Presentation Description: The AI revolution promises to reach into all aspects of our lives, with AI becoming a constant companion in day to day matters both personal and professional. However bringing forwards this sci-fi like vision of the future will be incredibly energy intensive and place demands on power semiconductors not typically seen before.
This presentation will summarise the challenges and opportunities for power semiconductors in this rapidly changing landscape. Data centers are the obvious place to focus. The sheer volume of energy needed to power the AI data centers of the future has led to discussion of them having their own dedicated nuclear reactor on sight. As each new generation of GPU demands more and more power, being able to efficiently get that power from the grip to the chip will be vital in keeping operating costs sensible and demand manageable. This starts with the AC-DC power supply but the challenge is also faced by the on-board bus systems. Finally, as AI shifts to the edge to allow for offline performance and enhanced personal security power semiconductors will be key to maintaining battery life in our devices.
In addition to their semiconductor analyst team, Omdia has analysts specialising in both data centers and consumer device. Their insight will be used in forming the data and discussion points in this presentation, giving attendees an opportunity to understand the view of the end user, which is critical in correctly tackling this tricky opportunity. The presentation will provide forecasts for use of power semiconductors in AI related applications, discuss the competition between silicon carbide and gallium nitride, and suggest the largest barriers to success.
