AI and the Energy Catastrophe – What is it and what can we do about it?
Andrew Walker, Vice President of Product, Spin Memory
The rise of AI has emphasized the importance of solid-state memory as a brake in system performance and power dissipation. This has led to a plethora of AI accelerator startups and new memory technologies. A growing concern is energy efficiency with the realization that a large part of energy conversion into heat takes place because of data transfers to and from memory. Projections of AI implementation suggest that AI datacenters could use more than 10% of the world’s energy generating capacity by 2025. In addition, IoT applications could demand more than the current US generating capacity. We need to address this problem.
With more than 30 years experience, Walker is widely known as a proven leader in the semiconductor industry, providing expert guidance specifically in the 3-D Flash memory technology segment. Walker has been particularly focused on 3-D Flash memory since 2000, and founded Schiltron Corporation to investigate and develop new 3-D memory technology. He has worked for several major semiconductor companies including Cypress Semiconductor, Artisan Components and Matrix Semiconductor. Prior to his work in Silicon Valley, Walker was with Philips Research Laboratories in Eindhoven, The Netherlands from 1985 to 1994 where he worked on CMOS and non-volatile memory research and development. He earned a Bachelor of Science degree in physics from the University of Dundee, Scotland and a Ph.D from the Technical University of Eindhoven, The Netherlands. He holds over 40 US patents, and is the author and co-author of more than 30 academic articles.