The FYUZ event in Dublin brought together leaders and innovators to explore the future of network disaggregation. While the core mission remains to create open, competitive ecosystems, the conversation has evolved significantly. This year, the focus sharpened on how disaggregation is not just about reducing costs but also about paving the way for advanced AI integration and new network capabilities. Let’s explore the key takeaways that are shaping the telecom industry’s path forward.
The Core Mission: Disaggregation Across the Network
Disaggregation continues to be a central theme, aiming to break down proprietary, monolithic network structures. The goal is to create standardised reference architectures for the Radio Access Network (RAN), transport, and even local networks like LAN and Wi-Fi. By fostering an open ecosystem, the industry hopes to drive competition, lower costs for operators, and provide greater choice in vendors and technology.
This year’s event, however, hinted at a shifting landscape. The number of exhibitors was noticeably lower, with some prominent companies opting for smaller meeting rooms instead of large booths. This may reflect a reallocation of budgets or a more focused, strategic approach to industry engagement. Despite this, the commitment to open standards remains strong, with significant developments underway.
Open RAN: From Theory to Large-Scale Deployment
A major topic of discussion was the deployment of Open RAN (O-RAN). While operators in Western countries are now deploying O-RAN at scale, a key question lingered: how does it perform against traditional RAN?
Operators confirmed they would not proceed with deployments unless O-RAN offered technical and commercial parity, or better. Yet, none were prepared to share specific performance metrics comparing the two. This suggests that while the business case for O-RAN is solidifying, the industry is still in a crucial phase of proving its real-world performance at scale. A landmark announcement from Deutsche Telekom, revealing a 30,000 RAN site Request for Proposal (RFP) for 2026, signals that major commitments are being made, accelerating the move toward open architectures.
AI: The New Frontier Enabled by Openness
Perhaps the most significant shift in the narrative at FYUZ was the morphing purpose of disaggregation. It is no longer just an end in itself but a critical enabler for Artificial Intelligence. The open and consistent interfaces inherent in disaggregated networks are the foundation upon which advanced AI applications can be built.
One could argue that the concept of “AI RAN” is only possible because of this open architecture. In a closed, proprietary system, AI development would be the exclusive domain of a few large vendors, operating “under the hood” without transparency. Openness democratises this innovation, allowing a wider range of players to develop and deploy AI-driven solutions that can optimise, manage, and secure the network.
Moving AI to the Edge
The conversation also challenged conventional wisdom about data management. The idea of relying on massive, centralised data lakes is becoming outdated. The process of collating and storing vast amounts of network data is slow and resource-intensive, often failing to provide timely insights.
A more effective approach is emerging: applying AI and machine learning models directly at the network edge, close to where the data is generated. This allows for real-time analysis and immediate action, a necessity for dynamic network management. Instead of waiting for data to travel to a central repository, insights are extracted instantly, enabling faster, more efficient operations.
Understanding the AI Landscape
It became clear that “AI” is often used as a catch-all term, causing confusion. Many people associate AI primarily with generative AI, like large language models. However, AI in the form of machine learning (ML) has been used in networks for years to identify patterns, predict failures, and optimise performance.
The consensus is that training massive, public data models on general internet data is not useful for specific telecom needs. These models lack the specialised network data required for meaningful application. The more practical and cost-effective strategy is to develop smaller, tailored AI models trained on specific telco datasets to solve distinct problems, from network security to energy efficiency.
The Emergence of Sensing as a Service
One of the most fascinating discussions at FYUZ centred on the network’s potential to become a sensing platform. Innovators showcased how both fibre optic cables and RAN signals can be used for more than just communication.
By analysing disturbances in these signals, the network can be used to “sense” the physical environment. This includes detecting the movement of objects, monitoring traffic patterns, and even identifying seismic activity. This “sensing as a service” capability opens up a world of new applications and revenue streams for operators, transforming the network from a simple communication conduit into an active participant in the physical world.
The Road Ahead for Telecom
The FYUZ event in Dublin painted a picture of an industry in transition. The foundational push for disaggregation continues, but its purpose is expanding. Open architectures are now seen as the key to unlocking the full potential of AI, enabling smarter, more responsive, and more efficient networks. As operators like Deutsche Telekom make massive investments in O-RAN, and as innovators push the boundaries of what’s possible with AI and sensing, the future of telecommunications looks more open and intelligent than ever before.