Could the GPUs that power artificial intelligence rise in price like they did in the crypto boom? Read the analysis here!
Given the monumental growth of ChatGPT and its incorporation into the search engine Microsoft Bing, some players began to worry. It turns out that even though the AI runs on a huge server and you can query it from the cloud, part of the process by which it works is a continuous learning algorithm that uses special GPUs. Could the increasing need for technology of this type destabilize the prices and accessibility of video cards as it already happened during the cryptocurrency mining boom?
The thing is, despite the coincidences, crypto mining and learning models are different. Traditional video cards will not be able to be used in large servers to provide power to an algorithm. For example ChatGPT it processes countless amounts of text per day, but this requires specially made plates. Thus, it is not the Nvidia 4090 that the bot requires but dedicated hardware made for servers.
According to experts, the artificial intelligences of the future will not run on their own hardware for consumers, but will be in the cloud controlled by a group of developers. And yet…
The question then is not about accessibility in the present but in the future: is it going to help Nvidia to create RTX 5090 boards when it has a limited number of implements, and surely it will get more money to sell to some server than to consumers? The problem in the future could be that, a relative stagnation of the technology because it will not be monetarily useful to produce all these plates.
What could benefit gamers specifically in the long run would be if other industry giants seize this opportunity and the relative “disappearance” of nvidia to make your own space. AMD already has its own highly competitive hardware, and Intel is increasingly betting on creating competent boards. With the advancement of artificial intelligence, you may not run out of boards, but maybe they are not Nvidia boards.