Nvidia and artificial intelligence, it’s a real love story. In recent months, Big Tech’s love for this technology has allowed the green giant to move into a whole new dimension, with a peak at more than 3 trillion in market capitalization which testifies to its insolent financial health. Recently, CEO Jensen Huang drove the point home again by stating that AI was now a fundamental component of their business, and they simply couldn’t do without it.
« We can no longer approach the graphic part of computing without artificial intelligence “, he said at the Communacopia + Technology conference organized by Goldman Sachs. We calculate a single pixel, then determine 32 more by inference making predictions from new data using an already trained AI model, Editor’s note), and the whole thing is stable, photorealistic, qualitative, with incredible performances “, the boss exclaims.
He also took the opportunity to tackle a criticism regularly leveled at machine learning, namely energy consumption. According to him, this is a preconceived idea that has no place, at least in the graphics field. He claims that the energy bill would be even higher if each pixel had to be calculated individually without using this technology.
« The amount of energy that you save… calculating a single pixel takes a lot of energy — that’s the nature of computing. But calculating a single pixel and then inferring 32 more pixels takes very little energy by comparison, and you can do it incredibly quickly. »
Nvid-IA preaches to its parish
This last point is probably the most relevant for average consumers. Indeed, at a time when the graphic fidelity of video games is really reaching a milestone, hardware sometimes has trouble keeping up; you now need an absolutely monstrous machine to run AAA titles in extreme resolution (4K or more) and with a very high frame rate. In this context, it is a safe bet that AI-based supersampling systems, such as Nvidia’s famous DLSS, will become more and more interesting and take up more and more space in the industry.
However, the wording remains debatable. In practice, there is no absolutely nothing stopping developers from designing graphics platforms completely devoid of artificial intelligence.
But as CEO of a trillion-dollar company, Huang obviously doesn’t think in terms of practical feasibility; it’s profitability who interests him… just like shareholders. And this distinction is important, because this was not a statement aimed at an audience of gamers and technophiles.
At this conference organized by Goldman Sachs, the CEO was addressing an audience of financial veterans who probably have no idea what DLSS is. So his remarks should be taken with a grain of salt, because his goal was not to report on a change in technology trends; It was very clearly a question of pandering to investors and shareholders for whom AI is above all a means of making a share portfolio grow.. It doesn’t take much of a step back to realize that one of AI’s biggest moguls was simply preaching to his parish, extolling the virtues of a technology in which he has a near-monopoly.
On the gamers’ side, what should be remembered from these statements is that AI-based technologies will continue to take up more and more space in Nvidia’s catalog. In the current context, there is only one scenario that could possibly change the green giant’s mind: a possible explosion of the ” AI bubble “, as some analysts predict. But in the meantime, gamers will have to get used to eating DLSS in every way, for better or for worse; it remains to be seen whether this trend will actually benefit gamers or whether Nvidia will seek to impose it against all odds, thus increasing its customers’ dependence on its technology.
The transcript of the conference is available at Seeking Alpha.
🟣 To not miss any news on the WorldOfSoftware, , .
