Is Moore’s Law Really Dead This Time? Nvidia seems to think so

Zoom in / The green lines make it faster.

When Nvidia rolled out its new RTX 40 series graphics cards earlier this week, many gamers and industry watchers were a little shocked at the asking prices the company was mounting on its latest top-of-the-line hardware. New heights of raw power have also arrived with new heights when it comes to the recommended price, which falls in the $ 899 to $ 1,599 range for the 40 series cards.
When asked about those price hikes, Nvidia CEO Jensen Huang told the assembled press to, in fact, get used to. “Moore’s Law is dead,” Huang said during a question and answer session, Digital Trends reported. “A 12-inch wafer is much more expensive today. The idea that the price of the chip will drop is a thing of the past.”

Sorry, how much does it cost?

In justifying Nvidia’s price hikes, Huang defended the raw power of the 40-series cards over past offerings. “The performance of Nvidia’s $ 899 GPU or the $ 1,599 GPU from a year ago, two years ago, at the same price, our performance with Ada Lovelace is enormously better,” he said. “Better off the charts”.

But those price comparisons are a bit misleading. True, the $ 1,599 flagship RTX 4090 is just $ 100 more than the $ 1,499 introductory price for the RTX 3090 in 2020. Adjusted for inflation (which has been unusually high in recent years), the The launch RTX 3090 would cost more than $ 1,700 in August 2022, making today’s RTX 4090 seem like a relative bargain.

When you look at the lower levels, the price increases start to look a little less reasonable. The RTX 4080 line, split into price tiers of $ 899 and $ 1,199, is significantly more expensive than the introductory pricing of $ 699 and $ 799 for the two tiers of RTX 3080 cards in 2020. Even taking inflation into account, those cards would have launched at around $ 800 and $ 900 in dollars today, well below what Nvidia is now asking for its RTX 4080 cards.

Furthermore, the 40 series so far has nothing that even comes close to the introductory price of the cheapest 30 series option offered by Nvidia in 2020: the RTX 3070, which cost $ 499 at the time (about $ 567 in today’s dollars. ). And Nvidia’s lower-end 4080 card is rumored to be on a less-performing chipset than its more expensive cousin, making it technologically more similar to the last-generation 3070.

Nvidia says the high cost of its new Ada Lovelace architecture comes with a lot of additional processing power.
Zoom in / Nvidia says the high cost of its new Ada Lovelace architecture comes with a lot of additional processing power.

Sure, it could be argued that the RTX 30 series cards were priced too low – those cards were exceptionally hard to find at the time of their MSRP launch at the time, largely due to the extreme interest of cryptocurrency miners. But the days of widespread GPU shortage are long gone, as the collapse in cryptocurrency prices and changes in the GPU cryptocurrency mining landscape have led to second-hand GPU prices falling steadily for months. Nvidia explicitly noted this trend in August, warning investors that the effects of the cryptocurrency market “could reduce the demand for our new GPUs.”

If we can’t do it, can anyone else?

Generational price comparisons aside, Huang’s general claim that “Moore’s Law is dead” is a bit shocking for a company whose daily bread has released graphics cards that roughly double in one year each year. comparable processing power. But the prediction is far from new, both for Huang, who said the same thing in 2019 and 2017, and for the industry at large, the International Technology Roadmap for Semiconductors formally announced it would stop chasing the benchmark. in the 2016 roadmap for development chips.
You could write a whole book about the implications of Moore’s law (and in practice Ars did), but the fundamental claim made by Intel co-founder Gordon Moore in 1965 remained extraordinarily prescient: the number of transistors per chip (keeping size / price relatively constant) roughly doubles every year (in a later update, Moore would change the doubling time to 24 months).
As Kevin Kelly explained in a 2009 article, however, Moore’s law is best understood not as a law of physics but as a law of economics and business motivation. Processing power keeps doubling in part because consumers expect it to keep doubling and finding uses for that extra power.

AMD's upcoming RDNA 3 chipset could increase the pressure to keep up with Moore's law.
Zoom in / AMD’s upcoming RDNA 3 chipset could increase the pressure to keep up with Moore’s law.

AMD

That consumer demand, in turn, pushes companies to find new ways to keep up with expectations. In the recent past, that market push has led to innovations such as tri-gate 3D transistors and manufacturing process improvements that continually shrink the size of individual transistors, which IBM can now push to just 2nm.

The point here is that Huang’s alleged “death of Moore’s law” is not entirely up to Nvidia. While Nvidia is no longer able to keep its processor power increases in line with constant prices, they’re not the only game in town. AMD, for example, is already anticipating that its soon-to-be-announced RDNA 3 cards could sport bigger-than-expected improvements in efficiency and overall processing power, thanks to some new chiplet-based designs.

While it’s too early to tell how AMD’s and Nvidia’s new chips will compare, this is the kind of market competition that has traditionally kept hardware manufacturers from becoming too complacent in pushing new frontiers of relative hardware power (see also : Apple Silicon versus older Intel-based Macintoshes). In other words, even if Nvidia can’t figure out how to keep up with Moore’s Law these days, someone else might.

Leave a Reply

%d bloggers like this: