A few years ago, we investigated the state of European cloud providers as a sovereign alternative to US big tech in the field of analytics. You can read that earlier article here, but the conclusion at the time was clear: European providers were simply not there yet.
This was mainly because modern analytics platforms need more than virtual machines and storage. They need managed data warehouses or lakehouses, scalable compute, serverless workloads, orchestration, monitoring, security, integrations and increasingly AI capabilities. At the time, the US hyperscalers were far ahead in offering these capabilities as mature, connected services.
Now, several years later, the question deserves another look. The world of cloud has changed, and so have the providers. European cloud is becoming a strategic topic for governments, public institutions and companies that want more control over their data, their suppliers and their legal exposure.
TL;DR: European providers are still not equal to AWS, Azure or Google Cloud in breadth, but they have become much more credible for high-scale analytics than many people may assume.
The discussion around European cloud providers used to be mostly technical. Could they run the workloads? Were the databases good enough? Did they have proper APIs? Could a data team work productively without constantly missing features from the major hyperscalers?
Those questions still matter, but Cloud infrastructure has become part of Europe’s broader discussion about strategic independence. For analytics teams, this is especially relevant because analytics platforms often centralise valuable and sensitive data in an organisation.
Customer data, behavioural data, marketing data, financial data, product usage data and operational data often come together in one analytics environment. That makes the cloud provider behind it a strategic dependency, rather than just a technical supplier.
Several recent developments have made this more visible. The European Commission has moved forward with sovereign cloud procurement for EU institutions and agencies. France has pushed cloud sovereignty through its “Cloud au centre” doctrine, which makes cloud the default model for many state digital services while emphasising cybersecurity and protection of citizen and business data. In the Netherlands, the government signed a framework agreement with STACKIT, making it easier for public organisations to use a European cloud provider under pre-agreed conditions.
At EU level, the Data Act has also strengthened the conversation around switching, portability and protection against unlawful third-country access to data. This means sovereignty is not only about where a data centre is located. It is also about who controls the provider, which laws apply, how easily workloads can be moved and whether customers can avoid becoming too dependent on one ecosystem.
A very recent development has added more uncertainty to this picture. On 29 June 2026, the US Supreme Court backed the dismissal of FTC Commissioner Rebecca Slaughter and overturned long-standing precedent around the independence of certain US agencies. The Federal Trade Commission has historically played an important role in the enforcement foundation of EU-US privacy arrangements, which means the dismissal is a significant development in these arrangements.
This ruling does not automatically invalidate the EU-US Data Privacy Framework. However, it does make the legal and political foundation feel less stable. For EU-based organisations, the practical takeaway is that legal risk can change quickly, especially when data is hosted or processed under non-EU jurisdiction.
That is why European cloud providers deserve a renewed look. The question is whether they can help organisations reduce strategic, legal and operational dependency while still delivering a modern data stack.
To compare European providers fairly, we need to define what “ready for analytics” means. A high-scale analytics platform is not just a database. It is a combination of services that work together reliably.
First, there needs to be a managed place to store and query large volumes of data. This can be a cloud data warehouse, a data lakehouse or a query layer on top of object storage. For many organisations, this is the heart of the analytics platform.
Second, teams need flexible compute. Modern analytics platforms include ingestion jobs, API endpoints, enrichment services, server-side tracking, reverse ETL processes, data quality checks and event-driven workloads. These are often best served by serverless functions, serverless containers, serverless jobs or a platform-as-a-service model.
Third, orchestration is essential. Data pipelines rarely run in isolation. Jobs need to trigger each other, wait for dependencies, handle failures and create visibility for data teams. Without orchestration, organisations often end up with fragile cron jobs and undocumented workflows.
Finally, AI is becoming part of the analytics stack. This includes summarisation, classification, embeddings, document processing, semantic search, data assistants and agentic workflows. For a sovereign data stack, it is increasingly important that these AI capabilities can also run within a European cloud context.
With that in mind, we looked at three European cloud providers that currently stand out: Scaleway, OVHcloud and STACKIT.
Scaleway is a French cloud provider and part of the Iliad Group. It has grown into one of the more developer-friendly European cloud options, with a clear focus on modern infrastructure, serverless services and increasingly data and AI.
