What France's AI Policy Means for European Businesses

Are You Still Working in the Cloud or Already Sovereign?

France has made it very clear in recent months where Europe is heading: away from uncontrollable dependency on US hyperscalers, towards technical and legal sovereignty — especially when it comes to Artificial Intelligence.

The French government's IT authority (DINUM) has issued a clear directive: central administrations are to gradually disengage from proprietary US cloud and office stacks and transition to European, sovereign alternatives. In parallel, the question of data sovereignty became very concrete before the French Senate: Microsoft was compelled under oath to admit that no complete guarantee can be given that data belonging to European customers will not be affected by US laws such as the CLOUD Act.

At the same time, the EU Parliament sent a clear signal with a vote of 471:68: Europe wants more control, more transparency, and more sovereignty in cloud and AI. The message is unambiguous: anyone planning digital infrastructure in Europe today can no longer ignore questions such as data sovereignty, governance, or the EU AI Act.

On the supply side, a new ecosystem is emerging. With "Euro Office", an initiative has been launched that aims to establish a truly sovereign office and collaboration suite "Made in Europe" — as a serious alternative to M365 and Google Workspace. Sovereign infrastructure is therefore no longer merely a political demand, but lived practice.

All of this shows: Europe no longer wants to be just a user of foreign platforms, but aims to build its own digital and AI infrastructure. And this concerns far more than just office software — above all, it concerns Artificial Intelligence.

What Is Actually Happening? DINUM, Senate, EU Parliament and Euro Office

To understand the dynamics, it is worth taking a closer look at the four key signals.

DINUM Directive: Sovereignty as an Architectural Principle

The French interministerial digital directorate (DINUM) marks a change of course with its directive. The goal is to reduce dependency on US hyperscalers in central government agencies and to migrate step by step to European, often open-source-based infrastructure. Sovereignty thus becomes a design principle of public IT architecture — not an "add-on" bolted on at the end of a project.

Microsoft Before the Senate: The Limits of Data Sovereignty

The hearing of Microsoft before the French Senate brought the legal reality home to many decision-makers. Even when data centres are physically located in Europe and "EU only" is contractually guaranteed, extraterritorial legal norms such as the CLOUD Act remain a residual risk. For critical infrastructure, the judiciary, public administration, or heavily regulated industries, this is difficult to reconcile with the claim of genuine data sovereignty.

EU Parliament 471:68: A Political Mandate for Sovereignty and Governance

With a clear majority of 471 votes in favour, the EU Parliament underlines its course towards greater digital self-determination. The resolutions address, among other things:

• stronger consideration of data protection, data sovereignty, and the EU AI Act in procurement decisions,

• the promotion of European cloud and AI infrastructure,

• higher requirements for transparency, auditability, and governance of AI systems.

Anyone introducing extensive AI systems in Europe will in future require an architecture that incorporates governance, compliance, and sovereignty from the very beginning.

Euro Office: Concrete Alternatives to Proprietary Office Stacks

With "Euro Office", an industry initiative is taking shape that is developing a sovereign office and collaboration environment based on European, often open-source technology. The goal is a suite in which organisations retain full control over data, metadata, access rights, and — looking ahead — embedded AI functions as well.

This creates precisely what many CIOs have been missing: a realistic alternative to "everything in M365", where sovereignty and compliance are not retrofitted as an afterthought, but are part of the foundational architecture.

The neuland.ai Perspective: AI Is Not a Tool — It Is the New Core of the Enterprise

At neuland.ai, we have been observing the same development that France and the EU are now framing politically for years in our own projects — with a focus on AI. Our central thesis is:

Artificial Intelligence is not a tool — it will become the new core of the enterprise.

Anyone who treats AI merely as another tool within an existing stack will inevitably end up in technological and organisational dependencies. Because AI fundamentally changes value creation, processes, and decision-making logic. In the future, companies will operate two "factories" at their core: one for their products, and one for AI.

This AI factory must not rest on intransparent, externally controlled platforms if sovereignty, compliance, and governance are to be taken seriously. This is precisely where neuland.ai comes in.

