Behind the success of ChatGPT, a future problem for PC gamers

Photo of author
Written By tsboi team

The Best Url Shortener, QR Codes, Bio-Profile Link & File Sharing Platform.

ChatGPT’s immense capabilities and databases have been emulated by tech giants. But its development has a cost in graphics cards that could well be the cause of a new shortage of GPUs.

The development of the artificial intelligence ChatGPT required 10,000 graphics cards//Source: Frandroid

The graphics card market is barely recovering from its post-Covid shortages and a new supply crisis is likely… and ChatGPT has a role to play in it. In any case, this is what the US online media prophesies. techradar. In question: the growing popularity of these artificial intelligences dominated in personal uses.

10,000 GPUs to create ChatGPT

In fact, in addition to their use in video games, GPUs (and more precisely their so-called Tensor cores, specialized in calculations for artificial intelligence) are also used in data centers around the world. One of the best known models is the Nvidia A100 GPU.

While the current demand for data centers does not cause a shortage in and of itself, it is the gigantic size of data used by AIs like ChatGPT that is causing fear in techradar a looming crisis. Because they are based on large swathes of what has been written on the Internet, these large algorithms are called “Large Language Models” (or LLMs). But to create and maintain the databases of an LLM like ChatGPT, 10,000 Nvidia graphics processors were needed, the specialized medium specifies. fierce electronics — plus a controversial subcontractor.

AI everywhere, graphics cards nowhere

Even if the need for GPUs on this scale remains rare, the risk would be to see these uses go mainstream if the integration of these giant AIs into our daily services follows the appetite of the tech giants. In early February, Microsoft released a new version of its Bing search engine that directly integrates ChatGPT.

However, according to two analysts cited by the American business magazine Forbesif Google wanted to keep up and embed similar AI, like the Google Bard, into its billions of daily search results, it would need more than 4.1 million Nvidia A100 GPUs at an estimated total cost of $100 billion.

One scarcity reminiscent of another

Therefore, such a demand would cause a significant increase in the price of graphics cards, sold to the highest bidders at the expense of smaller industries and the consumer.

A perspective reminiscent of earlier GPU shortages, which occurred following a sharp increase in demand following the Covid-19 pandemic, quickly followed by the development of cryptocurrency mining. The situation has returned to normal recently, thanks in particular to the evolution of the performance of Ethereum and the fall in the values ​​of several other major cryptocurrencies.

A crisis in the hands of Nvidia

The only glimmer of hope: As mentioned above, making such expensive and power-intensive AI widely available isn’t cost-effective in the long run. Access to the new version of Bing with ChatGPT today is still limited and, as the finger points Forbeseven Microsoft will have to look for a cheaper alternative.

But the development of these LLM algorithms will not disappear. So the choice remains in the hands of specialized GPU manufacturers like Nvidia – if they decide to target this market and play to its advantage, prices are likely to rise rapidly.

To follow us, we invite you to Download our app for Android and iOS. You can read our articles, archives and watch our latest YouTube videos.

Rate this post

Leave a Comment