Jensen Huang says Nvidia’s AI chips are outpacing Moore’s Law : US Pioneer Global VC DIFCHQ SFO Singapore – Riyadh Swiss Our Mind

Claims the company’s GPUs have seen a 1,000x improvement over the last decade

Nvidia CEO Jensen Huang has said improvements to the chip giant’s hardware are outpacing Moore’s Law.

In an interview with TechCrunch after his keynote speech at the Consumer Electronics Show (CES) in Las Vegas, Huang said “Our systems are progressing way faster than Moore’s Law.”

Nvidia CEO Jensen Huang with a DGX-2

Nvidia CEO, Jensen Huang– Sebastian Moss

He added: “We can build the architecture, the chip, the system, the libraries, and the algorithms all at the same time,” he told the news outlet. “If you do that, then you can move faster than Moore’s Law because you can innovate across the entire stack.”

Moore’s Law is the theory that the number of transistors on computer chips would double every two years. The term was coined by Intel co-founder Gordon Moore in 1965.

The death of Moore’s Law has been a much-debated topic in recent years, with chipmakers already pushing the limits of how small they can continue to make semiconductors. Huang himself declared Moore’s Law dead in 2022. He said: “The ability for Moore’s Law to deliver twice the performance at the same cost, or at the same performance, half the cost, every year and a half, is over.”

Whilst he seemingly did not repeat this claim to TechCrunch, the implication from Huang appears to be that we have moved beyond Moore’s Law, saying that where it had helped to drive down computing costs in the past, “the same thing is going to happen with inference where we drive up the performance, and as a result, the cost of inference is going to be less.”

Huang, who coined Huang’s Law in 2018, said that when it comes to running AI inference workloads, Nvidia’s latest data center superchip is more than 30x faster than its predecessor.

Huang’s Law refers to the Nvidia CEO’s theory that advancements in GPUs are growing much faster than that of traditional CPUs, an observation he made at Nvidia’s 2018 GTC conference where he said the company’s GPUs were “25 times faster than five years ago” and 20 times faster than comparable CPUs. Under Moore’s Law, the GPUs would have only had an expected 10x performance increase.

In his interview with TechCrunch, he doubled down on his claim from November 2024 that the AI industry is experiencing “hyper Moore’s Law.” According to Huang, Nvidia’s AI chips have seen a 1,000x improvement when compared to those being produced by the company a decade ago – although no metric for this claim was given.

https://www.datacenterdynamics.com/en/news/jensen-huang-says-nvidias-ai-chips-are-outpacing-moores-law/