Today, I'm sharing the tragic but lesson-filled story of Dindon. Dindon wasn't just any turkey; he was a small-scale poultry producer, passionate about his farm, but also a huge tech enthusiast.

Some years ago, Dindon told himself: "My farm must enter the modern era!" He dove headfirst into Agile and Scrum certifications. His grain shed became a "Backlog," every hen had a "User Story" for laying eggs, and he "iterated" on egg sizes.

For his inventory and online sales system, Dindon decided to go big. No more physical servers in the barn. He outsourced everything to the Public Cloud (Ireland region). He mastered IaC with Terraform. His infra was "abstracted," "scalable," and "serverless." Dindon felt powerful. He wasn't just a turkey anymore; he was a "Cloud Architect."

Then, Dindon received his first major Cloud bill during the peak Christmas season. €475 per VM, for 200 "various" VMs that mostly just stored turkey photos and processed a few orders. Meanwhile, his old On-premise setup used to cost him only €117 per unit — corn and electricity included. His turkey brain glitched. "What's the logic here?!" he gobbled across the farm.

If you see Dindon today, he's not "agile" anymore. He's plucked bare. The data egress fees, the complexity of his IaC, and the "scalability" of his Kubernetes cluster have devoured all his profits. He spends his days doing Extreme #FinOps:

Do you remember Dindon? Our friend who got plucked bare by the Public Cloud and those €95k/month bills? 💸 After a few nights crying over his last remaining feathers, Dindon had an epiphany. He stopped letting himself be blinded by buzzwords and decided to get rational. He realized a small poultry farm didn't need a "multi-region serverless" infrastructure in Ireland — that's like putting a Boeing jet engine on a hay cart!

Dindon launched his "Cloud Exit" strategy. His goal: a reliable, predictable, and human-sized infrastructure.

Instead of talking to algorithms or support teams on the other side of the world, Dindon turned to local and regional providers (Ecritel, OVHcloud, Scaleway, DRI, Claranet France, Infomaniak — The Ethical Cloud, IONOS...). The shock: a clear quote. No hidden fees, no probability calculus for the end-of-month bill. One server has one fixed price.

Dindon realized that 80% of his 200 "as-a-code" VMs were either useless or massively underutilized.

To truly understand why our friend Dindon ended up plucked bare (Episode I) before escaping to local providers (Episode II), we have to go back to where the drama all began.

In the beginning, Dindon was a happy digital farmer. He managed his servers on Kimsufi boxes using good old Cpanel or Plesk. It was simple; it was robust. He'd click a button, and boom — his site just worked. It was the glorious era of "if it ain't broke, don't fix it."

But one day, while scrolling through "Tech Evangelist" blogs on LinkedIn, Dindon felt a sting of shame. He told himself:

He started seeing his dedicated servers as "dirty" and "static." He craved ephemerality, abstraction, and prestige. He wanted Infrastructure as Code because that's what Silicon Valley unicorns do, right?

He finally felt "In." He felt "Scalable." He didn't realize yet that he had just traded a solution that cost the price of a bag of grain for a giant slot machine that was about to swallow every single one of his feathers.

Remember how our friend Dindon ended up completely plucked by the complexity and costs of the Public Cloud (Episodes I & III), before deciding to make a run for it (Episode II)? Today, discover how that escape turned into a smooth and successful migration, thanks to a true partner: the local provider Plannitel.

After his FinOps realization, Dindon met Plannitel. No more abstract buzzwords — just pure pragmatism. Plannitel didn't show up with a catalog of 400 complex APIs; they showed up with a methodology:

When D-Day arrived, Plannitel operated. It was precise, calm, and professional. They handled it from A to Z: secure data transfer from Ireland, configuration of dedicated Bare Metal servers and private instances, post-migration testing. Dindon no longer had to code in YAML just to know if his site was up. He had a human point of contact.

The go-live happened with absolute discretion. No panic, no surprise "404 Not Found" during the peak season. The site worked — plain and simple. Better yet, it was faster. Thanks to Plannitel, Dindon has his feathers back. He has a fixed, rational, and surprise-free cost, technical support that actually picks up the phone, and peace of mind to focus on his chemistry lab.

After the heroic rescue by Plannitel (Episode IV), many of you are asking: "What happens next? Is the honeymoon phase still going strong?" 🌙 The answer is yes. Dindon is no longer "Cloud prey"; he has become the king of his own digital farmyard.

