- Global data centre spend reached $57bn in 2024.
- New tools like 800G optics and AI traffic control aim to cut latency.
What happened: Decentralised growth creates new challenges
Global data centre investment hit $57bn last year. AI, cloud and edge services are pushing capacity to new limits. Ciaran Delaney, COO at EXA Infrastructure, said the old model of central hubs is fading. Operators now deploy sites in more regional locations. That adds pressure to link them with speed and resilience.
Latency is a key barrier. Data travelling between edge and core sites faces longer routes and congestion. Power availability also shapes where centres are built, raising risks that grids may strain under future loads. Rules like Europe’s GDPR add further limits, with strict data localisation requirements.
Operators are responding with dense fibre builds, managed optical fibre networks and redundancy models such as quadversity, which splits traffic across four paths. Open, software-based systems promoted by the Telecom Infra Project help avoid vendor lock-in.
Also Read: Lambda, EdgeConneX build AI data centres in US
Also Read: Blackstone boosts Aligned data centres with $1bn credit
Why it’s important
Connecting data centres is now as vital as expanding them. Without robust interconnects, users lose the benefits of decentralised processing. Next steps include 800G optical gear, which doubles throughput, and hollow-core fibre, which can cut latency by 40%. Multicore fibre adds bandwidth while saving space.
Edge and core integration remains key. Edge sites serve low-latency needs, while core hubs process data-heavy tasks. Aligning them boosts efficiency. AI will also play a dual role. It drives demand yet helps operators manage networks in real time, balance loads and spot faults early. As traffic grows, scalable and resilient interconnects will decide how well the digital economy keeps pace.
