What is the European Sovereign AI Cloud?

The European Sovereign AI Cloud, or ESAI, is a joint offering from delaware, Combell and imec. It combines AI governance and operations from delaware, dedicated GPU infrastructure in European datacenters from Combell, and AI model evaluation and optimization from imec.

Why European sovereign AI?

Sovereignty does not have to mean isolating your organization from the wider AI ecosystem. For ESAI, it means giving you more choice and control: greater visibility into where sensitive workloads run, the ability to use dedicated European infrastructure when appropriate, and the freedom to select different AI models depending on the task. The gateway makes that hybrid approach possible. Sensitive or suitable workloads can be routed towards your private environment, while external models can remain available where your policies allow them. You gain control without having to standardize every AI use case on one model or replace the tools your teams already rely on.

Do we need to replace the AI tools we already use?

No. ESAI is designed to add a governance layer around the AI landscape you already have. The AI gateway can route workloads towards private infrastructure or external models depending on your organization's policies and requirements.

What is an AI gateway?

The delaware AI gateway provides a central access point for AI requests from employees, applications and AI agents. Depending on your policies, it can route requests, remove personal data, enforce team budgets and create an audit log.

Does ESAI mean that all our AI workloads have to run privately?

No. Private infrastructure is one option within the architecture. Organizations can continue using external AI models for workloads where that makes sense, while routing workloads to dedicated European infrastructure where sensitivity, scale or predictability justify it.

How does imec help us choose AI models?

imec evaluates and optimizes AI model deployment using your organization's own documents and tasks. This allows models to be compared based on your real workloads rather than solely on generic benchmarks.

How does ESAI help with AI governance?

The AI gateway gives organizations visibility and control over how AI requests are handled. It can enforce policies, maintain an audit log and provide information about where data was processed, helping organizations document compliance with requirements including the EU AI Act, NIS2 and DORA.

How do we get started?

The starting point is an assessment of your current AI landscape: usage and spending, tools in use, where company data is processed, governance and compliance exposure, future cost development, and workloads that may be suitable for private European AI infrastructure.

European Sovereign AI Cloud (ESAI)

European Sovereign AI Cloud (ESAI)

AI adoption is accelerating. So are the questions around it. How much are you actually spending? Which AI tools are employees using? Where is sensitive company data being processed? And how do you keep control as AI becomes part of everyday work?

The European Sovereign AI Cloud (ESAI) brings together the expertise of delaware, Combell (part of team.blue) and imec to give organizations a controlled route to private AI infrastructure in Europe. It combines AI governance and operations, dedicated European infrastructure, and independent model evaluation and optimization.

Implement the AI strategy that make sense for your business and stay in control of your costs, data, and technology choices.

book the talk

Let’s talk about your AI landscape

Want to understand where private AI could make sense for your organization? Let’s look at your current AI usage, costs, governance and infrastructure together.

We can discuss:

how AI is being used across your organization;
where your AI costs are heading;
governance, data and compliance considerations;
which workloads could benefit from private European infrastructure;
where external AI services remain the better choice.

45' conversation, no obligation.

AI adoption is growing, but so is the risk of losing control

Nearly half of employees who use generative AI do so through a personal account (Netskope Cloud and Threat Report, 2026). More than a third of organizations do not track their AI use against a budget (Flexera, State of the Cloud 2026). And Gartner expects the cost of tasks performed by AI agents to increase more than fivefold through 2028.

That leaves organizations facing three increasingly connected challenges:

Keep AI spending in check

As adoption spreads across teams, applications and AI agents, costs can grow quickly and quietly. A governed environment gives you visibility into usage and lets you enforce budgets at team level.

Keep sensitive data under control

Banning personal AI tools without offering an alternative, is unlikely to stop employees from using them. Give them a secure, useful alternative instead, while retaining control over where company data is processed.

Keep your technology options open

Not every task needs the same AI model. ESAI lets you combine private AI models with external AI services based on your policies and requirements, without locking your organization into a single model or provider.

One governed gateway for your AI

The delaware AI gateway is one governed access point in front of the AI tools that your organization uses.

Requests from employees, applications and AI agents pass through the gateway. Based on your policies, it can:

• route requests to your private AI environment or an external model;

• remove personal data;

• enforce budgets by team;

• create an audit log of AI usage.

This gives CFOs, CIOs and CISOs a shared view of how AI is being used, while providing information that can support your compliance documentation for regulations such as the EU AI Act, NIS2, and DORA.

