Sovereign on paper, dependent in practice. That’s the contradiction many companies fall into: airtight data governance policies on one hand, and on the other, a single non-EU cloud provider running every critical workload. It’s the topic of the second episode of Bitrock Tech Radio, picking up where the first episode left off: the one where we explained why digital sovereignty has become a boardroom issue, not just an IT department one.
Today we’re leaving theory aside. We’re back with Franco Geraci, Head of Engineering at Bitrock, to find out how you actually build a sovereign infrastructure without giving up speed and innovation: which architectures, which tools, which method.
In the last episode we looked at why sovereignty matters so much today. So how can companies gain autonomy and control without sacrificing innovation and speed?
Let’s start with an honest premise: there’s no single right answer, only the ideal trade-off for each specific situation. The market is mainly shaped by three mainstream approaches, each defined by its own set of trade-offs.
Approach A relies on US hyperscalers in their sovereign versions. This model delivers the highest technological maturity, the greatest computing power, access to the most advanced AI models, and very fast implementation times. The level of control is high, though not absolute, since some latent exposure remains tied to the non-EU provider’s jurisdiction.
Approach B, on the other hand, operates within a fully European perimeter. This option drastically reduces legal exposure and relies on a fast-growing continental ecosystem; however, it comes with a trade-off in overall performance and capability, which end up slightly behind the overseas giants.
Finally, approach C adopts an entirely Italian perimeter, centered on the National Strategic Hub (Polo Strategico Nazionale) and local model providers. While this choice ensures maximum control and governance, it also means a wider technology gap and lower scalability.
These are three different balances between control, technological maturity, and speed. Fortitude’s distinctive value isn’t pushing one approach regardless of context, but helping the client think it through analytically. There’s no single right answer for the whole organization — it varies by individual business need.
Now that we’ve set the stage, let’s dig into the concept of “AI layer decoupling.” What does it mean, and how is it connected to achieving digital sovereignty?
Think of AI as a car engine. Right now, most companies build their technology solutions by welding this engine directly onto the application’s chassis. If they ever need to change the engine — for cost, performance, or sovereignty reasons — they’re forced to take the whole car apart.
Our strategy does the exact opposite. We put an intermediate layer between the business application and the AI models, set up as a sort of universal socket. The application only ever talks to this standard interface. Behind the socket, we’re free to plug in whichever engine fits best at any given time: the highest-performing US model for processing public data, or a European, open-source, or Italian model — possibly installed on-premise — for handling sensitive information.
This approach delivers an extraordinary benefit. Swapping out an AI model or moving a workload to a more protected perimeter stops being a massive rebuild-from-scratch project and becomes a simple configuration change, like turning a dial. This flexibility also makes it possible to quickly adopt the new, better models the market keeps releasing, without wasting past investments. It means eliminating vendor lock-in, gaining real operational freedom, and ensuring maximum business continuity even in the face of sudden regulatory changes, price swings, or geopolitical tensions.
In practice, how do Bitrock and the Fortitude Group bring all this into companies? With what tools?
Over time we’ve consolidated and structured a rigorous methodology built on four sequential phases.
The essential starting point is the sovereignty assessment. In this first phase we analyze the client’s data and workloads in detail to map their sensitivity level and determine which type of sovereignty applies to each one. This screening results in a transparent strategic map paired with a concrete work plan. Once the route is set, we move on to the second step: the architectural design of the solution. Then comes the third phase, the actual build of the infrastructure and software. The last piece is ongoing management, ensuring end-to-end coverage and support across the whole project.
Throughout the project’s entire lifecycle, the Fortitude Group stays true to four core guiding principles: system portability, reversibility of technology choices, absolute control over data, and auditability — that is, the ability to transparently demonstrate and trace, at any time, who accessed or modified a given resource.
What’s the precise role of each of the different units that make up the Group?
The Fortitude Group delivers its value through the synergy of four specialized units, each focused on a crucial part of the value chain. Bitrock brings method, architectural design, design, and enterprise-grade delivery, orchestrating digital transformation. ProActivity steps in to provide flexible technical capacity and expertise wherever projects need operational continuity and team scalability. Radicalbit is our component dedicated to governance and AI, providing the advanced software tools needed to manage, monitor, and keep production models under control (MLOps).
Last but not least is Waterstream, our native vertical solution built specifically to connect the Internet of Things (IoT) directly to the Apache Kafka data streaming platform. In simple terms, this technology collects real-time data streams from the field — like sensors and industrial machinery via the MQTT protocol — and feeds them natively into Kafka, eliminating the cost and architectural complexity of traditional integrations. This isn’t a generalist suite; it’s a specialized, vertical solution built specifically for the Industry 4.0, manufacturing, energy, and smart city markets.
The real differentiator lies in combining these skills: we offer verifiable control, full reversibility of choices, and a markedly faster time-to-market for the same organizational structure. This isn’t about isolated experiments, but a different way of building enterprise software solutions. We help the market move past abstract slogans and embrace a sovereignty that’s concrete, measurable, and provable through facts.
Thanks to Franco for joining us again and for helping us make sense of such a complex topic, guiding us through the key steps for turning digital sovereignty from an abstract concept into everyday business practice.
If you’d like to learn more about our approach, or want to figure out how to apply these principles to your own organization, get in touch — let’s take the first concrete step together.