EU's €8m Shield-6G project targets AI-native network security

The race to define 6G is well underway, and the European Union is making clear it intends to shape not just the performance of the next generation of mobile networks but the security principles baked into their foundations.

University College Dublin (UCD) has been named lead institution for Shield-6G, an €8m project funded under the EU’s Horizon Europe Smart Networks and Services Joint Undertaking. The project was the sole awardee under the highly competitive Horizon Europe call, a signal of both its ambition and the seriousness with which the European Commission is approaching the challenge of securing next-generation connectivity infrastructure.

Shield-6G is tasked by the European Commission with establishing the foundational security, reliability and resilience guidelines for 6G networks. In practical terms, it is an AI-driven cyber threat intelligence platform designed to shift the dial from basic network reliability to what its architects describe as an “unprecedented” standard of systemic resilience.

6G will not simply be a faster version of 5G, it is widely expected to form the connective tissue between cloud infrastructure, autonomous systems, edge compute and critical national infrastructure. Getting the security architecture right before commercial rollouts begin is not an afterthought; it is the entire point.

“6G will be far more than the next step in mobile connectivity; it will form the intelligent digital nervous system of future society, connecting people, industries, critical infrastructure and autonomous systems,” said Madhusanka Liyanage, director of UCD NetsLab and lead of the Shield-6G consortium. “Through Shield-6G, we are working to ensure that security, trust and resilience are embedded at the very foundation of this future.”

What Shield-6G actually does

The proposed architecture combines automated zero-touch security orchestration with privacy-preserving analytics, including federated learning, secure multi-party computation and differential privacy. The ambition is to allow multi-stakeholder networks to share threat intelligence and self-heal against cascading vulnerabilities without ever transmitting raw, sensitive data between parties.

That last point matters enormously in an environment where cross-border data flows face increasing regulatory scrutiny, and where telecom operators are simultaneously competing and needing to cooperate on shared infrastructure threats.

The consortium brings together 19 international partners spanning academic research institutions, multinational industrial companies and highly specialised SMEs.

The wider picture: standards, AI and the regulatory race

Shield-6G does not exist in isolation. The European standards community has been building towards this moment for some time, and the debate about how AI should be embedded into 6G is heating up across the industry.

Earlier this year, Capacity spoke with Ultan Mulligan, chief services officer at the European Telecommunications Standards Institute (ETSI), who made clear that AI will not simply be a feature of 6G but something far more fundamental. “AI is going to have a very significant input, evolving every bit as fast as, if not faster than, 6G networks,” Mulligan told Capacity. “It will very much be a driving force behind 6G.”

On security specifically, Mulligan was unequivocal. “Assurance in the quality and performance of AI systems and security in the data (whether for training, telemetry, digital twins, or similar) is crucial,” he said. ETSI published its first standard on cybersecurity for AI systems at the end of last year, and Mulligan indicated that more would follow as both the threat landscape and the technology itself continued to evolve.

“It’s an absolutely key component that must be addressed from the beginning,” he added.

That sense of urgency runs through the Shield-6G project too. According to Dell’Oro Group research published this month, cumulative 6G RAN revenue and wireless capex during the first six years of the commercial cycle are projected to exceed $100 billion and $500 billion, respectively, figures that signal just how much infrastructure and how much money will ultimately depend on the security architecture being defined right now.

UK dragging its feet on cloud regulation as reliance on hyperscalers deepens

Can Europe’s AI gigafactories close the gap with the US?

EU overhauls EuroHPC to power AI gigafactories and quantum build-out