The European Commission took a step on June 25, 2026 that could reshape the cloud computing landscape. In a preliminary finding, Brussels signaled that Amazon Web Services and Microsoft Azure should be designated as gatekeepers under the Digital Markets Act, bringing the dominant cloud platforms under the bloc's most stringent competition rules for digital services. The move directly targets the growing entanglement of artificial intelligence with core cloud infrastructure, aiming to prevent what regulators describe as AI-driven lock-in and to enforce strict new interoperability requirements.

This is the first time the DMA, which previously ensnared tech titans in search, social media, and advertising, has been extended to cloud computing. The preliminary finding, while not a final decision, marks a pivotal escalation. If confirmed, AWS and Azure will be compelled to open their ecosystems in ways that could fundamentally alter how businesses build, deploy, and migrate AI workloads.

The DMA defines gatekeepers as platforms with entrenched market power that serve as crucial intermediaries between businesses and consumers. Criteria include annual EU revenue of €7.5 billion, market capitalization of €75 billion, and at least 45 million monthly active end users in the EU. Both AWS and Azure comfortably surpass these thresholds. But the Commission is zeroing in on two pain points that have intensified with the AI boom: the inability to easily switch providers because of proprietary AI services and data gravity, and the lack of effective technical interfaces that allow rival clouds to interoperate.

AI Lock-In: The Core of the Commission's Concern

The explosive adoption of generative AI has cemented cloud platforms as the default environment for training and deploying machine learning models. Microsoft, through its deep partnership with OpenAI, has made Azure the exclusive cloud for the GPT family of models. Amazon offers Bedrock and its own Titan models. These proprietary AI stacks create a powerful gravitational pull: once a company's data is in Azure, and its applications are built around Azure AI services, the cost and complexity of moving to another cloud become prohibitive.

The Commission's preliminary view is that this constitutes a form of unfair advantage that rivals such as Google Cloud, smaller European providers, and open-source alternatives cannot overcome. By designating AWS and Azure as gatekeepers, Brussels aims to impose restrictions on self-preferencing and bundling. Specifically, cloud providers could be required to-
- Offer APIs and tools that allow customers to migrate trained AI models, datasets, and associated pipelines to other clouds without excessive fees or technical hurdles.
- Ensure that AI services deployed on their infrastructure can seamlessly interconnect with similar services from competing clouds.
- Prohibit exclusive tie-ins where certain AI capabilities are only available if you also use the underlying cloud infrastructure for other services.

Interoperability Rules That Will Reshape Enterprise Cloud Strategies

The DMA's hallmark is its focus on contestability and fairness. For AWS and Azure, this translates into a mandate for horizontal interoperability. Under the rules being contemplated, the two cloud giants would have to provide standardized, well-documented APIs that allow workloads running on their infrastructure to be managed and monitored via third-party tools, and to move data in and out without throttling or egress fees that act as a barrier to exit.

For Windows-centric enterprises, this could be particularly disruptive. Microsoft has spent years integrating Azure deeply into the Windows Server ecosystem, Active Directory, and the Microsoft 365 suite. The ability to lift-and-shift those workloads to another cloud has been deliberately limited. If the DMA forces open, businesses could gain the ability to run hybrid deployments across Azure and, say, Google Cloud, with unified identity management and no data transfer penalties. The long-term effect might even compel Microsoft to decouple its AI services from Azure in the way that Office was eventually untethered from Windows.

Amazon's AWS is equally affected. Its broadest AI offering, SageMaker, is tightly woven into the AWS fabric. The DMA could require Amazon to document and support equivalent functionality on other clouds, a massive engineering challenge that could level the playing field for startups and open-source alternatives like Hugging Face or Mistral AI.

Implications for Microsoft and the Windows Ecosystem

For Microsoft, a gatekeeper designation carries both operational and strategic weight. The company has spent the last decade transforming itself from a licensed-software vendor into a cloud-and-AI powerhouse. Azure currently accounts for over half of its revenue, and AI services are its fastest-growing segment. Any regulation that weakens the stickiness of the Azure ecosystem endangers that trajectory.

Windows, while no longer the center of Microsoft's universe, remains the dominant desktop operating system in enterprise environments and an important on-ramp to Azure. Microsoft has used Windows to promote Azure-specific features such as Windows Admin Center for Azure Arc, Azure Active Directory (now Entra ID), and deep links to Azure virtual machines. If the DMA compels Microsoft to provide equivalent management capabilities across clouds, the strategic value of Windows as an Azure feeder could diminish. In the short term, though, enterprises running Windows Server and Windows 11 will likely welcome easier multi-cloud operations.

The designation process now enters a consultation phase. AWS and Microsoft can challenge the preliminary finding, present arguments about market definitions, and negotiate the precise obligations. Final decisions are expected within six to nine months, with compliance deadlines typically 18 months from designation.

What Enterprise IT Leaders Should Do Now

The writing on the wall is clear: cloud lock-in is now a regulatory risk, not just an architectural one. Enterprises that have gone all-in on Azure or AWS for AI should start reevaluating their architectures. Those who embraced multi-cloud strategies early are better positioned, but the DMA could reduce the cost and complexity of diversifying. Key steps include-
- Conducting an audit of current cloud dependencies, especially around AI model training, inference, and data storage.
- Demanding that providers commit to publishing open APIs and migration tools, even before being legally required.
- Investing in cloud-agnostic orchestration layers like Kubernetes and Terraform, which can abstract away proprietary hooks.

Beyond AWS and Azure: A New Era for Cloud Regulation?

The DMA case against AWS and Azure is a test balloon for a broader regulatory theory: that AI and cloud are becoming a single, vertically integrated market. If the designation sticks, it sets a precedent that could eventually pull in other hyperscalers like Google Cloud, Oracle, and perhaps even the AI model-as-a-service platforms themselves. European digital sovereignty advocates have long argued that dependence on US-based clouds for critical AI infrastructure is a strategic vulnerability. The DMA's interoperability mandates could accelerate the development of a federated European cloud, such as Gaia-X, by lowering the switching costs.

But the road ahead is fraught with technical and political challenges. Defining interoperability for AI workloads is not trivial. A model trained on proprietary Azure hardware using InfiniBand networking won't perform identically on Google's TPU pods, and forcing exact parity may prove impossible. Regulators will need to work closely with engineers to set practical standards that don't stifle innovation.

The Bottom Line

A decade after the cloud wars began, Brussels is stepping in to rewrite the rules of engagement. The preliminary finding that AWS and Azure are DMA gatekeepers is not the final word, but it signals a tectonic shift toward open cloud ecosystems. For Windows shops, the promise of seamless multi-cloud AI could soon become a regulatory right rather than a vendor concession. The pressure is now on Microsoft and Amazon to convince regulators—and a wary market—that their AI empires can coexist with genuine competition.