Artificial intelligence seems to be the technology or, rather, phenomenon of the moment. A year or so ago it was Web3, but it has been overshadowed by other innovative and emerging technology. While many companies are making (or on the way to) a lot of money, who is making money is NVIDIAIt is running out of graphics cards for AI.
Being a visionary is not easy. Jensen Huang is arguably one of the greatest modern visionaries (well, he and the mythical sucks at him). Although we know his graphics cards from gaming, they are actually used for many other things like advanced math, architecture, medicine, etc.
For years they have been preparing for a field more typical of science fiction, such as artificial intelligence. This is how Tensor Cores were born, specific cores designed for AI that are currently the basis of DLSS. This technology, launched in February 2019, represents an advance of its time from which the company reaps benefits.
Years of development that pay off
The price of AI-specific graphics cards, such as NVIDIA A100 have risen in price by 37.5%. These graphics cards are not sold individually, they are integrated into specific servers which have become more expensive.
Specifically, they are mounted on the NVIDIA DGX A100 servers, which include six A100 graphics. These systems, when announced in 2020 they had a cost of 200,000 dollars per unit. Currently, they would be around $275,000 per unit due to strong demand for AI.
According to IDC China, this is not an isolated phenomenon, it is happening with other products of the company. Currently they also offer A800 exclusive to China, which are a cut down version of the A100. According to reports, the price of A800 is up 20%. But, as if that were not enough, the demand is so strong that there is no stock available.
Until February of this year the situation was good, the products being sent in a month, which is normal. Currently, the situation is so complicated that the minimum delivery time is three months. For some models (not specified) it is indicated that they will not have stocks until December.
To give us an idea, the NVIDIA V100which are the simplest for IA, rThey are already at $10,000. The H100on the other hand, exceed $35,000 per unit. In addition, the thing looks like these prices are low and will continue to rise.
TSMC, which makes the chips for these graphics cards, is working to ramp up production as fast as it can. The truth is that it is not as easy as turning on a printer, this process is slow and tedious.
Again, in the right place at the right time
For those of us who have been in the hardware business for so long, this is not a surprise. We have already seen in the past the brutal lack of stock of graphics cards for Ethereum mining. Such was the demand and madness that planes were chartered to transport the graphics cards.
History repeats itself, now, in another line. At the moment, it seems that it does not affect the graphics cards, for the moment. As this maelstrom continues and it is not possible to satisfy the demand, it is likely that companies will throw themselves at what there is. It cannot be ruled out that, in a few months, if AI continues to grow this excessively, we could see a shortage of gaming graphics cards.
We must not rule out that NVIDIA may reduce the production of gaming solutions in favor of AI products. In the end, it is a company and its function is to make money.