The rise of high-powered GPUs has compressed a decade of data center evolution into just a few short years. Standard facilities are rapidly transitioning into high-density AI factories, where individual rack loads are climbing to 100 kW and beyond—demanding a level of thermal extraction equivalent to cooling multiple single-family homes in a single rack footprint.

This sudden shift breaks traditional infrastructure rules. At these unprecedented densities, cooling is no longer a simple perimeter airflow challenge. Operators are now forced to navigate the physical weight of liquid-filled racks, the chemical management of direct-to-chip coolants, highly dynamic GPU power spikes, and intense grid and local water resource constraints.

Successfully scaling your infrastructure requires looking beyond standalone cooling components. It demands a holistic, resource-first strategy that integrates power, cooling, and regional environment constraints.
Whether you are retrofitting legacy data halls to extend their lifespan or mapping out a blueprint for a next-generation gigawatt campus, this session provides a practical roadmap.

Join Meher Nondekulia (Global Business Development Manager at Johnson Controls) and host Ella Wilkinson (BizClick) as they explore how to design resilient, scalable, and community-conscious thermal strategies for the next generation of AI workloads.