Freed from YAML hell and complex API debugging, Dindon has finally been able to devote himself to his true passion: his chemistry lab applied to agro-excellence. His turkeys are the most renowned in the region, and his online shop processes orders without breaking a sweat. He no longer chases "auto-scaling" that doesn't scale; he chases his own success.

Did you think Dindon was just going to sit back, count his savings (60% ROI, remember Episode V), and eat barbecues at Plannitel? Then you don't know the bird. 🍗

Once his feathers were fully grown back, Dindon looked around. What he saw saddened him: thousands of other businesses, SMEs, and mid-caps were still being silently "plucked" by hyperscalers. Dindon had an epiphany. His success shouldn't be an exception — it should be a blueprint. He transformed from a prosperous producer into a Local Cloud Activist.

Dindon called every "plucked" peer he knew. And so, the association "The Cloud Fall Guys" (Les Dindons de la Farce) was born.

The goal? To unite all the companies that realized, a little too late, that they were the punchline of a monumental joke orchestrated by the Public Cloud: unreadable and unpredictable bills, prohibitive Egress fees, and total loss of control. The association has become a powerful lobby for Digital Sovereignty, advocating for a return to local, human, and transparent partners.

But talk is cheap. It's time for action. Dindon and his association have announced the launch of the largest operation in European digital history: Operation "Dindon!". The concept: the greatest data repatriation of the century. The objective: to organize, plan, and provide technical and financial support for the massive repatriation of data from thousands of European companies — moving it out of the Public Cloud and into Sovereign and local Clouds.

Why the name "Dindon!" with that angry exclamation point? It is a vibrant tribute to a courageous Director at STIME (the IT arm of the Mousquetaires Group) who, back in 2011, faced with the exorbitant costs of Windows Server and Office licenses, decided to name his major infrastructure migration project: "Operation Dindon." The circle is complete.

Remember Dindon? Our tech farmer who had fled the Public Cloud at €475/VM to find technical and financial sanity (sovereign Bare Metal, real ROI). Well... Dindon relapsed. 😔

He had kept one last addiction hidden in a corner of his console: his "Managed" Kubernetes cluster. He had convinced himself: "It's managed, it's simple, it's just the modern compute layer, it won't bankrupt me." Bullshit. 🛑 "Managed" doesn't mean "Free" — it means "Opaque."

Bottom line: 104% in hidden extra costs.

The worst part? "Managed" is just an Ops illusion. They manage the control plane, but YOU are still the one wrestling with complex drop-down menus for NAT, LBs, and volumes at 3:00 AM.

Render unto Caesar what is Caesar's: hyperscalers invented brilliant abstraction. But they also invented financial vassalage. True #FinOps means reading the bill line by line, not blindly trusting the abstraction that suffocates you.

Thought Dindon hit rock bottom with Kube? 🛑 Today, we tackle the ninja of financial asphyxiation: The AWS NAT Gateway.

For "High Availability," the AWS manual advises 1 NAT per zone (AZ). Dindon has 3 AZs, so he clicked 3 times. 🖱️ What he didn't know? At the hyperscaler's, doing NAT (a simple iptables rule, free on Linux since 1998) isn't a feature — it's a luxury tollbooth.

For 10 TB/month of outbound data (updates, APIs, webhooks...):

€565/month. €6,700/year. Just to change a TCP source IP! ☕💥

On-Premise, a €150 Mikrotik router (one-time CAPEX) or a pfSense VM (€0) handles this without breaking a sweat. In the Cloud, the network is a taxi meter spinning with every byte breathed.

The giants' "High Availability" is often designed to multiply your paid resources. True #FinOps? Replace those 500-buck managed NATs with a simple EC2 instance (NAT instance) or a dedicated local server for the price of a subway ticket. Or better yet: #CloudExit.

Today, we're dotting the "i"s and putting instances on real hardware. We're going to talk about a touchy subject, but we'll do it with respect: bankruptcy by the Cloud and the great sovereign slumber.

Dindon has a confession to make. If he has become an ayatollah of FinOps and Bare-Metal, it's not out of ideology. It's because he has seen ships sink. I have seen with my own eyes how infrastructure can go from being an "innovation engine" to a "financial ball and chain."

Let's be fair: technically, what AWS, GCP, or Azure do is masterful. Their engineers are geniuses. But their economic model is a cash vacuum designed for unicorns doped on fundraising. The tragedy is that SMBs and traditional companies are copying these architectures. The result? Public Cloud becomes the leading cause of financial stress for CFOs. We no longer go bankrupt because we lack clients; we go bankrupt because our variable costs (Egress, NAT, API calls) grow faster than our revenue.

