Low Carbon Emission Design Optimising AI Computing
Technology Transfer & Commercialisation

Low Carbon Emission Design Optimising AI Computing

Silver Medal

Low Carbon Emission Design Optimising AI Computing

The digital age is increasingly reliant on artificial intelligence (AI). Robust hardware such as GPUs and FPGAs are required to support the tremendous processing power required by AI to realise its full potential. High-density power converters are needed to produce a large amount of power for AI computing algorithms. Large amounts of electronic waste are generated when faulty hardware is discarded.

ASTRI has invented high stability and thermal performance solutions, the Active Electromagnetic Interference Filter (AEI) and the Thermally Enhanced Interposers (TEI), to enable better thermal management and greatly reduce electronic waste. The AEF technology can reduce Electromagnetic Interference (EMI) noises and obtain a 30% volume reduction in the EMI filter. The TEI technology can enhance the thermal management of the power supply system and achieve a 40% improvement in thermal performance. These will bring down the usage of capacitors, inductors and heatsinks, lessening the space in the GPU board and impacting system.

This technology can also be applied to Electric Vehicles, New Energy, Smart Grids and many other industries.