right reflex is to know what you have first:
- how many virtual machines, on which operating systems, some of them out of support;
- which applications depend on which others;
- how much of the estate is actually used, and how much is oversized;
- what is backed up, and where.
Many teams only find these answers during the renewal negotiation, which is too late to carry any weight. Autonomy starts with a reliable, up-to-date inventory. That inventory is what lets you price an exit scenario, compare platforms on figures rather than convictions, and set your own pace.
Do not trade one dependence for another
Leaving VMware only to lock yourself into another platform solves nothing. Serious alternatives exist and they are mature: Proxmox VE, XCP-ng and others, often open source, often European. But the real gain in autonomy comes from reversibility: running several hypervisors side by side, migrating in waves rather than all at once, and being able to roll back if a step goes wrong.
This is the principle we apply at NaoShift: a single console above all your hypervisors, so that the choice of platform remains a technical, revisable decision rather than a lifelong commitment. An operations team keeps the same view, the same risk analyses and the same reports, whatever hypervisor runs its machines.
Data is a dependence too
Taking back control of your infrastructure also means controlling what leaves it. An analysis tool that stores the names of your servers, machines and addresses with a third party, or sends them as they are to an AI model, creates a quieter but very real dependence.
At NaoShift, two principles apply. Anonymisation happens at collection time: what is stored, and what an AI model receives, does not carry the names of your infrastructure. And collection goes through an outbound agent, read-only by default: no inbound flow to open, nothing to install on the hypervisors or inside the virtual machines.
Autonomy is also built locally
Tech Alternatives Nantes showed it: in Nantes, alternatives are not promises. Hosting providers, software vendors, integrators and professional networks already work together. That is why NaoShift joined French Tech Nantes this summer.
Choosing a supplier close to home is not folklore. It means being able to talk to them, weigh on their roadmap and know where your data lives.
What I take away
- The most expensive dependence is often the least visible one: infrastructure.
- Autonomy starts with measurement: an up-to-date inventory before any decision.
- Reversibility matters more than the choice of platform: several hypervisors, one console, migration in waves.
Thanks to La Cantine, La French Tech Nantes and Clever Cloud for the organisation, and to the other panellists for the quality of the discussion.
Frequently asked questions
What is Tech Alternatives Nantes?
A one-day event held on 18 September 2026 at La Cantine, Halle 6 Est, as part of Nantes Digital Week, organised by La Cantine x La French Tech Nantes and Clever Cloud. It brings together companies, associations and professional networks around alternatives to the large technology platforms, through a keynote, panels and field experience.
How can you reduce your dependence on VMware without migrating everything at once?
Start with a priced inventory of your estate, then run VMware alongside one or more alternative platforms under the same console. Machines then move in waves, least critical first, with a rollback possible at each step.
Where should you start when assessing a move away from VMware?
With measurement. The VMware exit file prices a scenario on your real estate, and the hypervisor comparator helps you weigh the platforms against each other.
Going further
Preparing to leave VMware, or simply want a clear view of your infrastructure? You can start for free or write to us to talk it through.