AI-Driven Demand Reshapes Networking Landscape: Investment, Innovation, and Infrastructure Debt Loom Large
Meta Description: Discover how the surging demand for AI is forcing a rapid evolution in networking infrastructure, with companies like Cisco, Nokia, and Lumen Technologies leading the charge.
The rise of artificial intelligence is creating a new certainty: massively increased workloads on network infrastructures. Enterprises and connectivity providers are already experiencing an unprecedented surge in network demand fueled by AI, and keeping pace with its continued growth will require significant investment in new long-haul networks capable of rapidly scaling capacity – particularly within datacentres.
The emergence of agentic AI-enabled applications is fundamentally reshaping datacentre requirements, prompting a rapid evolution in networking solutions. AI isn’t just using networks; it’s driving advancements, presenting a dual opportunity for both network innovation and operational transformation.
Research conducted earlier in 2025 highlighted the critical role of core fibre investment in supporting future AI growth. However, a significant obstacle remains: four-fifths of firms are delaying network builds due to existing infrastructure constraints. This bottleneck is becoming increasingly acute, with Zayo predicting AI-driven datacentre capacity to grow by 2-6x over the next five years, and Ciena reporting that AI capacity is doubling every six months, requiring the activation of hundreds of additional fibres connected to datacentres.
Looking ahead, Cisco identified two major forces reshaping the networking landscape as of October 2025: the increasing demands of AI agents – raising the bar for scale, security, and governance – and the growing threat of “AI infrastructure debt,” described as early warning signs of hidden bottlenecks that could erode long-term value.
Addressing these challenges, a leading research firm, Omdia, assessed in April 2025 that significant network evolution is necessary to deliver enhanced capabilities and drive continued growth in the global AI economy. The firm emphasized the need for new, advanced optical networks to meet advanced application and service requirements while adhering to tight capital expenditure (capex) targets. These networks, including all-photonic networks, also offer the benefit of lower power consumption per bit, supporting sustainability goals and reducing energy costs.
Crucially, realizing the full benefits of AI in networking requires a holistic approach – considering “networking for AI” alongside AI’s impact on networking.
Several key players are responding to this evolving landscape. Nokia is expanding and enhancing its datacentre networking portfolio to meet the increasing performance and scalability demands of AI workloads, leveraging AI itself to drive efficiency and reliability in its operations.
The pressure on network infrastructure is particularly acute given the energy demands of datacentres, the cloud, and graphics processing units (GPUs). However, experts now recognize that the network infrastructure – including routing, interconnects, and protocols – is becoming the primary bottleneck as AI workloads increase, due to concerns around heat output, cost, and energy usage. AI workloads, unlike traditional traffic patterns, require high-bandwidth, persistent east-to-west traffic, raising the question of whether European infrastructure companies can sustainably scale AI operations.
Cisco is actively modernizing its campus, branch, and industrial networks for the AI era, focusing on systems that simplify operations, enhance security, and scale for evolving business needs. These upgrades build upon the company’s launch of an AI-ready secure network architecture earlier in 2025, designed for automated deployment and security across distributed networks, meeting the high-bandwidth and low-latency demands of AI workloads at the edge.
Research indicates that companies proactively embracing network AI pilots are significantly more likely to move them into production and realize measurable value. The Cisco AI Readiness Index 2025 revealed that only 13% of businesses are fully prepared for AI, yet those that are are four times more likely to move pilots into production and 50% more likely to see tangible results. This underscores the importance of foresight and a strong foundational network.
A recent study from IDC emphasized that legacy infrastructure is insufficient to support the scale and complexity of current and future AI workloads, making network modernization crucial for AI success. Over 78% of companies now view networking capabilities as important or very important when selecting providers for GenAI infrastructure.
However, challenges remain. Research from RtBrick warned that network operators risk being overwhelmed by the combined demands of AI and streaming services within the next five years. The study identified issues not only with technology but also with people and processes, noting that consumer expectations are rising faster than networks can adapt, and that staff shortages are hindering network modernization efforts.
Despite these hurdles, the potential of AI in networking is undeniable. Experts agree that virtually any network can be improved by AI, leading to operational efficiencies, increased reliability, and enhanced user benefits. However, simply “switching on” AI is not enough; a strategic and well-planned implementation is essential.
Lumen Technologies is accelerating its network expansion to meet the growing demands of AI workloads, adding fibre miles and increasing network capacity across the US. The company aims to deliver the high-performance pipeline required to handle large volumes of data processing, effectively building the backbone for the AI economy.
Cisco has also unveiled simplification of its network operations, claiming to deliver exponential performance with next-generation devices while integrating security and enabling new business workflows. A recent IT networking leader survey cited by Cisco found that 97% of businesses believe network upgrades are necessary for successful AI and IoT initiatives. The company cautioned that the future will be defined by those who adeptly leverage AI and those who struggle to do so.
Finally, while the allure of AI is strong, research commissioned by Expereo revealed significant roadblocks to UK AI plans, including poor infrastructure, employee resistance, and unrealistic expectations. However, 88% of UK business leaders surveyed believe AI will be important to fulfilling business priorities within the next 12 months, suggesting a continued commitment to AI adoption despite the challenges.
