Gamma Technologies Company Overview

Aren't you designing the cooling and power of AI data centers separately? With the proliferation of generative AI, data centers are experiencing increased GPU density and power demand. In particular, in AI clusters where GPU loads can change dramatically in milliseconds, there are significant impacts on both the cooling systems and power systems. Such transient phenomena are not adequately captured by designs based on traditional steady-state assumptions, making it challenging to achieve both stable operation and PUE targets. In this webinar, we will introduce a method using Gamma Technologies' GT-SUITE to couple everything from chip and cold plates to CDU, chillers, UPS/BESS, and power systems into a single model. We will analyze how cooling, control, and electrical transients interact during sudden load increases and decreases, and explain design approaches aimed at achieving both stable operation and energy efficiency.

On December 2nd (Wednesday), we will hold a webinar on coupled simulations from chip to grid aimed at achieving PUE targets during sudden GPU load changes in data centers. We look forward to the participation of those responsible for the design and operation of AI data centers, cooling and thermal design, electrical and power systems, control, and facility planning, as well as equipment suppliers for CDU, chillers, UPS/BESS, and others.