Much of the discussion around artificial intelligence focuses on GPUs, data centres, and electricity demand.
But the real foundation of the AI economy is less visible.
Optical fibre.
Without massive expansion in fibre networks, metro, country, intercity, and subsea, the AI economy simply cannot function at scale.
AI Is Driving an Explosion in Data Movement
AI workloads generate and move far more data than traditional cloud services.
Large language models, training clusters, and distributed GPU environments constantly exchange enormous datasets between data centres.
Global data creation is expected to double within five years, with AI contributing significantly to this growth.
At the same time, global internet traffic continues to expand by roughly 25% per year, reflecting sustained exponential demand for network capacity.
However, while end-user traffic growth is significant, one of the largest emerging drivers is east-west traffic between data centres, the backbone traffic moving training data, models, and inference workloads between AI clusters.
Fibre Capacity Must Scale with AI Infrastructure
AI-driven data centres require extremely high-bandwidth optical interconnects.
Modern hyperscale facilities increasingly rely on coherent optical links operating at 400–800 Gbps, with multi-terabit aggregate throughput across fibre routes.
Research suggests future AI systems may require multi-terabit-per-second transmission per link to support distributed compute clusters.
This demand is driving rapid expansion across the fibre ecosystem:
The data centre interconnect market is projected to grow from $15.4B in 2025 to nearly $26B by 2030.
The global fibre-optic cable market is expected to grow more than 10% annually over the next decade.
From my own experience working on optical fibre build projects, demand for low loss, large core high fibre count cables has been increasing steadily, placing additional pressure on manufacturing capacity.
Demand from hyperscale AI data centres is already contributing to supply constraints in certain optical components.
Australia’s Opportunity
For Australia, fibre infrastructure is a strategic enabler of the AI economy.
International subsea cables, long haul terrestrial fibre routes, and metro interconnection networks will determine where hyperscale AI infrastructure can realistically be built.
Projects currently under development, including new subsea systems linking Australia directly to North America, aim to provide the bandwidth required for the next generation of AI data centres.
These emerging digital highways will support the growing data flows generated by cloud platforms, AI training clusters, and edge computing environments.
The Strategic Question
If AI becomes the defining infrastructure of the next decade, then fibre becomes one of its core transport systems.
Electricity powers AI.
Compute performs AI.
But fibre moves AI.
Countries that invest early in high-capacity fibre corridors, subsea connectivity, and data-centre interconnection networks will be the ones that attract the next generation of AI infrastructure.
The question is no longer whether fibre capacity must grow.
It is how quickly we can build it.
Footnote
This article is part of a series exploring the physical infrastructure behind the AI economy.
Disclaimer
The views expressed in this article are my own and are intended for general information and discussion purposes only. They do not represent the views of any employer, organisation, or client.
© 2026 Rodney Terry – Digital Backbone. All rights reserved.