Organisations Must Reorganise Around AI

The French DINUM directive places sovereignty at the centre of public IT architecture. Translated to the corporate context, this means: AI must not be "somewhere" in the IT landscape, but becomes the central operational layer around which structures, processes, and responsibilities are aligned.

It is not sufficient to introduce AI tools on a point-by-point basis or merely to optimise existing structures. Anyone who wants to scale AI productively, securely, and in a legally compliant manner must align their organisation around an AI-centric operating logic. Shadow AI and "Bring Your Own AI" are clear symptoms that such a central AI infrastructure is still missing.

AI Management and Orchestration Platforms Are the Key

Europe's political direction is unambiguous: without its own sovereign infrastructure, digital sovereignty remains mere lip service. This applies to AI in particular. Modern AI landscapes do not consist of a single model, but of dozens of applications, hundreds of assistants and agents, and countless data flows and workflows.

Without a central management and orchestration backbone, the result is:

• isolated solutions without unified governance,

• unclear data flows and responsibilities,

• high risks with regard to data protection, compliance, and trade secrets.

Our experience from many projects: this is precisely where AI initiatives fail after successful PoCs. The technology works — but the platform on which everything can be operated in an orderly, secure, and scalable manner is missing.

This is where the neuland.ai HUB comes in. It is not another point solution, but an orchestration and management platform for Artificial Intelligence — "Made in Germany" and available as a SaaS solution or as a sovereign on-premises deployment.

The neuland.ai HUB enables companies and public authorities to:

• orchestrate AI applications, assistants, and agents across the entire organisation,

• securely integrate different models and services (on-premises, cloud, open source, proprietary),

• technically implement regulatory requirements such as GDPR, EU AI Act, DORA, BRAO, or ISO 27001,

• centrally manage roles, permissions, data access, costs, and degrees of automation.

This creates precisely the layer that France and the EU are demanding politically: an infrastructure in which sovereignty, governance, and compliance are not "audited" on top after the fact, but are technically anchored at the core.

Our customers — from mid-sized enterprises to financial institutions and law firms — are already using the neuland.ai HUB today to transform isolated AI experiments into a scalable AI infrastructure. They recognised early on: only in this way does AI move from "nice to have" to a productive, secure, and legally compliant enterprise core. And only in this way can a company truly become an "AI First Organisation".

From AI PoC to Sovereign AI First Organisation: Your Next Step

Developments in France and at EU level mark a turning point: sovereignty, governance, and technical control are becoming mandatory — especially for AI. For companies, this gives rise to a clear course of action.

Instead of purchasing yet another individual AI tool, now is the time to build your own sovereign AI platform. The neuland.ai HUB provides you with a stable and scalable foundation for this — for initial PoCs as well as for enterprise-wide rollout:

• You retain control over your data, models, and workflows.

• You can systematically fulfil regulatory requirements.

• You create an architecture in which AI no longer runs "somewhere", but becomes the new core of your organisation.

If you want to:

• reduce dependencies on individual AI or cloud providers,

• bring order and governance to a growing AI landscape,

• or plan the step from PoC to productive, legally compliant AI usage,

then we would be glad to support you.

Test our demo version of the neuland.ai HUB or speak directly with our sales team. Together we will develop a roadmap for how you can consistently develop your organisation towards AI First — on a sovereign, secure, and future-proof AI infrastructure.

Image generated with the neuland.ai HUB.

France / DINUM Directive – Tech Sovereignty PlanThe Register: "France's tech sovereignty plan: Government IT orders move away from US cloud giants"www.theregister.com

Microsoft before the French Senate – Data Sovereignty / CLOUD ActThe Register: "Microsoft admits it cannot guarantee EU public sector data is safe from US laws"www.theregister.com

EU Parliament – Vote 471:68 to Strengthen Digital Sovereignty / GovernanceEuropean Parliament: Legislative Resolution (Document TA-10-2026-0022)www.europarl.europa.eu

Euro Office – Sovereign European Office SuiteNextcloud – Press Release: "Industry initiative launches Euro-Office as true sovereign office suite" nextcloud.com