According to Anagnostopoulos, optimizing edge data centers would allow for a more regionalized computing architecture. In order to increase efficiency at each level, his research would go beyond conventional optimization techniques and focus on "vertical solutions" for particular infrastructure sectors.
"Operational and embodied carbon footprints are impacted by a number of parameters," he stated. By utilizing renewable energy sources, improving power usage efficiency, and intelligently dispersing workload across different computing components, we are attempting to develop solutions to reduce the operating carbon footprint. Downscaling energy-efficient hardware accelerators, which are specialized units made to do certain operations more effectively than a general-purpose central processing unit, is another option to reduce the carbon footprint. These, specifically the subclass of tensor processing units, are essential for speeding up computationally heavy processes and improving the effectiveness of data processing.