New technology would boost energy efficiency at data centers

Artificial intelligence (AI) is becoming increasingly prevalent in everyday life, leading to a significant rise in data centers and the challenge of meeting their energy demands.
Associate Professor Pritam Das from Binghamton University has developed innovative technology to improve energy efficiency in AI data centers.
Das’s efforts are supported by the National Science Foundation and have received $100,000 from the University’s EXCEED program backed by an NSF Accelerating Research Translation (ART) grant to prototype and evaluate his inventions.
As traditional methods like Moore’s law reach their limits, AI processor designers are focusing on power efficiency, necessitating solutions to manage heat alongside increased computing capabilities.
Das’s new point-of-load converters offer a single-stage conversion of power with higher efficiency (up to 12% more) and faster conversion speeds compared to existing multi-stage converters.
His technology aims to reduce thermal stress and energy consumption in data centers, making it more feasible to handle the immense computational demands of modern AI systems.

You may have missed