The massive and growing energy consumption of AI data centers is unsustainable and threatens to bottleneck industry growth.
The existing US power grid infrastructure is insufficient to meet the projected energy demands of AI over the next decade.
Scaling analog computing has historically been a major technical challenge due to manufacturing variability.
Opportunities Identified
Developing a highly energy-efficient analog computing substrate could fundamentally disrupt the AI hardware market.
A new computing paradigm based on dynamic systems may enable breakthroughs in AI, potentially leading to models with a better understanding of causality.
The clear and present energy crisis creates a strong market pull for radical hardware innovation.