After the Invidia challenge in the field of intelligence .. Are Huawei heading towards graphics cards?

In an important technological step that may bear greater indications than just the artificial intelligence market, the giant Chinese Huawei began to deliver its new advanced system of artificial intelligence chips, known as Cloudmatrix 384, to its main customers in China.

This rapid development comes in a direct response to the increasing restrictions imposed by Washington on the export of advanced semiconductors, foremost of which is artificial intelligence chips from Invidia, which limits the ability of Chinese companies to obtain the technology necessary to develop huge artificial intelligence models, due to the increasing American restrictions.

Based on well -informed sources, Reports indicate that Huawei It is based on its headquarters, has already sold more than 10 groups of the Cloudmatrix 384, which is characterized by its ability to connect a very large number of chips together to create a massive computing power, and information indicates that the data centers of Chinese technology companies were one of the first beneficiaries of these shipments.

Industry analysts expressed their admiration for the speed that Huawei was able to develop and start charging the Cloudmatrix system that connects 384 processors dedicated to artificial intelligence to provide the mathematical capacity needed to develop the services and models of complex artificial intelligence, but the founder of Semianalysis has gone. Dylan PatelTo say that Huawei’s development of this system means that China now possesses an artificial intelligence system capable of outperforming competing systems from Invidia.

It is noteworthy that this achievement of Huawei comes at a time when the American export restrictions are witnessing additional tightening. Earlier this month, Inviteia announced that it would incur profit losses of $ 5.5 billion, after its H20 SIM, which she previously designed to comply with old restrictions, now requires a special license to sell it to Chinese clients, by a decision of former US President Donald Trump.

Huawei assures its local customers that its Cloudmatrix system is greatly outperforming the NVL72 system from the NVIDIA -a common artificial intelligence system consisting of 72 GB200 chip and used by giant American technology companies, in terms of total computing power and memory capacity according to the company’s presentation and informed sources.

Although the ASCEND 910C chips used by the Cloudmatrix 384 are the weakest performance on its own compared to the GB200 processors from Invidia, Huawei has compensated for this difference using a much larger number of chips (384 against 72), and linking them through a super knot using advanced light communication techniques, which greatly enhances the total performance of the group. Huawei claims in its presentations that Cloudmatrix surpasses NVL72 by 67% in computing and more than three times in the total memory capacity.

Analysts explain, such as Dylan PatelHuawei “compensates for the poor performance of the individual slide with advanced networks to enhance the performance of the group in the group,” noting that the company took advantage of its long communications experience to improve the performance of its systems.

Are Huawei heading towards the GPUS world?

The remarkable progress made by Huawei in developing complex and high -performance slides systems such as Cloudmatrix 384 not only serves the artificial intelligence market, but also raises important questions about its future path in the wider semiconductor market, developing advanced slide design capabilities, and systems engineering that connects hundreds of processors with high efficiency using visual technologies, and developing a special programmatic environment such as a system Its Mindspore as CUDA’s potential alternative from Invidia, all of which are highly intersecting with the requirements for designing and manufacturing graphics cards (GPUS) for games or professional workstation stations.

Although Huawei’s current focus is to bridge the gap in artificial intelligence chips in the Chinese market and develop parallel computing systems for artificial intelligence, its success in building a complex and strong system such as Cloudmatrix 384 opens the door wide to speculation, so Huawei will take advantage of these growing experiences and advanced technologies that it developed to compete with Invidia and AMD in the graphics card market Consumer or professionalism in the future?

The comparisons show that the Cloudmatrix 384 system suffers from some defects compared to the competitor of the Entepia, the most prominent of which is the consumption of much higher energy and higher operational costs due to the need for experienced engineers in Huawei alternative software, however, in an environment like China where energy and engineers may be available, this system becomes an applicable solution, especially in the absence of other advanced alternatives.

In the end, Cloudmatrix 384 is an important achievement of Huawei that reflects her ability to adapt and innovate under pressure. While its basic efforts are currently focusing on artificial intelligence, the great progress it achieves in this field makes the anticipation of any possible future step towards the world of graphics cards is very natural and interesting in the world of technology and games.