Building the Token-Efficient AI Factory

At the AIDC Workshop 2026 organised by IMDA and SGTech, Tho Lye Yit, VP, Design and Delivery at Bridge Data Centres, joined fellow industry leaders to discuss how data centre infrastructure must evolve to support the next generation of AI workloads.

AI Performance Is Redefining Infrastructure Priorities

As AI performance is increasingly measured by metrics such as tokens per watt and cost per token, infrastructure considerations must extend beyond traditional efficiency measures like PUE. Supporting AI at scale requires high-density power, advanced liquid cooling, dynamic GPU loads and long-term operational resilience.

Enabling AI at Scale

The rapid growth of AI is driving new demands across the data centre ecosystem. From power delivery and cooling to operational reliability, every layer of infrastructure plays a critical role in sustaining AI performance and efficiency.

As AI models continue to grow in complexity and scale, infrastructure must be designed to support increasingly dense and dynamic computing environments while maintaining reliability and operational continuity.

Building the Foundation for Next-Generation AI

The AI factory is only as efficient as the infrastructure behind it. At Bridge Data Centres, we are designing for the realities of AI at scale, where power, cooling and resilience directly influence performance and cost efficiency.

By focusing on these fundamentals, we are creating data centre environments capable of supporting the evolving requirements of next-generation AI workloads.