For a long time, businesses viewed the cloud as an infrastructure issue and data as a business issue. On the one hand, IT teams focused on environments, storage, computing power and application migration. On the other hand, business units were looking to make better use of their data to inform decision-making, improve processes or develop new services.
This distinction is becoming less and less relevant these days.
The boom in artificial intelligence is further accelerating the convergence of these two worlds. High-performance AI relies on accessible, reliable and properly managed data. However, this data must also be stored and processed in an environment capable of meeting requirements in terms of performance, security, compliance and sovereignty.
Data is only valuable if it can actually be put to use
“The key question is no longer simply how much data a company holds, but what it is able to do with it.” explains Malick Thiam, Sales Architect at Proximus NXT Luxembourg.
A company may have vast amounts of data without being able to derive any value from it. Data scattered across different systems, information that is difficult to access, inconsistent formats, a lack of governance or poor data quality: these are all obstacles to the use of artificial intelligence.
AI does not create value from data that is not accessible, contextualised and usable. So, even before we talk about models or tools, we need to consider the quality of our data assets
This reality puts data back at the heart of technology strategy. It also forces companies to take a fresh look at their infrastructure.
The Cloud is becoming the cornerstone of data strategy
For Jad Houdeib, Head of Cloud Consulting at Proximus NXT Luxembourg, the reasoning must work both ways.
A data strategy cannot be considered in isolation from cloud architecture. The environment in which data is stored and processed directly determines a company’s ability to utilise, secure and scale it. In practice, we still too often see cloud migrations carried out on an application-by-application basis, without any consideration of the data itself. This simply moves the silos rather than eliminating them.
One of the key benefits of the cloud is the ability to scale computing and storage resources rapidly in line with requirements. However, not all data needs to be stored in the same place.
Depending on business, regulatory or security requirements, a company may combine several environments: public cloud, private cloud, sovereign cloud, dedicated infrastructure, or hybrid and multi-cloud approaches.
“The issue is not about choosing between public and private cloud on ideological grounds. It is about building the architecture that best suits the organisation’s data, use cases and requirements. In practical terms, this begins with classifying data according to its sensitivity: what can be stored in the public cloud, and what must remain under enhanced control, or even in a sovereign cloud.” emphasises Jad Houdeib.
With AI, the boundary between data and the cloud is disappearing
Artificial intelligence makes this complementarity even more apparent.
AI models require significant storage, processing and computing capacity. Above all, however, they need access to the right data, at the right time and within a controlled environment.
This is particularly important for businesses wishing to develop applications using generative AI based on their own information: internal documents, knowledge bases, customer data, procedures or business history.
“A company may have access to the best AI model on the market. However, if it does not know how to connect this model to its own data in a secure and meaningful way, it will not be able to realise its full potential.” summarises Malick Thiam.
It’s a question of performance, but also of trust
This convergence of data, the cloud and AI also raises a fundamental question: that of trust.
Businesses need to know where their data is stored, who can access it and how it is protected. With AI, an additional question arises: what data is fed into the models, where are these models run, and who can access the results they produce?
These issues are of particular importance to organisations that handle sensitive data or are subject to strict regulatory requirements.
“Technological performance cannot be separated from security, governance and compliance. We see this very clearly with generative AI assistants: they can draw on all the documents to which the user has access. If access rights have never been reviewed, AI simply exposes what was poorly protected. A cloud architecture must enable the organisation to innovate, whilst also maintaining control over its data,” points out Jad Houdeib.
The cloud strategy thus becomes an integral part of the data strategy and, more broadly, of the AI strategy.
Shifting from a technology-centred approach to a platform-centred approach
For businesses, the real change therefore lies in moving away from thinking in silos.
Rather than launching a cloud project, a data project and then an AI project separately, it is becoming necessary to view the whole as a coherent ecosystem.
This involves, in particular, defining a suitable architecture, establishing data governance, securing access and data flows, as well as identifying use cases capable of generating tangible value for the business, whilst keeping costs under control: if poorly managed, data and AI applications can quickly drive up cloud costs.
This is the comprehensive approach adopted by Proximus NXT Luxembourg: from maturity assessments and data mapping to the design of public, private, hybrid or sovereign cloud architectures, including the securing of environments, governance and support for initial AI use cases.
“The Cloud, Data and AI are not three separate projects. They are the different building blocks of a single transformation. The ultimate goal remains the same: to turn technology and data into value for the business.” concludes Malick Thiam.
The question, therefore, is no longer whether to choose between data and the cloud. Companies must now learn to think of them together.

