Your attention is my accelerant, pay attention to it, learn more about finance, learn about finance, and understand finance.
As we all know, the United States is worried about the rise of China's AI industry, in order to protect its own interests, the United States first banned the export of Nvidia's H100 and A100 chips to the Chinese market the year before last, and then restricted the scope of the ban on more chips, such as H80, A800 and other models. In the face of this situation, Nvidia not only complained about the United States, but also found another way, and launched castrated chips H800 and A800 that meet the ban requirements to meet the market demand. However, the main problem with the castrated chip is the decline in performance, according to **, the most powerful H20 chip may have lost 80% of its performance compared to the original H100, leaving only 20%. In such a situation, many large manufacturers have begun to look for alternatives, among which Huawei's AI chips have become the target of many enterprises, especially the Ascend 910B. According to reports, the Ascend 910B has more than 5,000 units on order, and although its performance is slightly inferior to NVIDIA's A100, it can be compensated for by clustering and other means. In addition, Huawei's Ascend 910B series AI chips not only have cost-effective advantages, but are also domestic brands, so large manufacturers have turned to Huawei and refused to buy NVIDIA's castrated chips.
While Huawei's AI chips are being welcomed, how should NVIDIA respond? The most direct measure is, of course, to cut prices to compete directly with Huawei's Ascend 910B, and perhaps to win back some customers, after all, Nvidia has the advantage of the CUDA ecosystem. Recently, it was reported that Nvidia has begun to formulate a ** H20 chip for the Chinese market, with each H20 between $12,000 and $15,000, and some dealers have priced it at 110,000 yuan in China. Compared with the previous H100 price, this ** has been reduced by two-thirds, because the H100 once sold for more than 300,000 yuan. At the same time, Nvidia's pricing is close to the price of Huawei's AI chips. It can be clearly seen that Nvidia is indeed feeling a lot of pressure, after all, the company previously occupied 90% of the Chinese AI chip market, with a market size of about $7 billion. If it loses this market, it will be a huge blow to Nvidia. What's more, once the AI chips of domestic manufacturers such as Huawei rise, Nvidia will completely lose this largest market. Therefore, Nvidia had to put down its body and launch a ** war, trying to suppress Huawei's AI chips by cutting prices. The next question is, can the reduced price of Nvidia chips attract consumers to buy? In the past, everyone complained about the performance of Nvidia's castration chips, but no one would buy it if the price was not reduced, but now that ** has declined, will anyone still be willing to buy it?
Judging from the current trend, Huawei's AI chips have indeed achieved a certain advantage in the Chinese market, and Nvidia has to compete by cutting prices in order to cope with this situation. However, we might as well think about whether the price reduction can completely solve the problem faced by Nvidia. When customers buy AI chips, in addition to **, performance is also a very important consideration. Although NVIDIA's castration chip has declined in performance, it still has advanced CUDA ecological advantages, which cannot be completely replaced by domestic manufacturers such as Huawei. Therefore, although price reductions can attract some consumers, there is still some uncertainty about whether they can win market share.
On the other hand, Huawei, as a well-known domestic brand, has a strong influence in the Chinese market, and its Ascend 910B series AI chips perform well and have a high cost performance. Although it is slightly inferior to NVIDIA's A100, it can make up for some of the performance gaps through clustering and other means. In addition, Huawei's AI chips are reasonably priced and have a relatively low cost, making them more competitive. As a result, many large manufacturers have turned to Huawei, choosing to use Huawei's AI chips to replace NVIDIA's products.
To sum up, with the rapid development of China's AI industry, AI chips from domestic manufacturers are gradually rising, and NVIDIA's hegemony is facing challenges. In response to this situation, Nvidia can only compete with Huawei by cutting prices. However, there is still uncertainty about whether this battle can truly turn Nvidia's predicament. When customers buy AI chips, in addition to performance, the importance of brand recognition and ecosystem cannot be ignored. In the future development, whether Huawei and other domestic manufacturers can further improve the performance and competitiveness of their products, and whether NVIDIA can adapt to market demand, these are all issues of concern. In the end, the market will naturally choose the most suitable products and brands, and the needs and choices of consumers will ultimately determine the direction of the market.
If you like it, you can follow me, share financial advice regularly, and talk to you about financial topics.