On June 25, 2026, the European Commission reached a preliminary view that Amazon Web Services and Microsoft Azure — the EU's largest and second-largest cloud providers — should be designated gatekeepers under the Digital Markets Act (DMA). The announcement follows a seven-month investigation that began in November 2025. What makes this ruling architecturally significant is not the conclusion itself, but the route taken to reach it: neither service meets the DMA's own quantitative designation thresholds.
How the Threshold Exception Works
The DMA creates a rebuttable presumption of gatekeeper status when a provider meets all three quantitative criteria under Article 3(2): annual EU turnover of at least €7.5 billion in each of the past three fiscal years (or market capitalisation of at least €75 billion), at least 45 million monthly active end users, and at least 10,000 yearly active business users — each sustained for three consecutive years. Neither AWS's cloud division nor Microsoft Azure clears this bar as standalone cloud services.
Article 3(8) gives the Commission an alternative: a formal market investigation weighing qualitative factors — market position, network effects, vertical integration, switching costs, and data advantages — even where the numerical thresholds are absent. This is the mechanism now invoked for the first time against cloud computing services.
The Case for Designation
The Commission's argument is structurally grounded. Three US hyperscalers — AWS, Microsoft, and Google — collectively hold roughly 70 percent of the European cloud market by revenue, while EU-headquartered cloud providers have watched their share fall from 29 percent in 2017 to a plateau of approximately 15 percent. AWS and Azure together account for the majority of EU infrastructure-as-a-service and platform-as-a-service revenue. Against that backdrop, the qualitative gateway criteria — "significant turnover," entrenched user bases, high switching costs, and AI service offerings that the Commission describes as "a decisive factor in cloud procurement" — read as genuine market conditions, not regulatory sophistry.
Executive Vice-President Henna Virkkunen framed the strategic stakes directly: cloud has become "a prerequisite for AI." The Commission's logic is that whoever controls cloud infrastructure will shape the adjacent AI services layer — and that structural obligations are worth imposing now, before lock-in deepens further. That is a serious argument, and it reflects a real policy concern about the infrastructure dependencies European businesses are accumulating.
The Case Against, and Why It Deserves a Hearing
AWS has objected forcefully, and its arguments are not frivolous. The company contends that the DMA's threshold design reflects a deliberate legislative judgment: only platforms that intermediate between large masses of businesses and end consumers — app stores, search engines, social networks — exercise the kind of structural power the law was built to address. Cloud infrastructure provides technical building blocks rather than intermediating between businesses and end consumers in the manner of an app store or search engine, and AWS argues it simply does not fit the DMA model. More concretely, Amazon cites data that 70 percent of European customers already use multiple cloud providers, and that 85 percent of global IT spending remains on-premises — suggesting a market where customer mobility exists and alternatives are real.
There is also a regulatory-overlap problem. The EU's Data Act — already in force — introduces a phased elimination of cloud switching fees and mandates technical interoperability between cloud services. If those provisions address the lock-in concerns the Commission invokes, layering DMA obligations on top risks redundant requirements with uncertain incremental benefit.
The threshold design is not procedural fine print. It is the proportionality mechanism the legislator built into the DMA to prevent its hardest rules from being applied to firms whose market power does not warrant them. Invoking the qualitative route here sets a precedent that those thresholds are negotiable when political priorities are high enough — which should concern any large-scale infrastructure provider operating in the EU.
What the Obligations Would Mean in Practice
If the Commission issues a final designation — after AWS and Microsoft exercise their rights of defence — both firms would have six months to comply with DMA obligations as applied to cloud. These include: prohibitions on self-preferencing their own services within the cloud stack (for instance, favouring Azure AI or AWS Bedrock in default configurations that disadvantage competing AI services), mandated interoperability enabling customers to run workloads across competing cloud environments without technical barriers, and data portability requirements allowing businesses to extract and transfer datasets without penalty. These are substantive product-architecture changes, not compliance paperwork.
Why the Precedent Matters Beyond Cloud
This decision's significance extends past AWS and Microsoft. It establishes that the Commission is prepared to apply the DMA's qualitative pathway to infrastructure services — not just the consumer-facing platform intermediaries (app stores, search engines, messaging services) for which the original DMA case was most straightforward. A parallel investigation — due to report in May 2027 — will examine whether the DMA needs structural modification to address cloud-specific characteristics, suggesting the current designation is the opening move in a longer regulatory sequence.
The proportionate-regulation question the Commission must ultimately answer, and that AWS and Microsoft will press hard in their defence submissions, is whether cloud infrastructure exercises the kind of durable, self-reinforcing gatekeeping power that justifies the DMA's most intrusive tool — or whether it holds a large but competitively contested position in a market where the Data Act and natural switching dynamics are already working. The preliminary finding says the former. The analysis behind that answer will determine whether this ruling is a proportionate response to genuine market failure, or a precedent that stretches competition law beyond its proper reach.