For years, the AI infrastructure buying decision has largely come down to three things: price, performance and access to the latest models.
That calculation is starting to change.
For businesses operating in Europe, a different set of questions is increasingly entering the procurement conversation: where does our AI run, who controls the infrastructure, and what happens to our data once it gets there?
We recently ran a survey among our prospect and user base, with 342 people taking part, primarily across software and developer focused organisations. For 67% of respondents, sovereignty was a reason for choosing their AI infrastructure without price being a factor. By comparison, just 1.5% selected price without a sovereignty related reason.
That does not mean businesses have stopped caring about cost. It means sovereignty is increasingly being considered on its own terms, rather than simply as a consequence of getting a cheaper or better performing service.
AI sovereignty is becoming a business concern
For a long time, AI sovereignty was largely a conversation for governments and policymakers. It was about strategic autonomy, foreign technology dependencies and whether Europe could build enough of its own computing capacity.
As AI moves from experimentation into customer facing applications and business critical workflows, the infrastructure underneath it becomes a business concern.
Where is our data processed? Is it retained? Which jurisdiction governs the infrastructure? How dependent are we on a provider we have limited control over? What happens if that provider changes its terms, suffers a disruption or becomes subject to restrictions outside our control?
These aren’t abstract geopolitical questions anymore. They are procurement questions.
Sovereignty is increasingly about managing risk
The shift is visible beyond the AI infrastructure market. PwC’s 2025 EMEA Cloud Business Survey, based on more than 1,415 business and technology leaders across 26 territories, found that 82% of organisations were refining their cloud strategy in response to geopolitical and regulatory change, with sovereignty and trust becoming increasingly important considerations in cloud strategy.
For businesses, sovereignty is increasingly about managing risk: understanding where sensitive data goes, reducing unnecessary dependencies and having greater confidence over the infrastructure supporting critical workloads.
A European business can still want access to the world’s best AI models while demanding greater control over the infrastructure through which its data and workloads are handled.
Sovereignty can mean having greater choice and control over how those workloads are run.
Control matters when AI becomes business critical
That matters even more when AI is no longer an experiment but part of the business. Among respondents using AI in production or describing it as business critical, 97% cited EU data residency and 84% cited zero data retention.
The demand for control is also not confined to a handful of highly regulated industries. EU data residency was cited by more than 83% of respondents in every sector measured, including software and developer tools, legal tech, healthcare and finance. Zero data retention was particularly strong in legal tech, where 92% cited it as a reason for choosing their AI infrastructure.
Regulation is clearly part of this conversation, but it isn’t the only driver. Data sensitivity, operational resilience and vendor dependency are business concerns in their own right.
None of this means sovereignty can compensate for poor infrastructure. Businesses still expect reliability, performance and usability. A provider that offers strong data controls but cannot deliver a dependable service will not win serious workloads.
Sovereignty is one consideration among several
The next phase of competition in AI infrastructure will therefore involve several considerations at once. For businesses operating in Europe, sovereignty may determine which providers make it onto the shortlist. Price, performance and reliability will continue to shape the final buying decision.
That changes how AI infrastructure needs to be considered inside the business.
AI infrastructure is becoming part of business risk
As AI becomes embedded in how organisations operate, the infrastructure supporting it becomes part of the business risk. Questions about jurisdiction, data handling, provider dependency and operational control increasingly sit alongside the traditional questions about cost and performance.
The question for European businesses is no longer simply how much does our AI cost?
It is also where does it run, who controls it, and can we trust it with the work that matters?
