ChatGPT may cause another GPU shortage – sooner than you expect
Unless you live in a deep and cozy cave, you’ve certainly heard of ChatGPT. The AI-based text tool developed by the nonprofit organization OpenAI has become ubiquitous on the internet and heralds an explosion of AI development and widespread use.
To create and maintain the huge database of AI-analyzed data that ChatGPT requires, the creators of this tool apparently used an incredible 10,000 Nvidia GPUs to train the AI to generate responses. interesting and quirky.
With the growing popularity of AI, tech giants like Microsoft And Google are looking forward to integrating it into their own services. Forbes (opens in a new tab) estimated the cost to integrate AI into each Google search, estimating the 4,102,568 Nvidia A100 GPUs needed would bring the tech giant $100 billion — paying Nvidia of course, which is a numbers are really hard to wrap your head around even if it’s just a guess by Forbes.
Gamers among us will probably be reading this and sweating a bit, as it spells bad news for those looking to buy a GPU soon. We could see a graphics card shortage like the recent one due to increased demand for crypto mining, as companies will need multiple GPUs to train AI cores to make the most of the burgeoning craze. this present.
Analysis: Hello darkness, my old friend
This shortage fear will likely only affect Nvidia card sales as they have the most advanced AI cores on the market right now – however, that could change. If other companies like AMD and Intel see high demand and potential profit, they can work harder to catch up with Nvidia in AI.
On the other hand, if Nvidia turns more attention to the field of AI – and Blue Team definitely want – it could leave an opportunity for AMD to fill the consumer gaming space. The lack of Nvidia GPUs for gamers could push people to AMD graphics cards.
If the future demand for GPUs reaches extremely high levels, the availability of those GPUs will suffer greatly for casual gamers. We won’t know how far the demand can be collected and how many GPUs will be needed to train the AI models.
As mentioned above, we had to deal with a massive GPU shortage not too long ago thanks to miners and speculation in the crypto space, which gives us an idea of the shortage. How the new GPU could play out. Cryptocurrency miners bought as many GPUs as possible to mine currencies like Ethereum and any shares that were not used for mining were bought by speculators who blew the price up. Massive.
The shortfall only eased last year with the ‘crypto crash’ and Ethereum consolidation, and we still see Nvidia graphics cards selling for much more than MSRP. If Nvidia commits to producing more GPUs for AI training purposes, the price of graphics cards will rise again.