Digital sovereignty is not a question of complete technological independence, but of control over data, infrastructure and digital processes. That was the key message of the 126th Digital Dialogue, held on 19 October 2026 at the Raiffeisen headquarters in Raaba near Graz. Experts from the fields of research, the IT sector and public administration discussed how Austria and Europe can reduce their dependence on international technology conglomerates. The focus was on the capabilities of domestic data centres, open AI models and specific real-world applications.
Digital sovereignty requires choices
Matthias Rüther, Director of the DIGITAL Institute at JOANNEUM RESEARCH, made it clear that digital sovereignty should not be equated with complete self-sufficiency. Europe cannot produce all its semiconductors or all its software components itself. What is crucial, rather, is to have alternatives to large international providers and to be able to make decisions about its own digital infrastructure.
In addition to technological dependencies, he highlighted the security of international supply chains. He argued that it was not enough simply to certify individual components or suppliers. It is only the interaction of all components that determines whether an overall digital system is truly secure and resilient. Particularly in the face of geopolitical tensions, the ability to operate and further develop digital systems independently is becoming increasingly important.
Domestic infrastructure rather than dependence on hyperscalers
Wilfried Seyruck, Managing Director of the Raiffeisen Data Centre, demonstrated that Austria already has high-performance alternatives at its disposal. “Digital sovereignty requires local infrastructure – and it must be Austrian-owned,” explained Seyruck. A data centre located on Austrian soil alone does not guarantee digital sovereignty if it is owned by an international corporation and subject to its legal framework.
Seyruck is particularly critical of the concentration of the cloud market amongst a small number of large US providers, known as ‘hyperscalers’. These providers offer extensive computing and storage capacities worldwide. Under the US CLOUD Act, US authorities may, under certain legal conditions, also demand access to data stored by US companies outside the United States. This poses a risk to companies with sensitive business and research data.
Domestic data centres enable companies to operate their IT systems and data in Austria. High-performance infrastructure for artificial intelligence is now also available. This means that AI models can be run on local hardware and integrated into business processes without having to transfer sensitive information to external AI services.
Not every AI application needs the most powerful model
Wolfgang Pree, a professor at the University of Salzburg, used specific examples to demonstrate that the largest and most expensive AI models are not actually necessary for many operational tasks. In his view, around 80 per cent of industrial and commercial AI applications can be covered by open-source models, which cost a fraction of the price of leading commercial systems.
Pree explained that, whilst particularly powerful AI models are now capable of tackling highly complex scientific problems, these capabilities are not required for most business applications. What is far more important, he said, is to select suitable models for the task at hand and combine them efficiently.
The examples demonstrated that locally operated AI models are already capable of meeting numerous practical business requirements. For Pree, these developments open up opportunities not only to reduce costs, but also to modernise existing software and reduce technological dependencies.
Intelligence comes to the data
Christian Günther, managing director of the Styrian IT company Speicherkraft, demonstrated exactly how local infrastructure can be utilised in business processes. In collaboration with a regional data centre operator, the company has developed a solution that enables AI-powered document processing and process automation.
“The data stays where it is,” said Günther, describing the basic approach. The solution links information from various business systems, such as ERP software, databases, file storage systems and email inboxes. This enables documents to be processed automatically, invoice data to be extracted, and information to be consolidated for operational decision-making.
The necessary computing power is provided via a locally operated AI infrastructure. For Günther, this is a key advantage: companies can utilise powerful AI without relinquishing control over their data and processes to international platform providers. At the same time, the collaboration demonstrates that regional IT companies can work together to develop practical and cost-effective AI solutions.
Safety remains an ongoing task
A discussion followed the keynote presentations. Michael Stark, Managing Director of Comm-Unity EDV GmbH, made it clear that digital sovereignty did not only become important with the advent of generative AI. The company develops software for Austrian towns and municipalities and supports around 440 local authorities. In doing so, it processes sensitive information, including registration, personnel and administrative data. The majority of these solutions are hosted in a domestic data centre. However, data protection, reliable operation and digital sovereignty are inseparable elements of a responsible IT strategy.
Wilfried Jäger, Secretary-General of the Open Source Business Innovation Group (OSSBIG), expanded on this perspective by highlighting the importance of open-source software. He stated that open source is an essential prerequisite for breaking free from dependence on individual software providers. Openly accessible source code enables companies and public sector organisations to operate and further develop systems independently. However, this requires the necessary expertise and investment in their own operating environments.
Jäger also warned against assessing dependence on international cloud companies solely on the basis of economic considerations. Digital service providers have now themselves become part of geopolitical conflicts and could, as a result of political decisions, restrict or suspend services at short notice. This would give rise to risks that go beyond traditional economic dependencies. Digital sovereignty is therefore also a matter of risk management and long-term capacity to act.
The 126th Digital Dialogue thus made it clear that digital sovereignty can already be put into practice today. Domestic data centres, open AI models and regional technology partnerships offer businesses and public authorities concrete alternatives to global platform providers.
