Digital sovereignty is currently a constant presence at industry conferences. It appears in agendas, keynotes, and panel discussions, is written into strategy papers, and widely promoted at a political level. The term conveys a clear idea: control over one’s own IT, independence from individual vendors, and the ability to shape digital infrastructure on one’s own terms.
The problem is not the ambition. The problem is the reality in which that ambition is expected to be achieved.
While digital sovereignty is being discussed, the economic and technological conditions that underpin it are shifting in a direction that makes it increasingly difficult to realize. The market for infrastructure components is currently experiencing what can reasonably be described as a global memory crisis. Prices for RAM and storage are rising, availability is becoming less predictable, and planning reliability is deteriorating. What used to be a stable and manageable part of IT is turning into a strategic constraint.
This is not an abstract development. It is already visible in day-to-day operations. Projects are delayed because hardware is not available in time. Budgets that were stable for years are no longer reliable. Capacity planning becomes increasingly uncertain, as pricing and delivery timelines are difficult to forecast. Decisions that once relied on clear calculations now have to be made under uncertainty. Infrastructure is no longer just a foundation; it has become a variable.
A key driver behind this shift is the rapid expansion of AI infrastructure. Large US hyperscalers are investing in so-called AI factories at a scale that goes far beyond traditional IT demand. These systems require massive amounts of memory, computing power, and specialized hardware, and they secure these resources early and in volumes that are simply out of reach for most other organizations.
This is not a market malfunction. It is the market working exactly as expected. Companies that operate global platforms secure the resources required for their growth. The important point is not the intent, but the effect. Hardware is not only becoming more expensive but also structurally scarce. Capacity is increasingly concentrated among a small number of large buyers, while everyone else is left operating with tighter margins and reduced flexibility.
And this is where the real tension around digital sovereignty begins.
When digital sovereignty becomes a financial question
Digital sovereignty is often framed as an architectural or strategic decision. On-premise, cloud, or hybrid, there is a tendency to believe that independence can be achieved by choosing the right model. Under current conditions, this view is no longer sufficient.
The central question is no longer what is technically possible. The more relevant question is what remains economically viable.
Operating your own infrastructure today involves far more than purchasing hardware. It includes ongoing operations, maintenance, redundancy, security, and the necessary personnel. At the same time, the cost of that hardware is rising under the influence of a demand structure dominated by a few major players. The global memory crisis is no longer a theoretical concept; it is an operational reality.
There is also a secondary effect that is often underestimated. Infrastructure costs do not remain isolated. Rising prices for memory and compute power affect entire IT strategies. Projects are postponed, scaling plans are revised, and innovation initiatives are slowed down or cancelled altogether. Infrastructure, which was once an enabler, increasingly becomes a limiting factor.
This leads to a structural shift. Digital sovereignty is no longer simply a matter of design; it becomes a matter of prioritization. Not every organization can afford to build and operate infrastructure at the level implied by the concept. For large organizations with significant budgets, this remains an option. For many others, it becomes increasingly difficult to justify.
As a result, a reality emerges that is rarely stated explicitly: digital sovereignty is not equally attainable. It is becoming a function of financial capacity.
Or more directly: sovereignty is no longer an architectural question. It is an investment question.
The new dependency in a seemingly sovereign landscape
At the same time, a second dynamic is taking shape, one that is often overlooked. The providers from which organizations seek independence are the same ones best positioned to absorb and manage the current market conditions.
Hyperscalers are not only driving demand for infrastructure; they are also the actors most capable of mitigating its side effects. They can absorb rising hardware costs, aggregate capacity, and deliver infrastructure as a service. For many organizations, this is not a strategic preference but a practical necessity in the face of growing complexity and uncertainty.
This creates a structural effect that is rarely addressed directly: the same investments that contribute to market scarcity also produce the services that make this scarcity manageable for customers.
The result is not a contradiction, but a reinforcement. As operating infrastructure becomes more complex and expensive, outsourcing it becomes more attractive. The decision to move to the cloud is often not ideological; it is economic.
Dependency does not disappear in this process. It changes form. Instead of relying on internally managed systems, organizations become dependent on external providers whose infrastructure they no longer control. This dependency is not inherently problematic, but it is real, and often more difficult to reverse than it initially appears.
The frequently proposed shift toward European or national cloud providers illustrates this point. In many cases, it is a sensible move, particularly from a regulatory or data sovereignty perspective. However, it does not fundamentally change the structure. Infrastructure is still externalized, operations are still delegated, and dependencies remain. Digital sovereignty, in this context, is reduced to a question of geography rather than control.
The uncomfortable reality is therefore this: while digital independence is being promoted, new dependencies are forming at the same time, not because of strategic failure, but as a direct consequence of economic conditions.
Digital sovereignty is a luxury
Digital sovereignty is not the wrong objective. It addresses legitimate concerns around control, security, and long-term strategic flexibility. But it is increasingly discussed without fully acknowledging the conditions under which it must be implemented.
The global memory crisis, rising infrastructure costs, and the dynamics of AI-driven demand are reshaping those conditions. They make it clear that digital sovereignty is not purely a technical choice; it is inherently an economic one.
Perhaps the real challenge is not to achieve complete independence. Perhaps it is to remain capable of acting. Being capable of acting means being able to make decisions and to revise them when necessary. It means understanding dependencies, choosing them deliberately, and limiting them where they become critical.
In the end, digital sovereignty is not defined by the absence of dependency. It is defined by whether an organization retains control over its direction.
If your organization is navigating increasing infrastructure costs and growing complexity, a pragmatic approach, beyond ideology and buzzwords, can make the difference.
