The Sovereign Hyperscaler: T Cloud Public | EuropeanSwitch EuropeanSwitch

By EuropeanSwitch Editorial · 25 April 2026 · 9 min read

At a glance
In February 2026, Deutsche Telekom rebranded Open Telekom Cloud as T Cloud Public and pledged full core feature parity with US hyperscalers by year-end. The company calls it Europe’s answer to AWS, Azure, and Google Cloud. But the term “sovereign hyperscaler” is being used in two incompatible ways, one legal, and one technical. Buyers who conflate them will make the wrong procurement decisions.

What Deutsche Telekom actually announced

On 2 February 2026, Deutsche Telekom rebranded Open Telekom Cloud, a platform it has operated since 2016, as T Cloud Public. The Industrial AI Cloud, a Munich-based GPU infrastructure built with NVIDIA and data-centre partner Polarise, went live two days later.

The specifics from Telekom’s announcement:

80% of “core hyperscaler features” available today, 100% by end of 2026
Over 4,000 enterprise customers, including DAX-listed corporations
BSI C5:2020 Type 2 certification, zero-trust architecture, security-by-design
Data processing in German, Dutch, and Swiss data centres
An automated migration tool and professional services for customers leaving US hyperscalers

The framing from Dr Ferri Abolhassan, CEO of T-Systems: the era of choosing between “maximum functionality from overseas or European sovereignty” is over.

The Industrial AI Cloud deserves particular attention. Built in Munich’s Tucherpark on nearly 10,000 NVIDIA Blackwell GPUs, it is positioned as sovereign compute for workloads where US hyperscaler use is constrained by export controls, sector regulation, or contractual sovereignty clauses — defence, automotive R&D, pharma, and public sector AI among them. Announced anchor users include Siemens (integrating its SIMCenter simulation portfolio), Mercedes-Benz and BMW Group (for AI-powered digital twin simulation via Siemens), SAP (providing the Business Technology Platform layer in the joint “Deutschland Stack”), and Perplexity (for in-country AI inference). That is a different customer profile from OVHcloud or Scaleway, and the piece Telekom is pitching is sovereignty plus GPU scale rather than sovereignty alone.

That is a significant claim which warrants scrutiny.

Two definitions of “sovereign hyperscaler”

The phrase “sovereign hyperscaler” is being used in two incompatible ways in current European cloud marketing.

Definition one: legal jurisdiction. A cloud provider is a sovereign hyperscaler if it is incorporated in an EU member state, majority-owned by European entities, and immune to extraterritorial legal demands, particularly the US CLOUD Act. This is the definition applied in the European Commission’s €180M sovereign cloud framework contract awarded in April 2026. Under this definition, AWS, Microsoft Azure, and Google Cloud do not qualify, not because of capability gaps, but because their US parents remain subject to US law.

Definition two: technical capability. A cloud provider is a sovereign hyperscaler if it offers the breadth and depth of services that define AWS, Azure, and Google Cloud, compute, object storage, managed databases, container orchestration, AI/ML platforms, identity management, observability, and the hundred-plus ancillary services that make a hyperscaler a hyperscaler rather than an infrastructure provider. Under this definition, feature count and service maturity are what matter.

The 2025 ISG Provider Lens report for EU Sovereign Cloud Infrastructure Services, published in January 2026, applies a hybrid of these definitions. It named seven providers as Leaders: AWS, Deutsche Telekom/T-Systems, Google, Microsoft, Oracle, Orange Business, and OVHcloud. AWS received the highest score on portfolio attractiveness.

Under the Commission’s legal definition, four of the seven ISG Leaders do not qualify as sovereign at all. Under ISG’s capability definition, the US hyperscalers score as well as or better than their European counterparts.
Both definitions are defensible and measure different things. But when a provider, or a buyer, uses “sovereign hyperscaler” without specifying which definition applies, the term becomes a marketing artefact rather than a technical category.

What T Cloud Public’s parity claim actually means

Deutsche Telekom’s “80% of core hyperscaler features today, 100% by end of 2026” claim is specific enough to interrogate.
“Core features” is the load-bearing phrase, AWS alone offers more than 200 distinct services. Azure offers a similar number. Google Cloud somewhat fewer. No reasonable interpretation of “parity” implies matching this breadth, and Telekom is not claiming it does. The pledge covers core infrastructure: compute, storage, networking, managed databases, container services, basic developer tooling, and a defined AI services layer via the Industrial AI Cloud.

What this excludes is the long tail: specialised analytics services, industry-specific managed platforms, mature identity federation across thousands of third-party SaaS tools, and the ecosystem of marketplace integrations that hyperscaler buyers often take for granted.

