New compute systems

Artificial Intelligence is expected to propel the semiconductor ecosystem forward in the coming years. As the need for compute for both training and inference keeps growing, so does AI energy usage. By 2030, AI datacenter electricity consumption will have outgrown the yearly production in Arizona tenfold. New Compute Systems will therefore have to solely focus on providing a superior energy efficiency.

With 2D transistor density scaling slowing down, one must resort to Cross Technology Co-Optimization (XTCO). Only a holistic approach through system scaling challenges (Compute Density, Power Delivery, Thermal, Memory Bandwidth) will drive the technology roadmap towards energy efficient Compute scaling.

A CMOS 2.0 system delivers next-level energy efficiency, combining 2D scaling, fine pitch 3D Integration, Heterogeneous compute tiers and a relentless focus on energy efficiency transistor technology.