And you do not need to abandon the AI services you already use. The gateway adds governance around your existing AI landscape and routes workloads towards private capacity where sensitivity, scale or predictability justify it.

delaware: build it

We assess how AI is being used across your organization, determine where greater control can add value, and design and operate the environment.

The delaware AI gateway turns your governance policies into practice, controlling how AI requests are handled across private and external models.

Combell: run it

Private AI workloads run on dedicated GPU servers operated by Combell, part of team.blue, in European datacenters.

That gives you dedicated capacity for private AI workloads and clarity over where those workloads are processed, rather than relying exclusively on consumption-based external AI services.

imec: optimize it

Through Workload Studio, imec.aistack evaluates AI models on your actual documents and tasks, rather than generic benchmarks alone.

That helps identify the models and infrastructure configuration that offer the right balance of quality and cost for your use cases. And because the private environment uses openly available models, you retain the freedom to switch as better options emerge.

Start with your own AI reality

You don't need to move everything to private infrastructure. And you don't need to start with hardware. Start with the facts.

How much are you spending on AI today? Which tools are employees actually using? Where is company data being processed? And which workloads would genuinely benefit from private European infrastructure?

To determine the right approach, we look at:

current AI usage and spending;
tools used across your organization;
data, governance and compliance exposure;
future AI cost development;
workloads suited to private European infrastructure.

From there, we determine where governance gives you value, where private AI makes economic or strategic sense, and where external AI services remain the better choice.

book the talk

Built from a challenge we experienced ourselves

delaware has seen first-hand how quickly enterprise AI consumption can grow. Nearly all delaware employees worldwide use approved AI in their daily work. Between May and August 2026, our AI usage increased by an average of 41% per month.

As our own AI tooling costs grew, we worked with Combell and imec on dedicated AI infrastructure. ESAI brings that same approach to organizations facing similar challenges around AI consumption, governance and control.

Ready to take control of your AI landscape?

The first step is understanding where you stand today. Together, we assess your current AI usage and spending, governance and compliance exposure, future cost development, and the workloads that could benefit from private European AI infrastructure.

From there, you can make an informed decision about where private AI makes sense and where your existing AI services should remain part of the mix.

Benjamin Jacobs

Director
Connect with Benjamin on LinkedIn

book the talk

Frequently Asked Questions

The European Sovereign AI Cloud, or ESAI, is a joint offering from delaware, Combell and imec. It combines AI governance and operations from delaware, dedicated GPU infrastructure in European datacenters from Combell, and AI model evaluation and optimization from imec.
Sovereignty does not have to mean isolating your organization from the wider AI ecosystem. For ESAI, it means giving you more choice and control: greater visibility into where sensitive workloads run, the ability to use dedicated European infrastructure when appropriate, and the freedom to select different AI models depending on the task.
The gateway makes that hybrid approach possible. Sensitive or suitable workloads can be routed towards your private environment, while external models can remain available where your policies allow them.
You gain control without having to standardize every AI use case on one model or replace the tools your teams already rely on.
No. ESAI is designed to add a governance layer around the AI landscape you already have. The AI gateway can route workloads towards private infrastructure or external models depending on your organization's policies and requirements.
The delaware AI gateway provides a central access point for AI requests from employees, applications and AI agents. Depending on your policies, it can route requests, remove personal data, enforce team budgets and create an audit log.
No. Private infrastructure is one option within the architecture. Organizations can continue using external AI models for workloads where that makes sense, while routing workloads to dedicated European infrastructure where sensitivity, scale or predictability justify it.
imec evaluates and optimizes AI model deployment using your organization's own documents and tasks. This allows models to be compared based on your real workloads rather than solely on generic benchmarks.
The AI gateway gives organizations visibility and control over how AI requests are handled. It can enforce policies, maintain an audit log and provide information about where data was processed, helping organizations document compliance with requirements including the EU AI Act, NIS2 and DORA.
The starting point is an assessment of your current AI landscape: usage and spending, tools in use, where company data is processed, governance and compliance exposure, future cost development, and workloads that may be suitable for private European AI infrastructure.

European Sovereign AI Cloud (ESAI)

Let’s talk about your AI landscape

AI adoption is growing, but so is the risk of losing control

Keep AI spending in check

Keep sensitive data under control

Keep your technology options open

One governed gateway for your AI

Three partners, one sovereign AI offering

delaware: build it

Combell: run it

imec: optimize it

Start with your own AI reality

Built from a challenge we experienced ourselves

Ready to take control of your AI landscape?

Benjamin Jacobs

Frequently Asked Questions

Related content

AI

IT