Photo source: Flickr
Major tech companies are increasingly focusing on the development of “sovereign” artificial intelligence (AI) models to enhance competitiveness by prioritising local infrastructure. Data sovereignty emphasises that individuals’ data should be stored within their own countries or regions.
“Sovereign AI is a relatively new term that’s emerged in the last year or so,” said Chris Gow, Cisco’s EU public policy lead.
Currently, many large language models (LLMs), such as OpenAI’s ChatGPT, rely on U.S.-based data centres, raising concerns among European politicians about dependence on American technology and its impact on competitiveness and resilience.
Origins of AI Sovereignty
The concept of data sovereignty has gained traction in Europe partly due to regulatory pressures like the GDPR, which mandates secure handling of user data. A 2020 ruling by the European Court of Justice invalidated an EU-U.S. data-sharing agreement, prompting the establishment of the EU-U.S. Data Privacy Framework to facilitate safe data exchange.
Filippo Sanesi from OVHCloud highlighted a growing demand for European-based infrastructure as businesses recognise the importance of local compliance.
“We see more and more companies understanding the importance of having your data locally and under a specific jurisdiction,” Sanesi stated.
Industry-Driven Initiatives
Gow emphasised that the push for sovereign AI is primarily industry-driven rather than regulatory. Companies are establishing more data centres in Europe to support cloud-based AI solutions. He added that “the AI workload uses 20 times the bandwidth of a traditional workload,” highlighting the need for skilled workers.
In Italy, the launch of Italia 9B introduced an LLM trained on Italian language data, which aims to produce results grounded in local culture and history. Nvidia’s David Hogan stated, “Sovereign AI is about reflecting the values of an organisation or, equally, the country that you’re in.”
Regulatory Influence on Localisation
Regulations have not fully propelled localised AI infrastructure yet, but they have influenced industry perspectives. Sanesi noted that GDPR has sparked interest in regional data processing.
Local tech firms are also embracing AI sovereignty. For instance, Berlin-based Ecosia and Parisian Qwant announced a partnership to create a European search index focused on improving language results.
French telecom operator Orange is exploring collaborations to develop a smartphone-based “sovereign AI” model that reflects local languages.