And while the real economy gets siphoned by untraceable OPEX bills, what are our champions of "Digital Sovereignty" doing? They are sleeping. 😴 Our politicians make magnificent speeches with the phrase "Sovereign Cloud" in size 72 font. But on the ground? No massive tax incentives to relocate data, no real education on Total Cost of Ownership (TCO).

And our local hosters? You have the steel. You have the datacenters. You have the prices (sometimes 10× cheaper in CAPEX). But where is your voice?

Public Cloud is not a scam; it's a luxury product. If you need instant global elasticity, pay the premium. But for 80% of workloads (classic databases, stable backends, web), staying on a hyperscaler is like using a private jet to go buy a baguette. True "Scale" is profitability.

Today, a change of program. Before judging local hosters or the State, Dindon has decided to audit the root of the evil: the company itself. Because the sinking of the Cloud is never a technical accident; it is always an organizational suicide.

Welcome to the autopsy of an internal disaster, a clinical pattern that our industry repeats in loops.

Imagine a management committee organizing a major "workshop" to "rationalize the infrastructure." On the table, the numbers are lunar: hundreds of thousands of euros burned each month with hyperscalers, coupled with six-figure bills for "modern" SaaS licenses.

Yet, in the shadow, this same company possesses its own treasure: physical racks in local Datacenters, half-empty, perfectly equipped, and already amortized, costing ten to twenty times less to run.

Faced with this absurdity, the field Ops proposes the obvious: let's use the empty space of our Bare-Metal to repatriate our Kubernetes clusters and divide the production bill. The reaction from the top? Scorn. The CEO, blinded by the ego of his own "vision," brushes the proposal aside with a wave of his hand: "We'll see later, let's stay on the roadmap." The CTO, for his part, is silent — he is the archetype of the "Technical Director" who clings to derisory micro-tasks to justify his place, but who validates without batting an eye the waste of millions in OPEX.

The field engineer does not give up. He produces a quantified study, multiple exit scenarios, quotes in support. The response? Management plays for time. For a whole week, his requests for a simple 15-minute slot are dodged. To force the door of his own management committee, the Ops literally has to slam his fist on the table and demand, by a frontal email in front of other employees, an emergency meeting.

The meeting is finally wrested, but the sanction falls: a recasting and an oral warning from management. Laying numbers on the table is no longer seen as financial optimization; it is perceived as insubordination. The accounting truth has become the sworn enemy of managerial pride.

When the top of the pyramid (CEO, CTO, CFO) chooses blindness by ego, what does the rest of the company do in the face of the financial wall?

Faced with denial, our field Ops decides to break the law of silence. He takes the debate to the internal public square: emails, meetings, face-to-face. He warns that the ship is heading straight into the iceberg. The objective is not to make a publicity stunt for his ego, but to save the profitability of the company and, by extension, the jobs of his colleagues.

At first, colleagues encourage behind the scenes. But from the first management recasting email, it's a stampede. All departments withdraw from the debate. Support is whispered in secret, at the coffee machine. Faced with the risk of confronting management, the fear of losing their "cushy job" prevails over the instinct of collective survival. The whistleblower is left to isolate himself alone on the battlefield.

The engineer then makes a choice: to take the oath to remain the "veracious witness" of this intellectual bankruptcy, even if it must cost him his place. Before the ego of management and the complaisant silence of employees, he is inevitably pushed to the exit. The inconvenient witness is eliminated, Cloud waste continues, and colleagues keep their "cushy job"... waiting for the next social plan.

The testimony of Operation Dindon now turns toward our local players. This is not about blaming the giants for their excellence, but about taking a sincere look at our own failings. Sovereignty cannot be a mere slogan; it must be a promise of competence.

The market follows efficiency. The call for a common front — ironically named "Operation Dindon" to trigger a wake-up call — was an invitation to protect our SMEs. Ignoring this extended hand is not just a management error; it is a failure in the duty of economic fraternity. To remain silent while our companies bleed financially is to become complicit.

The developer is the one who prescribes. Offering them dated portals and obscure documentation is an injustice. Hyperscalers respect their time with fluid APIs. Technical justice starts here: if we want a return to local iron, let's make the path passable instead of a calvary.

The market wants managed services, not just raw metal. Leaving our companies with no alternative face-to-face with proprietary lock-in is a failure of counsel. Similarly, brandishing labels to mask a technical lag lacks sincerity. Patriotism must not be a shield for incompetence, but a driver of excellence.

