Nvidia and AMD
Exciting advances in generative artificial intelligence have seen rapid adoption of the technology by many tech, industrial and commercial companies, and “graphics chips” have become the most in-demand component on the market today, as these chips are considered booming “Oxygen” for development. AI systems are still breathing today.
And any company that really wants to compete in the new world of artificial intelligence will have to turn to graphics processing units, because without the power these chips provide, most of the services of generative artificial intelligence would not exist.
At present, Nvidia has a dominant position in the artificial intelligence chip market, with a market share of more than 80%. This advantage is due to the fact that the company has two super chips, namely the A100 chip and the H100 chip. It is the most powerful of all artificial intelligence chips currently on the market, which he describes as “legendary,” the best chip ever made by a human brain in the field, and sells for $40,000.
But with AMD’s MI300X graphics chip coming later this year, Nvidia’s absolute dominance in the AI chip market appears to be coming to an end.
AMD has long been known for making traditional computer processors, so if the MI300X chip succeeds in gaining market confidence, the company will enter a new market for it and try to break Nvidia’s dominance in the field.
But the question remains, does the MI300X, which AMD says is the most advanced in AI to date, really have the specs to break Nvidia’s dominance?
Rudy Shoshani, an expert in digital transformation, said in an interview with Sky News Arab Economy that the background of evaluating the Super MI300X chip is not only related to the technical specifications of the chip, but also must take into account that it is AMD’s first artificial intelligence chip in the field of chipsets, This means that the company does not have enough experience in this field compared to Nvidia, which suggests that the MI300X is superior to Nvidia’s H100 chip in some specifications, but no one knows how the MI300X will perform and interact with the operating system, and the This is considered critical in the market adoption of AMD chips.
Shoshani said that the MI300X uses up to 192 GB of memory, which means it is more suitable for artificial intelligence programs than the H100 chip, which uses up to 80 GB of memory, and pointed out that this thing makes the AMD chip in fact, the chip does not have a so-called Transformer Engine, That engine, found in the competing H100 chip, provides a six-fold performance boost and can significantly accelerate artificial intelligence performance.
The Practical Advantage of “Nvidia”
Shoshani explained that so far, the MI300X has only a “preference” in the description because many issues related to the chip’s performance have not yet been announced, while the Nvidia chip has a “preference” in terms of market confidence and speed, as well as success in this field. Historically, Nvidia is expected to intend to update the specs of the H100 chip before the MI300X hits the market, to prevent any other company from claiming to have the strongest specs.
Shoshani confirmed that all of the above does not mean that the MI300X has no hope of success, as it depends on whether the chip can convince companies if it proves through its performance that it is a “big deal” given that breaking Nvidia’s dominance is a “big deal”. Word”, which will not be realized in the foreseeable future, because the degree of market adoption of it can be seen in the first half of 2024. Competing with each other in order to achieve innovation in artificial intelligence technology, will not abandon Nvidia’s products in favor of products that know nothing about their performance, at least in terms of perspective.
Technology analyst Khaled Moussa confirmed in an interview with Sky News Arab Economy that it is too early to say how well AMD will be able to compete with Nvidia in artificial intelligence chipmaking because of the difficulty of pulling the market. very big. Nvidia is the backbone of this industry and excels in many technical elements of the space.
Musa believes that AMD’s MI300X chip could become the second-best option in the market behind Nvidia’s chipsets, especially if it’s cheaper in the market than the H100 ($40,000), noting that many startups buy Sorry. It bears the cost of Nvidia chips because of the recent problem in the market, so AMD offering artificial intelligence chips at an affordable price may help the company gain a foothold in this industry, if the MI300X proves its ability to operate against the competition , You must know that AMD has not disclosed the price of its new products.