European design expertise instead of technological dependence – The French cloud model as an economic strategy

No more Microsoft dependency: What Germany now needs to learn from the French IT strategy

A persistent dogma in European digital policy is that a modern, efficient state cannot function without the dominant infrastructures of Amazon, Microsoft, and Google. The common assumption is that anyone who mentions cloud computing must be referring to hyperscalers – a notion that often leads to billions of euros in dependencies in German government offices. However, a look across the Rhine reveals that this narrative is not only politically short-sighted but also economically flawed.

While Germany is increasingly entrusting its administrative digitalization to US corporations, thereby entering into a profound vendor lock-in, France is demonstrating the opposite with its platform apps.education.fr. What began as a response to the pandemic and underwent a strategic realignment following the catastrophic fires at OVH data centers is now a European flagship project: a state-run, open-source cloud infrastructure that serves hundreds of thousands of users daily, manages millions of data records, and remains entirely under democratic control.

The following analysis reveals why the supposed cost advantage of hyperscalers is, upon closer inspection, pulverized by hidden fees (such as egress costs) and strategic dependencies. It shows how France, instead of paying licensing fees, is investing in building its own technological expertise and why digital sovereignty in 2025 is no longer a romantic utopia but a stark economic necessity. It is a comparison of two schools of thought: the German model of convenient but risky leasing of third-party technology, and the French path of arduous but rewarding technological emancipation.

When reality disproves established myths: An analysis of the paradigm shift in government digitalization

The history of modern cloud infrastructure is often told as a story of inescapable dependency. In this narrative, American hyperscalers play the role of technological saviors, for whom there is no alternative. Anyone who wants to operate their own infrastructure is portrayed as less efficient, less secure, and fundamentally less competitive. This narrative has become so entrenched in Germany that it encounters hardly any resistance – it is treated as an economic reality, not a political decision. However, the French implementation of Apps.education.fr demonstrates something fundamentally different: that digital sovereignty is not a technological utopia, but a question of institutional freedom of choice.

With this platform, the French Ministry of Education has chosen a technically ambitious path that is symptomatic in several respects of a different European understanding of digital infrastructure. The platform now provides over 337,000 users with independently operated cloud storage solutions, regardless of its technical complexity. The project's history is informative: founded in 2018, it was rolled out to tens of thousands of users within seven days during the COVID-19 pandemic, but subsequently faced a critical test. The fire at the OVH data centers in Strasbourg in 2021 marked a decisive turning point. As a European cloud provider, OVH had long been the guarantor of an alternative to the American hyperscalers. When the flames destroyed the SBG2 data center, approximately 18 percent of the IP addresses served by OVH lost their availability. A total of 3.6 million websites went offline, including critical French government websites. For France, this was not only a technical disaster, but also a strategic lesson: dependence on a single European provider is just as questionable as dependence on hyperscalers.

The French response was precise and long-term oriented. The Ministry of Education drew the logical conclusion and moved operations entirely to its own, directly controlled data centers. Apps.education.fr became a test institution for a new cloud philosophy. This decision cannot be dismissed as a romantic commitment to open-source ideals, but must be understood as a rational economic strategy: The state consciously chose to prioritize expertise and control over price.

The invisible constellation: Why European price comparisons are systematically distorted

To appreciate the economic dimension of this decision, one must first understand how the hyperscalers' cost model works. The major cloud providers – AWS, Microsoft Azure, and Google Cloud – calculate their costs based on a service model billed via Operational Expenditure (OPEX). Customers pay for every gigabyte of computing power used, for every outgoing data transfer, for transactions, and for a multitude of additional services. This granularity in cost calculation has proven effective, but it is also structurally asymmetrical: it rewards providers for intensive use and penalizes churn.

The reality of egress costs is revealing. AWS charges $0.09 per gigabyte for outbound data transfer, Google Cloud $0.05, and Microsoft Azure also $0.05. These fees act as a financial barrier, making platform migration more expensive. For large workloads, data transfer alone can quickly run into the millions—a mechanism that achieves customer retention through cost structures, not technological superiority.

The CloudStack study on Total Cost of Ownership provides concrete figures for Frankfurt as a reference location. For an identical compute workload over three years, the study calculates the following expenses: AWS costs $8.1 million, Microsoft Azure $9 million, and Google Cloud Platform $10.2 million. The same architecture based on on-premises CloudStack with colocation in the Frankfurt region costs $4.6 million, including hardware, data center operations, and personnel costs. This is just under 46 percent of the AWS price, calculated over all three years.

Such figures are not marginal. They also explain why European governments are suddenly reconsidering controlling infrastructure. The funding advantage is real, especially when workloads are stable, predictable, and continuous—precisely the profile of a public administration system. Hyperscalers have traditionally argued that their scalability, service quality, and technological edge justify this cost difference. But that argument weakens when alternatives have proven their worth.