Our sovereignty sleeps in our datacenters, locked by a lack of audacity. No one will save the local sector in its place. The courage to unite and the sincerity to recognize its lags are the only keys.

The Operation Dindon audit rises today to the highest levels of the State. After analyzing the responsibilities of the local sector, the demand for justice compels us to examine public action. This is not about building a biased prosecution, but about placing the mirror of reality in front of stated ambitions.

Strategic autonomy is a noble ideal, fiercely defended in official speeches. However, the physics of Datacenters does not feed on eloquence. Declaring sovereignty while neglecting massive investment in hardware infrastructure creates a gaping dissonance. The truth of the Iron demands concrete actions, at the risk of seeing ourselves reduced to the status of digital tenants.

Affixing a local seal (SecNumCloud) on a technology whose engineering and kill switch are located across the Atlantic is a legal construct, not technological independence. Sincere counsel forces the question: how can we demand our SMEs repatriate their data to sovereign infrastructures when the State itself entrusts its crown jewels (like the Health Data Hub) to hyperscalers? Authority is built on example, not on exemption.

The State legitimately waters our startups through generous subsidies (BPI, French Tech). Yet, in the absence of FinOps guardrails, these millions of euros too often end up absorbing exponential Cloud bills, de facto financing foreign R&D. Where are the tax incentives to encourage the return to CAPEX and investment in local Bare-Metal? The silence on this economic lever is deafening.

Sovereignty cannot be bought with rhetoric; it is forged in the steel of our own infrastructures. Legislating against data lock-in (Egress fees) and offering tax incentives for companies to return to Bare-Metal are the true levers of an assumed and lasting independence.

This manifesto is drawing to a close. The French diagnosis (Episode XII) was merely the symptom of a broader failing architecture: the strategic flaw of the European Union.

While Europe discourses on compliance and the EUCS label, a financial asymmetry takes hold. The European legislator managed to regulate telecom "Roaming" down to the cent, yet looks away when data is shackled by prohibitive egress fees. The Data Act remains an abstraction in the face of the reality of irrevocable 36-month "commits." Europe thus lets its SMEs be bound by general terms and conditions of sale that neutralize all technical freedom.

Recovery plans and European funds (Horizon Europe) legitimately irrigate our technological fabric. Yet, the financial mechanics often end up draining this capital towards the San Francisco Bay. By refusing to impose local infrastructure quotas to obtain these subsidies, Europe indirectly finances its own technological trade deficit.

Persisting in this legislative inertia, after being alerted to the reality of TCO and proprietary lock-in, changes the nature of the failure. This is no longer administrative sluggishness; it is accepting the dependence of our engineering by omission.

This concludes the Operation Dindon audit. To the media, lobbyists, or podcasters: the media spectacle and sterile debates are politely declined. The withdrawal from this public circus is definitive. However, the posture of the "truthful witness" remains. The technical fight will continue where it began: in the engineering trenches, at the heart of the Bare-Metal, facing the Radix.

The entirety of the "Operation Dindon" manifesto (Episodes I to XIII) is released into the public domain. These texts may be reused, shared, or adapted by anyone, without authorization. Technical truth has no owner, and the urgency is now collective.

The Operation Dindon diagnosis is coming to an end. After analyzing cost structures, strategic dependencies, and political levers, the time has come for a technical assessment and the closing of this chapter. This final episode is intended as a solemn synthesis, driven by the demand for truth and sustainability.

Infrastructure is not just an accounting concept or a software abstraction; it is the physical foundation of our economies. The current trend towards total outsourcing has revealed its limits: loss of control over operational costs, increased complexity of transfers, and an erosion of field skills. The sustainability of a company rests on its ability to accurately arbitrate between service elasticity and asset solidity.

Digital sovereignty cannot be a mere slogan. It is built on precision engineering, valuing local hardware and the optimization of existing resources. The philosophy of returning to the "Iron" (Bare-Metal) and to technological recycling is not a step backward, but a step forward towards sounder and more responsible management. Strategic autonomy begins where dependence on infrastructures whose variables we no longer control ends.

Sovereignty is a choice. If I was harsh with companies, it is because their responsibility is decisive: neither the State nor Europe will act for them. As an employee myself, I know that the survival of our tools depends on our own audacity. It is now up to organizations to transform these findings into lasting decisions. Operation Dindon ends here, giving way to silent and rigorous action at the heart of the systems.