“In 10 years, computers will be doing this a million times faster.” The head of Nvidia does not believe that there is a need to invest trillions of dollars in the production of chips for AI

by alex

It’s enough to just continue to develop technology

Despite the fact that Nvidia is now almost the main beneficiary of the growing interest in AI, the head of the company, Jensen Huang, does not believe that additional trillions of dollars need to be invested in the industry. 

If you just assume that computers will never get faster, you might come to the conclusion that we need 14 different planets, three different galaxies, and four more suns to power it all 

Huang commented on the current initiative of OpenAI head Sam Altman, who is now seeking investment of $5-7 trillion to build a large number of new chip factories exclusively for AI accelerators. According to the head of Nvidia, it will be enough to simply continue to develop new architectures and technologies at the same pace as they are, and then the increased capacity will cover the increased needs without the need to exponentially increase the number. 

«Через 10 лет компьютеры будут делать это в миллион раз быстрее». Глава Nvidia не считает, что нужно вкладывать триллионы долларов в производство чипов для ИИ

Haung logically assumes that in the future a gigantic number of new factories could lead to an oversupply of chips and, as a result, a major economic crisis that will affect the whole world. 

READ
From bat drones to a whaler and a rocket scolopendra: “Rostec’s Battle Monsters” can turn into a computer game

«Через 10 лет компьютеры будут делать это в миллион раз быстрее». Глава Nvidia не считает, что нужно вкладывать триллионы долларов в производство чипов для ИИ

The head of Nvidia claims that in four to five years the AI ​​market in the form of data centers alone will reach $2 trillion

Remember that architecture performance will improve, so you can't just rely on buying more computers. You should also assume that computers will become faster and therefore the total amount you will need will not be as large 

For example, we can cite the same Nvidia accelerators. The V100 GPU in 2018 had a performance of only 125 TFLOPS, while the modern H200 delivers almost 2000 TFLOPS (FP16). However, in parallel with this, it is still very difficult to predict what the demand for chips for AI will really be in the next five to ten years. 

One of the greatest contributions we've made has been improving computing and artificial intelligence a millionfold in the last ten years, and so whatever the demand you think will power the world , you have to consider the fact that computers are also going to make it a million times faster in the next ten years

You may also like

Leave a Comment