The French platform had 100 million files and 330,000 daily active users by the end of 2025. This is not a pilot project, not an academic exercise. This is a complete administrative service with millions of daily interactions, powered by open-source software, with complete control over its own infrastructure.

Architectural complexity as myth: Federal systems and identity management on a large scale

A common argument against government-run cloud infrastructure is that the technical complexity is too high. Who has the expertise to manage millions of identities, federate heterogeneous systems, and simultaneously maintain security standards? This argument deserves serious consideration—it's not an unreasonable objection. But empirical reality shows that it's an exaggeration.

Apps.education.fr operates with 1.2 million identities in its target architecture. This is a large-scale, true identity management system. The platform is federated, meaning there are multiple instances that need to work together—Nextcloud clusters in several locations, integrated with legacy systems like Tchap (a French chat solution) and Zimbra (an email system). This integration is not trivial. But it has been resolved. The ministry processed 150 tickets with the Nextcloud developer over 18 months—a pace that is more pragmatic than perfectionistic. It's the mentality of a government agency that anticipates setbacks and addresses them systematically, rather than aiming for perfection from the outset.

The scalability of open-source solutions is often questioned. Linux, Kubernetes, Docker, and PostgreSQL are criticized as if they were hobbyist projects. This is historically inaccurate. The core of these software ecosystems is now part of the critical infrastructure of thousands of organizations. LinkedIn runs on Linux kernels, Netflix operates millions of containers with Kubernetes, and European banks rely on databases like PostgreSQL. The fact that this software is open-source doesn't make it smaller or less powerful—it simply means that the code is available for inspection and that no one is dependent on the goodwill of an American company to obtain security patches.

The architecture chosen by the French Ministry of Education for Apps.education.fr is deliberately conservative in its ambition. It uses CEPH as a distributed storage system (the same system used by Facebook, Dropbox, and other massive operations), Apache web servers for the frontend, Redis for caching, and Galera clusters for the database. None of these components are experimental. They have all been tried and tested for decades in installations millions of times larger. The complexity lies not in the individual components, but in their orchestrated interaction—and for this kind of compositional engineering, robust best practices now exist.

The narrative of dependency and its tacit reversal

One phenomenon in the German debate is noteworthy: The risk of vendor lock-in is constantly invoked for European or in-house solutions, while it is almost completely ignored when it comes to hyperscalers. This is analytically inconsistent. Vendor lock-in exists with Microsoft Azure, AWS, and Google Cloud just the same – it's simply structurally less transparent because the integration is deeper.

If a company deeply integrates its application logic into Azure's proprietary services—if it uses Microsoft Cognitive Services for AI, Azure SQL Database with its specialized features, and Azure DevOps for its CI/CD pipeline—then switching to alternatives isn't impossible, but it is massively expensive. The exit costs include not only data transfer (which can amount to millions with AWS), but also the redesign of integrations, retraining teams on other tools, and the lengthy transition period during which two systems run in parallel.

The hidden costs of this dependency have become the subject of intensive research. A European Cloud study documented that European cloud providers have significantly lower egress costs on average than the hyperscalers. While AWS charges $0.09/GB, many European providers charge nothing or a fraction of that. This is not a marginal saving—for large-volume transfers between applications, it can amount to thousands or millions. A company that moves data within a sovereign cloud infrastructure pays nothing extra for this. A company using Azure pays for every transfer between servers in different zones.

France observed and analyzed this cost structure and made a strategic decision: it did not want to accept the lock-in risk of others. Instead of becoming dependent on a hyperscaler not changing its pricing policy, not shutting down, not modifying its services – instead of submitting to the geopolitical mood in Washington or a CEO's change of strategy – it decided to retain control itself.

The French model shifts the dependence on external providers to internal expertise. This is not the same thing. Internal expertise must be nurtured, updated, and developed. But it is under state control. It cannot be shut down from the outside or suddenly made more expensive.

Our EU and Germany expertise in business development, sales and marketing - Image: Xpert.Digital

Industry focus: B2B, digitalization (from AI to XR), mechanical engineering, logistics, renewable energies and industry

More about it here:

• Xpert Business Hub

A topic hub with insights and expertise:

• Knowledge platform on the global and regional economy, innovation and industry-specific trends
• Collection of analyses, impulses and background information from our focus areas
• A place for expertise and information on current developments in business and technology
• Topic hub for companies that want to learn about markets, digitalization and industry innovations

Geopolitical reality and its economic consequences

The current geopolitical situation makes the question of control concrete. The United States has repeatedly signaled that it views its cloud infrastructure as an instrument of its foreign policy. The Cloud Act of 2018 grants the FBI the right to demand access to data stored by US companies, regardless of where that data is physically located. Microsoft has repeatedly defended this legal position and affirmed that it complies with such demands, even if European data protection laws would prohibit it.

