The real risk behind the Anthropic decision is not AI. It is control.

Two weeks ago, the US government imposed export restrictions on some of Anthropic’s most advanced AI models. The result was simple. Foreign users could no longer access them. And because that separation could not be enforced cleanly in real time, access had to be restricted broadly.

This was widely covered. But most discussions stayed at the surface. They focused on AI. They missed the underlying pattern.

This is not an incident. It is a signal.

What happened here is not an isolated technical issue. It is a structural reality: A government can decide who gets access to critical digital capabilities. And the companies providing those capabilities must comply.

That principle has existed for decades in areas like encryption and defence technology.

It now applies to AI. And there is no reason to assume it will stop there.

The uncomfortable question

If access to advanced AI can be restricted, what about the rest of your stack? What if similar controls are extended to:

• Cloud platforms
• Identity providers
• Security tooling
• Collaboration environments
• Data platforms

These are not optional tools. They are the operating layer of your organisation.

The economic asymmetry

European organisations are in a structurally weak position. Not because they lack technology. But because they lack leverage. Today, most organisations:

• Pay significant amounts to US vendors
• Store their data in those environments
• Build processes and workflows on top of them
• Feed those systems with the data that trains next-generation AI

And then:

• Pay again to use the capabilities that are built on top of that data

If access to those capabilities can be restricted then you are funding systems you do not control. And may not always be able to use.

A world of different speeds

The deeper risk is not a sudden shutdown. It is divergence. A world where:

• Certain jurisdictions have continuous access to the most advanced capabilities
• Others experience delays, restrictions, or limitations
• Innovation compounds faster on one side

All because of control.

The relevant question is not “will this happen again”

The relevant question is:

What would actually break in your organisation if access to parts of your stack is restricted?

• Which systems are critical?
• Where are you dependent?
• What is reversible, and what is not?

Most organisations do not have a clear answer.

A practical starting point

You do not solve this by replacing everything. You start by making the situation visible.

• map dependencies
• identify critical risks
• understand what can realistically change

That is the purpose of a Sovereign Workplace Assessment: clarity first, decisions second.

If this topic is relevant

If this raises questions internally, it is worth making it concrete. We help organisations map their dependencies and identify where they lack control.

→ Book a 30-minute session to start understanding your exposure.