Skip to content

What Is AI Sovereignty? And Why Is It Relevant Right Now?

Who has control over the AI your organisation uses? The answer to that question determines whether your data is safe, and whether you will still have access to your system tomorrow.

By soev.ai · 10 December 2025 · 5 min read · Sovereignty

Abstract visualisation of digital sovereignty

AI sovereignty sounds like an abstract concept. Something for policymakers and technologists. But in practice it comes down to a very concrete question: who has control over the AI your organisation uses?

The answer to that question determines whether your data is safe, whether you have vendor lock-in, and whether you will still have access to your system tomorrow if a foreign vendor changes its terms.

The definition

AI sovereignty means that your organisation has full control and flexibility over:

  • Which data which AI model sees, and which it does not.
  • Whether AI models are trained on your data.
  • Which model you use and how it is configured.
  • Where the infrastructure is located: in the Netherlands, in the EU, or on-premise.
  • Who the vendors are and what their terms are.

Sovereignty is not an all-or-nothing choice. It is a combination of architectural decisions that together determine how much control you have.

Why this is relevant right now

In early 2026, Anthropic, the maker of the AI model Claude, declined a contract with the US Department of Defense. The Pentagon wanted Anthropic to make its technology available for all lawful purposes, including scenarios such as mass surveillance and autonomous weapons systems. Anthropic refused and was designated a 'supply-chain risk'.

OpenAI did sign a similar contract shortly afterwards.

Lex Lubbers presents a sovereign AI solution at Gradient Demo Day

This exposes a fundamental vulnerability: if your AI platform is in the hands of a foreign company, you do not set the boundaries. Political pressure, legislation or an acquisition can affect how your AI systems work, without you being able to do anything about it.

The Dutch perspective

The Netherlands is hooked up to an American tech drip. The sale of Solvinity, which includes the infrastructure of DigiD, to the American company Kyndryl made this painfully clear.

AI does not make this dependency smaller, but exponentially greater. Whoever controls the AI layer ultimately controls the system.

The 2026-2030 coalition agreement acknowledges this. Digital autonomy and AI development are explicitly named as priorities. The government wants to reduce strategic dependencies in cloud.

AI sovereignty in practice

A consultancy that works with confidential client files can use soev.ai as a fully sovereign AI platform. The data stays in Dutch cloud, the model is open source and auditable, and there is no dependency on a single vendor.

If the political situation changes tomorrow, or a vendor changes its terms, the firm has all options open. That is AI sovereignty. Not an ideal, but a design choice.

Frequently asked questions

Is open source the same as sovereign?

Not necessarily. Open source means the code is transparent. Sovereignty goes further: it is also about where the model runs, who manages the infrastructure and whether you can migrate your data.

Can a small organisation be sovereign too?

Yes. Soev.ai is available as SaaS in Dutch cloud, even for small teams. You do not need your own data centre.

Want to know more?

Visit soev.ai/platform or read how custom AI implementation gives organisations full control.

Curious how this works on your data?

How a first call goes

  1. [01]
    You tell us what you are running into

    Which systems, which data and which rules.

  2. [02]
    We show how it works for others

    With the measurements to go with it.

  3. [03]
    Together we decide what phase 1 can be

    One use case, with a benchmark as the yardstick.

45 minutes · online or on site · no obligation