This is not speculation. Microsoft has responded precisely to this risk with its "Sovereign Cloud" offerings. Google has announced a "European Sovereign Cloud." These offerings are therefore an implicit admission that standard cloud services are not under European control.

For Germany, this reality is particularly pressing. The federal government has decided to outsource billions of euros to Microsoft contracts – sometimes without open tendering, via so-called Microsoft-specific terms and conditions. For example, Autobahn GmbH wanted to spend 60 million euros on cloud services for four years and initiated a tender that was effectively open only to Microsoft. After interventions by competitors derailed the tender, the terms were reformulated. But the pattern remains consistent: Germany pays billions, while data sovereignty resides in San Francisco.

France has not chosen this path. This doesn't mean that France is ignoring the hyperscalers. But it has made a different decision for the public sector, for education, for critical infrastructure: it is retaining control. Apps.education.fr is merely a symptom of this fundamental approach.

Competitiveness in the digital age as a competitive advantage

A frequently overlooked dimension of this decision is its impact on long-term competitiveness. A Boston Consulting Group study shows that French companies and organizations invest significantly more in digital innovation than their German counterparts: 28 percent of French C-level executives plan to invest 30-50 percent of their budget in technology, compared to only 18 percent in Germany. Even more telling is the focus: French organizations invest more heavily in customer-centric platforms (12 percent) and business innovation, while German companies prioritize infrastructure modernization (32 percent). This isn't wrong, but it's reactive rather than proactive – solving existing problems instead of creating new opportunities.

If France builds expertise in cloud management, open-source integration, and distributed systems within its own administration, it will also create a pool of know-how that can be leveraged in the private sector. The people who run Nextcloud for a million users can pass this knowledge on to French technology companies. They can move to startups or found consulting firms. This technology transfer is automatic – it arises from organized practice.

Conversely, when a country completely outsources its digital infrastructure to external providers, it loses these competencies. German IT departments in government agencies and companies become administrators of Microsoft instances, not architects of sovereign systems. They acquire expertise in Microsoft services that only Microsoft can utilize. The nation's technological competence is stored in proprietary, non-portable formats.

France has interpreted this path differently: as an investment in its own capabilities. And these capabilities are becoming a strategic advantage in a world where digital sovereignty is increasingly critical.

The truth about costs: Why simple comparisons are misleading

A hasty conclusion drawn from cost comparisons is that on-premises is always cheaper than cloud. That's wrong. But it's also wrong to say that cloud is always more expensive. The truth depends on the context.

For startups and volatile workloads, cloud computing is rational. The flexibility comes at a price, but this price is justified for many applications. For stable, large, predictable workloads—like an education management system for an entire country—the TCO calculation is different. Over a five-year period, on-premises or private cloud can be significantly cheaper.

The French decision in favor of Apps.education.fr wasn't made because someone had done a cost analysis. It was made because the OVH debacle had shown that even European providers can fall victim, and because the issue of control had become paramount. But a cost analysis would have supported this decision as well.

A simple example: 1.2 million users with 100 GB of storage each equals 120 petabytes. With AWS, egress costs alone would amount to several million per year, even with intensive use. With Apps.education.fr, these costs do not arise – they are absorbed by the existing infrastructure. This is not a theoretical advantage; it is a structural advantage that pays off year after year.

The irony of the European cloud initiative Gaia-X and its limited impact

Interestingly, the European debate has responded to this French reality with initiatives like Gaia-X – a project that aims to standardize and federate European cloud infrastructure. Gaia-X has been underway since 2019 and is intended to create a data infrastructure that meets European data protection standards and enables interoperability.

Gaia-X is a welcome initiative. But it also illustrates a European dilemma: France didn't wait for a Europe-wide initiative; it simply took action. Apps.education.fr has existed since 2018. Gaia-X was founded in 2019 and is still in its pilot phase. French pragmatism—acting, not waiting—has led to practical results, while European coordination is still being debated.

This doesn't mean that Gaia-X is pointless. It simply means that national initiatives sometimes yield results faster than European harmonization projects. And it means that countries willing to act nationally have a first-mover advantage.

The German paralysis and its structural causes

Germany finds itself in a peculiar situation. The findings are clear: 91 percent of German companies are dependent on non-European technology providers. 60 percent expect this dependence to increase. 89 percent are calling on the federal government to boost competitiveness. And yet, investment patterns remain unchanged. Germany invests less in digital sovereignty than France and focuses more on infrastructure modernization than on innovative business models.

