Skip to content
Hintelia Hintelia
Back to Governance

Sovereignty

Sovereignty Is Choice

Sovereignty begins when dependence becomes a conscious choice.

Hintelia

6 min read

In this article

The problem

In a 2026 IBM survey of 1,000 senior executives, 91% said they did not fully understand their organisation’s dependencies across AI vendors, models and infrastructure. Without that visibility, assessing costs and planning for disruption becomes harder.

These dependencies can accumulate through ordinary operating decisions. Your data was stored wherever your vendor chose to host it. A model you cannot inspect became load-bearing because it was the easiest to start with. Prompts and feedback can contain your people’s know-how; how a provider may retain or use that information depends on the service and its terms. A contract is renewed each year without reassessing whether it still meets the organisation’s needs. Each decision may have been reasonable on its own. Together, they can create dependencies whose costs and exit options have not been fully assessed.

The resulting exposure can take three forms:

  • Commercial. Changes to pricing or usage terms can increase costs. An organisation with few viable alternatives may have limited room to negotiate.
  • Political. Sanctions can restrict access to specific cloud services, as US restrictions on certain software services supplied to Russia illustrate. Gartner forecasts that more than 75% of European and Middle Eastern enterprises will move virtual workloads to environments designed to reduce geopolitical risk by 2030, up from less than 5% in 2025.
  • Operational. If a key vendor changed terms, restricted service in your region or withdrew tomorrow, would you know which operations were affected, for how long, and at what cost?

There is also a question of control over proprietary knowledge. Microsoft CEO Satya Nadella describes this concern as the Reverse Information Paradox: “You essentially pay for intelligence twice, once with money, and again with something even more valuable: the proprietary knowledge you must reveal to make that intelligence useful.”

Nadella warns of a potential imbalance: “The seller learns more and more about you as you use what you purchased, while you learn very little about what the seller is learning in return.”

That outcome is not inherent in every AI service. For example, OpenAI states that API data is not used to train its models unless the customer explicitly opts in. Data retention and training are separate questions, so both need to be assessed against the service’s terms and technical controls.

Nadella argues for a real trust boundary, with an organisation’s data, evaluations, memory and adapted models under its control. He also cites Palantir CEO Alex Karp’s view that customers “want to know they own the means of production, and it’s not being transferred to someone else.” These are arguments for retaining control of organisational knowledge, rather than evidence that every provider uses customer data in the same way.

Our point of view

Sovereignty means choosing your dependencies with a clear understanding of their cost and a tested way out.

The debate keeps dying on the same two rocks. On one side, sovereignty is treated as self-sufficiency, with everything in-house and everything local. That approach is impractical for many organisations. On the other, sovereignty becomes a buzzword on a slide or a procurement checklist, without changing how dependencies are managed.

Both positions overlook the dependencies every organisation relies on, from platforms and suppliers to energy grids and payment systems. Eliminating every dependency is neither practical nor necessary.

A dependency you have assessed, costed and can replace gives you greater control over the associated risks; it does not remove them. A service that is affordable and reliable may still be difficult to replace. The concern is dependence that has not been examined or actively managed.

So we hold every AI dependency to three tests:

  • Chosen.

    Someone in your organisation weighed it against a real alternative and said yes, for reasons they can still articulate.

  • Priced.

    You understand the costs and risks beyond the invoice, including what data is shared, how proprietary knowledge is protected and how switching costs affect your negotiating position.

  • Exitable.

    You have tested an exit plan, with realistic estimates of the time, cost and operational disruption involved.

These tests help distinguish dependencies you can manage from those that need further attention. Where a dependency meets them, you have a stronger basis for continuing to use the service. Where it does not, the gaps should inform your next decisions.

Understanding where your data sits, how your models are used and what it would take to switch providers can help you adopt better AI sooner, on terms your organisation has chosen.

What we do

We make your dependencies deliberate. Four moves:

  1. Map. We build the picture of your position across the four dimensions that decide sovereignty (data, model, knowledge and exit), with jurisdiction behind each. Most leadership teams have never seen this picture assembled. It takes weeks, and it changes the conversation in the boardroom immediately.
  2. Price. We quantify what each dependency really costs and what it would cost to lose: a repricing, a tariff, an export restriction, a vendor withdrawal. We use your cost and operational data to help you weigh the available options.
  3. Close. We fix the gaps that fail the three tests, and only those. In-house or open models where the economics justify them. Contracts designed for exit, with the exit rehearsed rather than assumed. Where a dependency passes, we say so, and you keep it with a clear conscience.
  4. Transfer. We build the judgement and the instruments inside your teams, where they stay after we leave. A sovereignty programme that depends on us forever has just replaced one unconscious dependency with another.

The payoff

Take the three exposures and turn each one around.

A price rise becomes a negotiation, because you have a costed alternative on the table and the vendor knows it. A tariff or an export rule becomes a rerouting exercise, because you already know which workloads move, where to, and in how many days. The board’s question, what happens if this vendor disappears tomorrow, gets an answer in one meeting, because you have rehearsed it.

There is also a benefit beyond resilience. When a platform offers a better fit, portable data, workable exit terms and internal expertise can make it easier to adopt.

The aim is to reduce dependency costs and exposure to geopolitical disruption while making it easier to adopt new capabilities.

With sovereignty, your organisation sets its own direction and has the means to follow it.

References

Hintelia

You bring the ambition; we bring the structure.

Contact us