Connectivity and Data Centers: The Backbone of AI and Sustainable Digitalization

The rapid ascent of artificial intelligence (AI) as a transformative technology has brought with it unparalleled demands for connectivity and data infrastructure. Robust digital infrastructure is critical in enabling AI applications and fostering sustainability. In the words of DE-CIX's CEO Ivo Ivanov, “The performance of AI applications... depends on the ability of data centers, networks, and interconnection platforms to work in unison,” ensuring seamless, secure, and resilient data transmission.

AI’s transformative potential hinges on its ability to access and process vast amounts of data efficiently. Modern data centers provide the computing power and storage capacity needed for training and implementing AI models, while connectivity ensures the rapid and reliable flow of information between systems.

The interplay between AI and connectivity

Connectivity can be defined as the ability to establish and maintain reliable, high-performance connections for data transfer between various digital entities like data centers, cloud providers, Edge locations, IoT devices, and end-users. It encompasses the physical infrastructure (cables, routers, switches) and the logical aspects (protocols, routing, security) that enable data to flow seamlessly between these entities.

Digitalization has driven a more than tenfold increase in computing power demand over the past decade. This is matched by advancements in connectivity, such as edge computing and Internet Exchanges (IXs), which reduce latency and ensure high bandwidth for real-time AI applications.

High-bandwidth, low-latency connections are essential for the successful development, deployment, and operation of AI systems. AI models, especially large language models (LLMs), require massive datasets for training and operation. This necessitates high-bandwidth connections to facilitate the efficient transfer of data. During inference – AI in action, as Ivo Ivanov calls it – low-latency connectivity is crucial to ensure rapid data access and responsiveness, particularly for time-sensitive use cases like autonomous vehicles and financial services.

Dr. Or Sela, F5, explains that modern AI deployments often utilize a distributed infrastructure, with components spread across on-premise data centers, edge locations, and public clouds. Robust connectivity solutions are essential for seamlessly connecting these disparate elements, ensuring data flows efficiently between them. This distributed approach also enhances redundancy and resilience, as AI workloads can be shifted between locations in case of disruptions.

Gabriel Willigens, ITENOS, sees increasing demand for robust and resilient connectivity solutions that can support AI workloads across various environments, including on-premise data centers, private clouds, and public clouds. This approach ensures businesses can choose the most suitable infrastructure for their AI applications while maintaining seamless data flows and high performance.

Data sovereignty and resilience

In an interconnected world, data sovereignty and resilience are pivotal. Distributed data infrastructures, bolstered by edge computing, enable data to be processed closer to its source, enhancing both latency and compliance with local regulations. These distributed systems minimize single points of failure, ensuring robust data resilience essential for critical AI applications.

As Andreas Weiss, Managing Director of the eco Association, explains, “Sovereignty, first and foremost, means self-determination. Whether people, companies, or nations, a free Europe is one in which everyone can decide for themselves – not just in matters relating to their own lives, but also regarding information, data, and intellectual value at the core of business models.” According to Juan I. Hahn, Head of eco’s Mobility Competence Group, data sovereignty is the foundation for this trust. 

After all, the key question in digital transformation is: Who controls the data? Resilience, achieved through strategies like redundancy, provides a concrete foundation for this self-determination, especially in the digital sphere.

Energy-efficient data centers: A sustainability driver

This focus on data sovereignty and resilience is particularly important as the rise of AI and digitalization underscores the need for sustainable data infrastructure. Globally, data centers consume around 1% to 2% of total electricity, varying by region. In Ireland, they account for over 20% of national electricity use, expected to grow to 32% by 2026. Energy-efficient data centers have achieved remarkable reductions in energy consumption per workload – 12 times lower in 2020 compared to 2010. Advanced cooling systems, efficient hardware, and AI-driven energy management contribute to these gains. For example, AI can detect operational inefficiencies in data centers and suggest energy-saving adjustments, as highlighted by Volker Ludwig of the Alliance for the Strengthening of Digital Infrastructures in Germany in an interview on the sustainability potential of data centers.

Despite increasing workloads, the global average Power Usage Effectiveness (PUE), a measure of data center efficiency, has remained steady at 1.56 in 2024. However, new facilities using advanced cooling and high-density IT configurations are achieving PUEs as low as 1.3, according to the Uptime Institute Global Data Center Survey 2024.

Innovations like waste heat recovery are turning data centers into contributors to the broader energy ecosystem. By redirecting waste heat to local heating networks, greenhouses, or adjacent buildings, data centers can simultaneously reduce their carbon footprint and support community energy needs.

The role of digital infrastructure in sustainability

The “Digital Transformation for More Sustainability“ study by the eco Association, the eco-founded Alliance for the Strengthening Digital Infrastructures in Germany, and Arthur D. Little shows how digital technologies not only consume energy but also drive efficiencies across sectors. By 2050, consistent digitalization could lead to significant CO₂ savings, with IoT applications alone cutting emissions in rural and industrial sectors by up to 39%. Such advancements align with the goals of the Paris Agreement, underscoring the dual role of digital infrastructures as both enablers of AI and stewards of sustainability.

The path to a sustainable future

The path to a sustainable future lies in the harmonious integration of digital innovation and environmental stewardship. Coordinated efforts between industry stakeholders and policymakers are required to balance the growth of data center capacity with sustainability goals. Efficient authorization procedures, competitive electricity pricing, and robust energy efficiency regulations are necessary to support sustainable growth.

In conclusion, the synergy between connectivity, data centers, and AI is pivotal for driving digital innovation and achieving sustainability targets. These infrastructures form the backbone of a digital future where AI can flourish without compromising environmental goals. With continued advancements in energy efficiency and a commitment to resilient, sovereign data systems, the Internet industry can rise to meet the dual challenges of digitalization and climate change.

For more insights on sustainable digitalization and energy-efficient data centers, keep up with news and events here: https://international.eco.de/topics/datacenter/ and subscribe to the dotmagazine newsletter.

Roland Broch is Senior Project Manager Digital Infrastructures and the contact person for the Data Center Expert Group at eco – Association of the Internet Industry.