LIVE WEBINAR 11 NOV - Fibre at AI scale: rethinking connectors and cable management for high-density data centres

As hyperscale data centres race to deploy AI clusters with tens of thousands of GPUs, a critical infrastructure question has moved from the roadmap to the urgent present: how do you fit enough fibre into the same physical space whilst keeping it manageable at scale?

This webcast will  examine this challenge directly with operators, infrastructure architects, and connector manufacturers who are grappling with it now.

The density problem

AI clusters need far more fibre in the same space than traditional data centre networks. Every GPU requires multiple optical connections. Every connection requires a connector. Every connector requires cleaning, inspection, and management throughout its lifecycle. The connector formats available: MTP/MPO, CS, SN; each have different trade-offs in density, cost, reliability, and maintenance burden.

Yet there has been surprisingly little industry conversation about which connector strategy actually works at AI scale, and what the real operational costs are once the infrastructure is live.

The webcast brings together operators, infrastructure specialists, and connector suppliers to explore the practical decisions driving deployment today.

What you'll learn

The session will cover:

• Connector selection: How to choose between MTP/MPO, CS, and SN connectors for each network layer, and where each makes sense
• Operational reality: The real cost of higher density—cleaning, inspection, polarity errors, and installation time
• Infrastructure strategies: Practical ways to fit more fibre into existing racks and pathways without a complete forklift upgrade
• Lessons from the field: What operators are learning from early AI deployments about what scales and what does not
• Supplier engagement: What to ask connector and cable manufacturers about loss budgets, lead times, and standards support

The discussion will address the full spectrum of participants in fibre infrastructure: data centre operations leaders planning fibre for GPU cluster build-outs, network architects evaluating connectivity strategies, structured cabling specialists managing installation and deployment, and procurement teams weighing cost, reliability, and supply chain risk.

Featured speakers

The webcast features industry experts from leading fibre optic and testing companies, moderated by Fibre Systems. Speakers include representatives from AFL, US Conec and VIAVI Solutions, bringing operational experience from hyperscale deployments and testing validation perspectives.

Why this matters now

The shift from traditional data centre connectivity to AI-scale fibre infrastructure is creating a genuine skills and specification gap. Many operators, architects, and even some manufacturers are still operating with assumptions built around traditional networking densities and connector management practices.

AI clusters are operating at 10x to 100x higher fibre density than legacy systems. That scale change fundamentally alters which connector formats are viable, which management practices work, and what total cost of ownership actually looks like.

This webcast is designed to close that gap, bringing together the people making these decisions now and the suppliers building the solutions.