Nvidia's Huang Says His AI Chips Are Improving Faster Than Moore's Law - Slashdot
Jan 08, 2025 - tech.slashdot.org
Nvidia's AI chips are advancing at a pace that surpasses Moore's Law, according to CEO Jensen Huang. Moore's Law, which historically predicted a doubling of transistor density annually, has been outpaced by Nvidia's thousand-fold chip improvement over the past decade. Huang attributes this rapid advancement to Nvidia's ability to innovate across the entire stack, including architecture, chips, systems, libraries, and algorithms.
Huang emphasizes that this accelerated progress will lead to reduced costs in computing, similar to the historical impact of Moore's Law. Specifically, he predicts that as performance increases, the cost of inference will decrease, driving further advancements in AI technology.
Key takeaways:
Nvidia's AI chips are advancing faster than Moore's Law, according to CEO Jensen Huang.
Huang claims Nvidia's chips have improved thousand-fold over the past decade, surpassing Moore's Law's predictions.
Innovation across the entire stack, including architecture, chips, systems, libraries, and algorithms, enables faster progress.
As performance increases, the cost of inference is expected to decrease, similar to the historical impact of Moore's Law on computing costs.