The most important analytics development is Scaleway Data Warehouse for ClickHouse. ClickHouse is a strong analytical database for high-volume, low-latency workloads such as product analytics, event analytics, BI dashboards, observability and log analysis. A managed ClickHouse service gives Scaleway a much more serious analytics foundation than it had a few years ago.
The clickhouse setup doesn’t offer the same level of ‘managed’ as BigQuery or Snowflake. Teams still need to think carefully about ingestion, modelling, storage and workload patterns. ClickHouse can be extremely powerful, but it rewards good architecture. Scaleway’s offering should therefore be seen as a strong analytical engine, not as a fully abstracted “just run any query and forget the rest” warehouse.
Scaleway is also strong in serverless compute. It offers Serverless Functions, Serverless Containers and Serverless Jobs. For analytics, the Jobs product is especially relevant because it can support asynchronous and resource-intensive tasks such as data processing, transformations and batch operations.
This makes Scaleway a good fit for a modular analytics architecture. A team could use object storage for raw data, managed ClickHouse for analytical queries, Serverless Jobs for ingestion and processing, and Serverless Containers or Functions for APIs and event-driven workloads.
The main gap is orchestration. Scaleway offers useful building blocks, but it does not yet appear to have the same obvious managed Airflow-style orchestration layer that many data teams expect. In practice, teams may bring their own orchestrator, such as Airflow, Dagster or Prefect, or use scheduled jobs for simpler workflows.
Scaleway also has a growing AI layer through its Generative APIs. These make it possible to use AI models through European-hosted infrastructure, which is useful for classification, summarisation, semantic search and internal assistants. The offering is promising, although it should be seen more as a sovereign AI inference layer than as a complete agentic AI development platform.
Overall, Scaleway is a strong option for technical teams that want modern cloud primitives, good serverless options and a performant analytics database. It is probably most attractive when the team is comfortable composing its own architecture.
OVHcloud is one of the most established European cloud providers. It has a broad infrastructure footprint and has been part of the European cloud conversation for many years. For analytics, the most relevant development is OVHcloud Data Platform.
Rather than only offering separate infrastructure components, the Data Platform aims to provide a more integrated environment for data integration, storage, preparation and analytics applications. This can be attractive for organisations that do not want to assemble every part of the stack themselves.
OVHcloud’s direction also fits the broader movement toward lakehouse architectures. Its platform refers to lakehouse capabilities and the use of open table formats such as Apache Iceberg. Open formats like this can help reduce lock-in and make it easier to separate storage, compute and governance.
Where OVHcloud feels less complete is in standalone serverless microservices. It is possible to run workloads through Kubernetes and the broader Public Cloud ecosystem, and the Data Platform includes some workflow capabilities, but teams looking for a very simple Lambda or Cloud Run-style experience may find the setup less direct.
This matters in analytics because many modern data stacks rely on small supporting services. Examples include webhook processors, server-side tagging endpoints, enrichment APIs, consent-aware routing and operational activation workflows. These workloads do not always require heavy infrastructure, but they benefit from simple deployment and scaling.
In our own work, this is why we created a library for deploying server-side Google Tag Manager on OVHcloud more easily. The platform can support this type of workload, but the developer experience is not always as frictionless as the equivalent setup on a US hyperscaler.
OVHcloud also has a growing AI offering. Its AI Endpoints service gives access to pre-trained models through APIs, including open-source LLMs and generative AI models. This makes it relevant for text classification, translation, summarisation, document processing and AI-assisted analytics use cases.
The main caveat is operational polish. In our own experience, removing resources and fully cancelling accounts can be harder than expected and may require support involvement. For small experiments, that is annoying. For larger analytics environments, clean lifecycle management matters because data teams create and remove many resources over time.
OVHcloud is therefore a good fit for organisations that want a broad European provider, a more packaged data platform and a path toward lakehouse analytics. It may be less ideal for teams that need a highly refined serverless developer experience around many small services.
STACKIT is the cloud provider from Schwarz Group, the parent company behind Lidl and Kaufland. That background makes it interesting. STACKIT grew from the infrastructure needs of a large European operating business, and it is now being positioned as a sovereign cloud platform for external customers.