For organisations running standard enterprise workloads, virtualised compute, containerised applications, conventional relational and NoSQL databases, GPU-backed AI inference, T Cloud Public’s roadmap is credible and the existing 4,000-customer base supports that. For organisations whose architecture depends on AWS-specific services like Lambda@Edge, DynamoDB Global Tables, or SageMaker Ground Truth, or their Azure and Google Cloud equivalents, “parity” will not mean drop-in replacement.

This is not a criticism of T Cloud Public. It is a matter of being precise about what parity means and what it does not.

Why T-Systems was not on the Commission’s framework

An obvious question, given Deutsche Telekom’s sovereignty positioning: why was T-Systems not among the four consortium leads on the Commission’s €180M sovereign cloud framework, alongside Post Telecom, Proximus, OVHcloud, and Scaleway?
The Commission has not published the per-bidder evaluation scores, and Deutsche Telekom has not publicly commented on whether T-Systems bid. So any answer is inference rather than fact. But three structural observations are worth making.

First, the Commission awarded a framework to four consortium leads covering specific workload categories. A framework is not the entire addressable market. Absence from a framework does not imply disqualification from future tenders, it reflects the specific requirements of a specific procurement.

Second, T-Systems’ traditional strength is in managed services for large enterprises and regulated sectors, not in the self-service public cloud model that AWS, OVHcloud, and Scaleway have optimised for. The Commission framework emphasised scalable self-service infrastructure. That plays to different strengths than the T-Systems business has historically prioritised.

Third, the rebrand and feature-parity expansion announced in February 2026 is, by Deutsche Telekom’s own admission, a work in progress. A 2025 procurement process evaluating 2025 capability would have evaluated Open Telekom Cloud as it then was, not as the expanded T Cloud Public plans to be by end of 2026.

The Commission framework and T Cloud Public’s expansion are running on different clocks. The interesting question is not why T-Systems was absent from the April 2026 framework, but whether it will be a credible contender for the next one.

Who should care, and how

For procurement teams evaluating cloud infrastructure in 2026, the T Cloud Public expansion changes the calculus in three specific ways.
For regulated industries with strict EU-jurisdiction requirements, financial services under DORA, healthcare, public sector, and T Cloud Public joins a shortlist that already includes OVHcloud, Scaleway, and Exoscale. The relevant question is not whether T Cloud Public qualifies as sovereign under legal-jurisdiction criteria but whether its service breadth matches the specific workloads being migrated.

For enterprises running substantial AWS, Azure, or Google Cloud workloads considering a partial or full sovereign migration, the automated migration tool Deutsche Telekom has announced matters more than the feature parity pledge. In practice, migration friction rather than feature gaps tends to be what stalls hyperscaler-to-European-provider projects. The tool’s quality will determine whether “sovereign migration” moves from boardroom slide to production reality.

For organisations that are not in regulated industries and have no specific sovereignty obligation, the sovereign hyperscaler question is still worth engaging with, but on commercial rather than jurisdictional grounds. European providers compete on price, regional data centre density, and freedom from US-dollar billing exposure. Those are reasons to evaluate them. “Sovereignty” is not the only frame.

For providers outside the named leaders including the growing cohort of mid-sized European cloud companies, StackIT among the, the expansion of T Cloud Public raises the bar. Competing with OVHcloud and Scaleway is one challenge. Competing with a Deutsche Telekom platform targeting full hyperscaler feature parity is another.

What this doesn’t mean

Three precisions worth stating.
It does not mean European sovereign cloud has solved the hyperscaler problem. T Cloud Public at end of 2026 will be one credible option among several. The ecosystem gap, marketplace integrations, third-party SaaS federations, specialist managed services remains.

It does not mean AWS, Microsoft, and Google Cloud are losing the European market. Their European revenue is growing, and ISG rates them as Leaders in sovereign cloud capability. What is changing is the share of regulated and public sector workloads accessible to them. That is a meaningful but bounded shift. It does not mean “sovereign hyperscaler” is a stable category. The term will continue to mean different things to different audiences. Buyers who ask providers to specify which definition they are claiming under legal jurisdiction, technical capability, or both, will make better procurement decisions than buyers who treat the phrase as self-explanatory.

This analysis is editorial. EuropeanSwitch is an independent directory and does not advise on procurement. Providers mentioned in this article are listed in our directory on the basis of European headquarters and European majority ownership.

The Sovereign Hyperscaler: What T Cloud Public’s Expansion Signals