Why this paralysis? One reason lies in the institutional structure of decision-making. Large IT projects in Germany are often planned according to principles of risk avoidance, not risk optimization. An open-source project is considered risky because there is no single point of authority to which problems can be addressed. A project with Microsoft is considered secure because Microsoft exists and contracts are in place. The fact that this risk assessment is irrational—that vendor lock-in with Microsoft is often greater than with open-source software—is systematically ignored.

A second reason lies in path dependency. Germany opted for Microsoft ecosystems decades ago, and this decision perpetuates itself. People who use Windows learn Windows. Companies running on Azure build expertise in Azure. Switching would mean devaluing this expertise. This is a genuine cost argument, but it's an argument for status quo bias, not for rational optimization.

France does not have this path dependency. Or rather, it has structured it differently. By building public-sector IT on open-source foundations, it creates new paths that do not lead to US providers.

Digital sovereignty as a strategic concept and its economic reality

The debate surrounding digital sovereignty is often framed in moral terms: as if it were about national honor or ideology. This is a misunderstanding. Digital sovereignty is economically rational. It's about maintaining control over strategic infrastructures and not being dependent on a foreign company changing its terms of service or on a foreign state having a different conflict of interest than one's own.

A country that controls its cloud infrastructure can:

– Enforce data protection standards without a California court imposing a different weighting
– Optimize costs without a company raising prices due to over-reliance
– Target innovation by providing non-proprietary infrastructure
– Build resilience by not relying entirely on the availability of foreign infrastructure
– Create jobs in the technology sector by establishing skill requirements that are not limited to a single product

The French model has all these features. It's not perfect. Nor is it the right model for all applications – some workloads rely on specialized cloud services offered only by hyperscalers. But for core functions, for administration, for education, for critical infrastructure, it is rational and increasingly mandatory.

The lack of a design alternative: An analysis of the Franco-German contrast

The fundamental difference between France and Germany lies in the question: Is digital infrastructure something to be shaped or something to be destined for? France answers: something to be shaped. Germany increasingly answers: destiny.

This answer is not inevitable. It is the result of decisions: decisions in finance ministries, IT departments, and tendering processes. It is the result of path dependencies that have become self-perpetuating. But it is not technically necessary.

Apps.education.fr demonstrates that an alternative approach is possible. Using open-source software, federal architectures, pragmatic expectations of perfection, and continuous improvement instead of relying on vendor support, states and administrations can operate infrastructure that is controllable, cost-effective, and sustainable.

Anyone who doesn't want to go down this path should say so clearly: We want the hyperscalers because we need their innovative power, because we don't want to build up our own expertise, because we're willing to accept the risk of vendor lock-in. That would be honest. Instead, the argument is that there is no alternative. That's wrong. The alternative exists and operates with a million users in a democratic country with high data protection standards.

Conclusion: Control is a choice, not a necessity

The analysis of Apps.education.fr and its context leads to an uncomfortable conclusion: Europe's dependence on American hyperscalers is not technically unavoidable. It is a political decision. Countries willing to invest in infrastructure control can do so. Countries that do not pay the price—not only in costs, but also in control, security risks, and lost skills development.

France has taken a different approach. With Apps.education.fr, it operates a cloud infrastructure supporting one million people, based on open-source software. The platform is complex, but not rocket science. It works. It's cheaper than the hyperscalers. And it's manageable.

Germany could do the same. The technology exists. The expertise can be developed. The costs are comparable or lower. It's a matter of decision. And that decision won't be made in San Francisco, but in Berlin.

☑️ Our business language is English or German

☑️ NEW: Correspondence in your national language!

Konrad Wolfenstein

I would be happy to serve you and my team as a personal advisor.

You can contact me by filling out the contact form or simply call me on +49 89 89 674 804 (Munich) . My email address is: wolfenstein ∂ xpert.digital

I'm looking forward to our joint project.

☑️ SME support in strategy, consulting, planning and implementation

☑️ Creation or realignment of the digital strategy and digitalization

☑️ Expansion and optimization of international sales processes

☑️ Global & Digital B2B trading platforms

☑️ Pioneer Business Development / Marketing / PR / Trade Fairs

Benefit from Xpert.Digital's extensive, fivefold expertise in a comprehensive service package | R&D, XR, PR & Digital Visibility Optimization - Image: Xpert.Digital

Xpert.Digital has in-depth knowledge of various industries. This allows us to develop tailor-made strategies that are tailored precisely to the requirements and challenges of your specific market segment. By continually analyzing market trends and following industry developments, we can act with foresight and offer innovative solutions. Through the combination of experience and knowledge, we generate added value and give our customers a decisive competitive advantage.

More about it here:

• Use the 5x expertise of Xpert.Digital in one package - starting at just €500/month