Gelsinger said that Moore’s Law works for Intel, not like Huang and NVIDIA

Two of America’s most important semiconductor designers disagreed on Moore’s Law. Intel CEO Pat Gelsinger said at yesterday’s eventthat the half-century-old rule of thumb is alive and well, and NVIDIA co-founder and CEO Jensen Huang last week reportedthat it no longer works.

Image source: intel.com

In the mid-sixties, co-founder and former head of Intel, Gordon Moore, formulated, and a decade later corrected the rule that the number of transistors on a chip doubles every two years, increasing its performance. To increase the number of transistors on a chip, it is necessary to make them smaller, and this requires intensive development of manufacturing technologies. For many years, Intel has been a leader in manufacturing technologies and its processors have the highest density of transistors in the world, but in recent years TSMC and Samsung have taken the initiative, which today work at 5nm standards, while Intel is still working with technologies 7 and 10 nm.

One of Intel’s key goals under Gelsinger is to return to world leadership in chip manufacturing, and for this the company will have to take five consecutive steps in improving technology standards in four years, although each of them would normally take two years. For this, new lithography technologies are being actively developed, the RibbonFET architecture is being introduced – they will help to further compact transistors on a chip, and technical processes will soon begin to be denoted not by nanometers, but by their tenths – angstroms.

  Image source: intel.com

NVIDIA chips are produced by TSMC, which today has the most advanced manufacturing technologies and is the world’s largest chip manufacturer – NVIDIA is left with only development. Instead of Moore’s Law, the company’s CEO Jensen Huang talks about the strategy of “accelerated computing”: he considers artificial intelligence technologies to be the most demanded tasks today, which NVIDIA GPUs can handle more effectively than others. But in fact, the head of NVIDIA justified the increase in prices for new generation video cards with the “dead Moore’s Law”.

It looks like the market is on NVIDIA’s side: over the past five years, Intel shares have fallen in price by 28%, while NVIDIA securities, by contrast, have risen in price by more than 180% – even taking into account their fall by 58% in 2022. The market capitalization of NVIDIA is almost three times that of Intel.

If you notice an error, select it with the mouse and press CTRL + ENTER.

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.