For analytics, STACKIT’s most interesting service is its managed Dremio offering. Dremio is a data lakehouse platform that allows teams to query, manage and govern data across sources through a SQL interface. It supports open lakehouse patterns and is especially relevant for organisations that want to keep data in object storage while still enabling fast analytical access.
This gives STACKIT a serious analytics story. Dremio can act as a semantic and query layer on top of a modern data lakehouse, which fits well with organisations that want openness, portability and governed access.
However, teams should validate the current maturity carefully. Some documentation still refers to preview-stage limits for Dremio, including limits around active instances and engine sizes. That does not make it irrelevant, but it does mean production teams should check capacity, SLAs, support and roadmap before placing mission-critical workloads on it.
STACKIT stands out more clearly in orchestration. STACKIT Workflows is a managed service based on Apache Airflow. This is highly relevant for analytics teams because Airflow is widely used to schedule, monitor and manage data pipelines. A managed version reduces operational overhead and gives teams a proper foundation for pipeline orchestration.
For application runtime and microservices, STACKIT offers Cloud Foundry. This is a managed platform-as-a-service that helps teams deploy, scale and manage applications without handling every infrastructure detail themselves. It is not exactly the same as function-as-a-service, but it can work well for teams that need to run APIs, services and cloud-native applications in a managed European environment.
STACKIT also has a notable AI portfolio. Its AI Model Serving service provides managed hosting for AI models, including large language models, and offers an OpenAI-compatible API. That compatibility matters because many modern AI tools already support OpenAI-style interfaces, making integration easier.
Overall, STACKIT may be the most surprising provider in this comparison. Its combination of Dremio, managed Airflow and AI model serving makes it very relevant for a modern sovereign analytics stack.
European cloud providers are moving from theoretical alternatives to usable solutions for many high-scale analytics scenarios. They especially fit when organisations are willing to design around open, modular architecture rather than expecting a one-to-one replacement for US hyperscaler services.
Scaleway is strong for developer-friendly analytics platforms built around ClickHouse, serverless jobs and European-hosted AI APIs. OVHcloud is strong for broader infrastructure needs and organisations that prefer a more packaged data platform. STACKIT is strong for lakehouse architecture, managed orchestration and sovereign AI.
The biggest remaining gap is ecosystem maturity. AWS, Azure and Google Cloud still offer more managed services, more integrations, more documentation, more marketplace solutions and more operational shortcuts. European providers are catching up, but they are not identical substitutes.
This means architecture decisions need to be more deliberate. A strong European analytics stack will likely use open formats, portable orchestration, containerised workloads, object storage, managed analytical engines and AI APIs that avoid unnecessary lock-in.
For many organisations, the best next step is to start with a real sovereign analytics pilot, rather than fully migrating their stack. Choose one meaningful workload, such as server-side tracking, product analytics, customer segmentation, marketing attribution, reporting automation or document classification, and build it on a European provider. Then measure performance, cost, developer experience, support and operational friction.
That kind of proof of concept will give a much clearer answer than a theoretical comparison.
When we last reviewed European cloud providers, the conclusion was that they were not ready to replace US hyperscalers for serious analytics work. In 2026, that conclusion needs to be updated.
European providers still do not match the hyperscalers in breadth, but they have made real progress in the areas that matter for analytics. Managed analytical databases, lakehouse services, serverless compute, workflow orchestration and sovereign AI APIs are now available in more credible forms.
At the same time, the strategic need has grown. Regulatory uncertainty, EU data policy, public procurement, geopolitical tension and questions around EU-US data transfers all make cloud sovereignty more relevant for EU-based organisations. The sensible conclusion is that European cloud providers should now be part of the architecture conversation.
For high-scale analytics, the question is no longer only whether European providers are ready. The better question is which parts of your data stack can already run sovereignly, and what your organisation would gain by making that move.
For many teams, the answer will be more than expected.
Are you looking to take the first steps into sovereign European clouds, but not sure how / where to start? Feel free to reach out to us through support@cloudninedigital.nl for a practical conversation on how to move forward.