The New Infrastructure Era: Intelligent Networks for an Intelligent World
On day two of the 19th edition of the Telecom Review Leaders’ Summit, held under the theme ‘Tech Intelligence Beyond Mobility’ from 10–11 December 2025 in Dubai, UAE, the ‘New Infrastructure Era’ panel brought together senior voices from across the telecom and technology ecosystem to examine how artificial intelligence is reshaping networks, infrastructure strategy, and value creation.
The panel featured Ashley Woodbridge, CTO META at Lenovo; Dr. Mahmoud Sherif, Head of Innovation & Technico Business Development at du; Raed Aoude, Regional Director for MEA & CIS at ANDREW; Sabah AlKubaisy, Associate Partner at Salience Consulting; Stelios Savvides, CTO at Vodafone Ventures; and Roque Lozano, SVP Network Infrastructure for MEA at Nokia. Moderated by Jessica Bayley, Assistant Content Director, Telecom Review Group, the discussion reflected an industry navigating both unprecedented opportunity and structural disruption.
As the AI era places growing pressure on legacy networks, the panelists agreed that a new infrastructure paradigm is emerging, defined by flexibility, modularity, and proximity to compute. Existing networks, originally designed for predictable voice and data traffic, are ill-suited to AI’s highly dynamic and unpredictable workloads, which introduce new requirements around logging, latency, and massive, distributed compute. As a result, GPUs and accelerated processing are moving beyond centralized data centers and closer to the edge, driving a convergence of physical and digital infrastructure.
This shift will be complementary rather than disruptive, enabling operators to dynamically scale coverage and capacity based on real-time use cases and workloads, while accelerating automation across applications ranging from core network operations to call centers. Ultimately, AI is demanding new components, processes, and skills to deliver ultra-low latency, edge-proximate compute, greater efficiency, and significantly improved customer experience.
From Predictable Traffic to Unpredictable Demand
Sabah AlKubaisy argued that the push toward a “new infrastructure era” is being driven by fundamentally different AI use cases, particularly the rise of generative and agentic AI that is embedded “within the network itself,” enabling capabilities such as “self-healing” and “self-optimized” networks rather than AI being confined to employee- or consumer-facing applications. Operators must work not only with technology vendors, but also with governments and regulators to align innovation with policy and long-term investment goals.
“This is where a lot of partnerships are needed,” he said, noting that innovation today is as much about business models and ecosystems as it is about technology. When asked what part of the infrastructure lifecycle is experiencing the biggest transformation due to AI, Dr. Sherif pointed clearly to operations as the area experiencing the most immediate transformation. “Operations is the highest priority,” he said. “The reason being that it fits AI models very well.”
Operations is followed by service and customer-facing functions, where AI virtual assistants in call centers are already proving “quite effective.” Planning and capacity design come next; while this area remains “a little bit tricky” and “more artistic” in nature, requiring human judgment, AI is increasingly playing a meaningful role, particularly in capacity forecasting, reinforcing a clear hierarchy of impact from operations, to customer engagement, and finally to planning.
Elaborating on the parts of the network AI is expected to manage completely independently, and agreeing with Dr. Sherif, Stelios Savvides reinforced this operational focus, describing a future where networks respond automatically to faults and performance issues. “You won’t need your operations team to notice a fault,” he said. “It will be done automatically.” By intelligently controlling what runs during outages, extending battery life, and reducing overall consumption, even “1%, 2%, 5% improvements on energy consumption” translate into significant OpEx savings, while simultaneously supporting broader sustainability goals.
Planning and Building Networks with AI
AI is reshaping infrastructure planning, build, and maintenance in two fundamental ways: by automating design and by decentralizing intelligence. According to Ashley Woodbridge, “The networks of the future are going to be completely designed by, or assisted by, AI agents,” enabled by emerging “world models or universe models” that can generate physics-accurate environments from a simple text prompt.
Woodbridbge identified that computer vision is emerging as the “last untapped resource” via cameras in mobile devices and the field and processing data centrally is proving inefficient, accelerating a move toward edge GPUs, mobile compute, and metadata generation on site.
