Achieving AI Sovereignty: Why Your AI Project Needs a Smart Data Store

Paul Speciale
Aug 7
3 min read

What does “AI sovereignty” actually mean?

Artificial intelligence (AI) is changing our everyday lives. But how much control do companies actually have over their own AI systems?

AI sovereignty means that a company develops, operates, and controls its AI itself – without hidden dependencies on large tech corporations or insecure clouds.

However, just because an AI model was trained in Europe doesn't mean everything is secure and under control. True sovereignty begins with the data—and thus with storage.

According to a Gartner study, by 2027, around 70% of companies will place great importance on data sovereignty and sustainability when selecting AI services.

What is object storage?

ingredient	Explanation
Data	The actual content, e.g. a PDF file, an image or a log file
Metadata	Additional information such as category, author, creation date, confidentiality
Object ID	A unique identifier for quick and secure identification
structure	No folder structure required – objects are stored flat and are accessible via IDs
Advantages	Scalable, searchable, automatable – ideal for AI and large data sets

Illustration of a secured digital file system with cloud, folder, padlock, and shield icon on a dark background. Symbolizes data security, access control, and digital — Ein moderner Objektspeicher als Grundlage für souveräne KI: Die Illustration zeigt die kontrollierte Speicherung und den Schutz sensibler Daten durch eigene Infrastruktur.

Why storage is crucial

Most AI systems require vast amounts of data—text, images, sensors, logs. This data not only needs to be stored somewhere, but also:

be sure
be quickly findable
be clearly structured
and above all remain under your own control

A modern object storage system is designed precisely for this purpose. It functions like an intelligent, digital filing cabinet—but for millions of files simultaneously.

AI systems aren't solely based on the computing power of large models—their quality depends significantly on the data they process. Without a reliable storage infrastructure, security gaps, inconsistencies, or biases arise. Access to current, structured corporate data is particularly crucial in RAG (Retrieval-Augmented Generation) systems: The storage thus becomes an active source of knowledge in the AI's decision-making process.

Typical problems - and how to solve them with smart storage

problem	What often happens	Solution with modern storage technology
Data is somewhere “in the cloud”	Nobody knows exactly where or with whom	Own storage with clear access rules
Access is not controlled	Anyone can read or change everything	Fine allocation of rights - who can see/do what?
Changes are not comprehensible	There is a lack of overview	Automatic log: Who did what?
Data protection rules are violated	Data leaves the country	Local storage - in Switzerland or the EU

What a smart storage device can do

function	Simply explained
Indexing	Content is automatically made searchable
Metadata	Additional information such as “subject”, “confidentiality” or “last modified”
Connection to AI systems	AI can specifically access suitable data
Vector search	AI also finds similar content – not just exact matches
API connection	Other programs can access it automatically
Governance mechanisms	Precise control over which data may be stored, processed or deleted and for how long

Why it is dangerous to give up control

Many companies blindly store their data with cloud providers without knowing what's really happening there. This can be expensive:

Data protection violations
Dependence on a provider
Unexpected additional costs
Lost knowledge when data is difficult to find

A good example is a car manufacturer that controls the design of its vehicle, but the software, data, and sensors come from third parties. Then it's no longer truly its own product.

Conclusion: If you want to create AI, you have to have control over your data

When companies work with AI, they need a solid foundation - just like a house needs a good foundation.

An intelligent object storage ensures that:

Goal	What the memory contributes
Keep control	Data remains within your own access
Comply with laws	Data protection is technically feasible
Make better use of AI	The right data is available at the right time
Creating trust	Transparency and traceability for all processes

Don't forget regulatory requirements

In the EU, but increasingly also in Switzerland, AI and data processing are under special scrutiny. Data protection laws such as the GDPR and the new Swiss Data Protection Act (nDSG) require technical documentation of where data is stored and how it is used. Modern object storage enables companies to meet these requirements technically through audit trails, encryption, and granular access rights – without slowing down operations.

About the author

Paul Speciale is Chief Marketing Officer at Scality, a leading provider of cyber-resilient storage solutions for enterprises and governments worldwide. He brings over 20 years of experience in technology marketing, including in cloud computing, object storage, and scalable IT infrastructures. Prior to Scality , Paul held senior leadership roles at companies such as Appcara, Amplidata, and Oracle.

He lives and works in California.