With the restriction of the export of many of NVIDIA's chips, the AI computing power market is facing unprecedented challenges. This situation not only led to the announcement of an increase in computing service fees by companies such as Winner Technology and Zhongbei Communications, which are the concept stocks of computing power leasing, but also triggered a situation in which prices were raised in the industry.
According to a recent interview with the Financial Associated Press, due to the impact of the 1 times increase in the fee of Nvidia A100, the price increase in the industry has been heard endlessly. It is the norm for the graphics card to rise by 2 times, and even the 4090 has risen by 5 times a while ago, but then it has come down. This has led to a significant increase in the cost of AI computing power.
Nowadays, AI computing power has become the focus, not only attracting attention in the industry, but also attracting enterprises with GPU resources to make cross-border layouts.
A person from a leading listed company in cyber security said that they are also doing computing power leasing now, mainly based on A cards, and there is still a little margin but not much. By cooperating with NVIDIA and domestic high-end chip manufacturers, these cybersecurity companies continue to optimize the multi-computing power system to meet customers' needs for AI large model training.
However, due to the impact of export restrictions on Nvidia chips, the AI computing power market has been in short supply. In the irrational situation of the market, computing resources continue to be tight, resulting in a large number of computing service charges. Some business people say that the market for computing hardware such as graphics cards is rising, and this trend is likely to continue.
In response to this phenomenon, some companies have begun to prepare for a rainy day and stock up in advance to deal with possible market shortages. And some researchers are starting to look for alternatives to reduce their dependence on Nvidia chips.
In the process of responding to Nvidia's chip export restrictions, domestic companies have shown a high degree of flexibility and adaptability. They quickly adjusted their strategies and actively explored the development direction of diversified computing power. By strengthening cooperation with domestic and foreign chip manufacturers, these companies have accelerated technology research and development and product upgrades, and continuously improved the performance and reliability of multiple computing power.
Among them, Inspur Information is a typical example. They have built a G7 multi-computing platform, which is compatible with 15 kinds of AI chips at home and abroad.
This open and inclusive strategy has enabled Inspur Information to achieve significant advantages in the field of AI computing power. They not only cooperate with internationally renowned chip manufacturers, but also actively cooperate with domestic emerging chip manufacturers to jointly promote the development of multiple computing power technologies.
In addition to Inspur Information, many domestic enterprises are also actively deploying diversified computing power fields. Through independent research and development and technological innovation, they have gradually broken the monopoly position of foreign companies in the field of AI chips. The products of these companies have not only achieved breakthroughs in performance, but also won market recognition and praise.
Although, in the period of computing power bottleneck, diversified computing power has become the focus of many manufacturers. However, the development of multiple computing technologies has not been smooth sailing. Enterprises are still facing great challenges in the face of problems such as long development adaptation cycle, large investment in customized development, and long service migration time.
Therefore, they still need continuous investment and efforts in technology research and development, talent training, market expansion, etc.
Despite this, industry insiders generally believe that diversified computing power is the future development trend of the AI field. They said that with the continuous advancement of technology and the continuous growth of market demand, diversified computing power will become the mainstream choice. At that time, the negative impact of Nvidia's chip export restrictions will gradually weaken.
In general, the limited export of NVIDIA chips has indeed brought a lot of impact to the AI computing power market. But at the same time, domestic enterprises have also shown a high degree of flexibility and adaptability. They have gradually adapted to this change and achieved good results by actively responding and flexibly adjusting their strategies.
However, for some start-ups and small and micro enterprises, this challenge can be a little more acute. They may not have enough money and resources to cope with changes in the market, and may even be forced to exit the market because they cannot afford the cost pressure.
So, with the continuous development and improvement of multiple computing technologies and the continuous growth of market demand, what will the AI computing market look like in the future?Everyone is welcome to share their thoughts.