For Aoude, hybrid infrastructure should indeed form the foundation of future networks, as “there is no one access technology spanning across in an optimal way” for indoor, outdoor, and mixed environments.
In this new era, “the next wave of networks will not be about the access technology itself,” said Aoude; rather it will be about how intelligently and invisibly these layers are orchestrated in the background to enable autonomy across environments.
The Future of Public-Private Network Synergy
Reflecting on Vodafone’s collaboration with Iraq’s Ministry of Communications on a new 5G network, Savvides emphasized that public–private collaboration can be a powerful catalyst for national network transformation.
Drawing on personal experience from startup initiatives in Qatar and Oman, he noted that the public sector often brings “a vision” and “the seeds of opportunity in the market,” but success depends on partnering with organizations that have the expertise and execution capability to turn that vision into reality. A critical lesson, he said, is assembling “the right team,” combining strong business development and engagement skills with deep technical and strategic know-how to build a viable plan and business case together.
Lozano concurred, noting that “no one can consider the future a sustainable future without meaningful digitalization of all the sectors in the world.” He highlighted that private companies bring speed and innovation, deploying high-speed cables, new digital models, and AI-enabled capabilities, while governments enable scale and direction.
Building on this, AlKubaisy stressed that public–private partnerships must align with policy while helping operators avoid repeating past mistakes, particularly those made during the rise of OTTs, by focusing not only on technical innovation but also on commercial and business model innovation.
The Role and Impact of Agentic AI in the 6G Era
As the industry moves toward 6G, AI agents are set to fundamentally transform network operations by shifting intelligence, decision-making, and monetization to the forefront of network design.
Dr. Sherif noted a clear change in the 6G conversation, observing that recent 3GPP contributions are no longer focused solely on technical specifications. “For the first time, it’s not only about the technology specs of the network. There is also a whole section and contributions about how 6G will be monetized,” he noted.
This focus on monetization explains why edge computing is becoming so central to 6G architectures. In this future, each customer may be supported by a personal AI agent, interacting in natural language to dynamically optimize performance and outcomes, while on the operator side, “the head of operations” could be supported by multiple AI agents managing networks autonomously.
Addressing the risks in AI-based networks, Lozano underscored that technology inherently presents both opportunity and threat, noting that, “As this technology becomes more sophisticated, so too will the risk, and so too will the defense.” He stressed that these risks are not confined to telecommunications, but represent “an entire global issue” with implications for productivity, enterprise services, data, and, ultimately, people’s lives.
Sustainability Under Pressure
Sustainability emerged as both a responsibility and a challenge, with Aoude highlighting multiple layers of impact, starting with material efficiency. He advocated for “less waste, less plastic, and the optimization of copper” to be applied not only to network components like antennas and cables but also to robots and other infrastructure.
A key focus for ANDREW is designing for longevity, with sites intended to last five, 10, or even 15 years, minimizing the environmental burden of frequent installations and expansions, an approach he illustrated by noting a tower in Kuwait which, after nearly a decade, still stood strong. However, he also cautioned, “We’re starting from a much higher benchmark, so we’re adding around 50% more power requirements, and we’re optimizing by 5–10%.” The challenge, he concluded, is balancing legacy network concepts with these new sustainability imperatives, aiming for “resident energy” efficiency over the next two-to-three years while supporting AI-driven growth.
Woodbridge emphasized that “telecom has always been quite challenging to decarbonize,” and added that AI-driven workloads are pushing energy requirements to unprecedented levels, with racks now consuming 110 kilowatts and campuses approaching gigawatt-scale, far beyond traditional data centers.
Addressing this requires innovation across the hardware stack, including strategies like liquid cooling, which can cut cooling costs by 40% by leveraging water’s thermal dynamics.
“It’s not just only about achieving the fastest sustainability gains; it’s more about long-term sustainable investments,” reiterated AlKubaisy.
As the discussion drew to a close, the panel converged on a shared understanding: the role of the network is fundamentally changing. What was once designed to move traffic efficiently must now support intelligence, autonomy, and innovation at scale.
link
