NVIDIANVDA+10.7% since this episode
Sacks is skeptical that the efficiency paper translates to reduced demand, arguing AI energy and data center needs will keep growing — especially with the coming robot revolution — and prefers product launches over academic papers as signals.
I think we're going to need a lot more power, a lot more electricity. I think that's pretty well known. We haven't even gotten to the robot revolution yet. That's coming in the next 5 years. That's going to be energy intensive.” ⚑
Friedberg highlights a new GPU architecture paper showing 70,000x energy reduction vs. H100, suggesting AI inference could shift to edge devices and dramatically reduce data center and energy demand — potentially threatening NVIDIA's current dominance in AI infrastructure.
if this thing's even close to being correct, then you could run the most kind of sophisticated LLMs in a robot. Without it needing to be run out of a data center going forward.” ⚑