Altnets has published a white paper on the infrastructure demands created by artificial intelligence, arguing that connectivity is becoming a key constraint alongside computing power.

Rapid growth in AI use, data centre construction and digital traffic is increasing pressure on fibre networks, backhaul links and supply chains, according to the report. It frames the shift as a practical issue for integrators that design, source and deploy communications infrastructure.

Among the figures cited, global active data centre capacity is forecast to rise from 24.4GW in 2025 to 147.1GW in 2035, based on ABI Research data. The paper also cites JLL estimates that AI workloads could account for about half of total global data centre capacity by the end of the decade.

Mobile traffic is another part of the picture. Citing the Ericsson Mobility Report, the white paper says global mobile data traffic is expected to more than double by 2031 to about 310 exabytes a month, rising to 482 exabytes a month when fixed wireless access traffic is included.

The UK energy system also features heavily in the analysis. Proposed AI-related data centre projects currently seeking grid connections in the UK could require around 50GW of electricity capacity, the report says, exceeding Great Britain’s current peak demand.

Infrastructure strain

This combination of higher computing demand, rising traffic volumes and greater electricity needs is shifting attention beyond chips and servers. Future growth will depend on the availability of fibre routes, optical connectivity, backhaul architecture and supply chains able to support larger, more complex deployment programmes, the paper argues.

Altnets describes the publication as the second paper in a wider series on changes in the connectivity market. The first examined fibre shortages and ways to reduce supply disruption; the new edition focuses on AI as a driver of network demand and infrastructure planning.

The argument comes as governments and businesses place greater weight on digital infrastructure as a factor in economic competitiveness. Data centre growth has already prompted debate in the UK and elsewhere over land use, power supply, planning delays and the role of telecoms networks in supporting increasingly data-heavy applications.

For integrators, the challenge is widening from component procurement to designing networks that can scale over the longer term, the report suggests. Organisations will need to focus on resilience and deployment planning as demand for interconnection and transport capacity rises.

A standalone line in the document sums up the company’s position: “The AI boom is not just a computer story; it is a connectivity story,” said Andy Ainsley, Commercial Director, Altnets.

Supply chain focus

The paper also highlights supply chain resilience as a factor in how the market responds. As more infrastructure is needed to connect data centres, mobile networks and fixed networks, the availability of fibre products and related equipment could become more important to delivery times and costs.

Its intended audience includes organisations involved in building and expanding communications networks, from independent internet service providers to large data centre operators. Altnets is based in Brighton and works with European manufacturers on telecoms infrastructure products.

Ainsley’s background spans telecoms and law enforcement. He studied Manufacturing Engineering at Nottingham Trent University, began his telecoms career at AFL, later worked in UK law enforcement, and then returned to the sector in senior sales and business development roles at AFL, Neos Networks and Hexatronic before joining Altnets.

The report reflects a wider shift in industry discussion around AI. Much of the public debate has focused on semiconductors, model training and the energy intensity of large-scale computing, but network operators and infrastructure suppliers have increasingly highlighted the less visible systems needed to carry growing traffic between users, data centres and cloud platforms.

That includes long-haul fibre, metro links, campus interconnects and backhaul networks connecting distributed sites to core infrastructure. If AI applications continue to spread across consumer services, enterprise systems and mobile networks, demand on those links is likely to rise with them.

In the paper’s closing argument, Ainsley links current investment decisions to longer-term network readiness. “The industry is moving into a new era where network resilience and infrastructure readiness are becoming just as important as capacity itself. Organisations that invest in scalable connectivity and long-term infrastructure strategy today will be better positioned to support the demands of tomorrow’s AI-driven economy,” he said.