The US side directly pushes the reversal! The phenomenon of NVIDIA's counterattack has emerged, and foreign media have lamented that the dust has settled
Over the years, the development of flagship CPUs of Intel and AMD has been the background, and the development of NVIDIA's flagship GPUs has always been the existence of the little brother. However, with the rapid popularity of artificial intelligence around the world, Nvidia has also ushered in a turnaround.
GPU chips dedicated to improving computing power began to sell well, directly making NVIDIA's market value rapid**, and even AMD and Intel began to increase investment in GPUs. But the United States has directly promoted this reversal, and the trend of NVIDIA may be turning again!
Although there is still a gap in the current chip manufacturing level of the mainland, the mainland is the world's largest semiconductor market, and basically 70% of the world's chips will flow to the mainland. Therefore, the American chip industry giant is very important to the mainland market.
Since U.S. companies account for about half of the world's total integrated circuit shipments, Intel, NVIDIA, Qualcomm and other companies have a considerable share of the mainland market, and almost all of their revenue comes from the mainland, accounting for about 30%, and some are even higher.
For example, Qualcomm's revenue from mainland China once reached 671%, and Nvidia's revenue from mainland China also reached 47%.
However, in order to obstruct the development of China's high-end integrated circuit industry, the United States seems to have taken no account of the interests of these American companies, and in recent years has continuously tightened export control regulations and restricted US integrated circuit companies from shipping high-end integrated circuit series to the mainland market.
Later, it also extended its spearhead to the field of artificial intelligence, directly informing Nvidia that it could not ship A100 and H100 chips to Chinese companies.
However, Nvidia did not want to put pressure on the mainland market, so it modified the A100 and H100 chips and launched the mainland-specific GPU chips A800 and H800, which meet the export regulations and can continue to be sent to Chinese companies without license restrictions.
The reason why Nvidia is doing this is because the demand for GPU chips in the mainland market is very large. The mainland's artificial intelligence industry is developing very rapidly, and even in the development of the global artificial intelligence industry, the mainland is an indispensable and important part.
Some time ago, the first Global AI Security Summit was held in the UK, and as early as September, the UK sent a formal invitation to China. British Prime Minister Rishi Sunak has publicly stated that inviting China to the summit is"Absolutely"。
Tesla founder Elon Musk spoke with Sunak in London and agreed that there would be no point if China did not participate.
This is enough to see the importance of China's artificial intelligence development in the world, but the United States is afraid that it does not want to see that some time ago it updated the chip rules and continued to expand the restrictions on artificial intelligence-related chips, and even special versions of chips were banned from export.
The point is that the US side can't wait, and this plan will not take effect until 30 days, and within a few days, it will directly notify the restrictive measures to take effect immediately.
The United States must directly push Nvidia to reverse the situation, and finally waited for the hot GPU chips, which could have fought a turnaround battle and completely abandoned them"Little brother"to catch up with the counterattack of old giants such as Intel and AMD.
This time, Nvidia suffered significant losses due to restrictions imposed by the United States, first of all, the loss of chip shipments. The previous A100 and H100 were restricted from shipment, and now even the modified A800 and H800 cannot be shipped normally.
Speaking of which, chip modification is not so simple, not only to modify the chip design, but also to communicate with TSMC and modify the relevant content of chip manufacturing, which also increases the cost to NVIDIA, it did not expect to be restricted again when it was not sold well.
The point is that the restrictions in the United States came very suddenly, and Nvidia modified a lot of chips that had already been produced.
And they are all special editions released specifically for the mainland market, and they can be sold if they are not restricted**, but after the U.S. restrictions are upgraded, Nvidia still has to consider how to deal with it, and naturally it can only reduce the price of sales in other regions.
Second, the loss of market share decline. NVIDIA is worried about declining market share.
The United States took the initiative to cede NVIDIA's market share, which is bound to quickly fill the Chinese companies in the research and development of GPU chips in the mainland. Some time ago, Huawei's Ascend 910B also showed its strength, and it is said that the computing power is no less than that of NVIDIA A100.
You must know that Nvidia CEO Jensen Huang previously said that there is only one mainland market in the world, and the mainland market is irreplaceable. Since the world can no longer find such a large market, the restrictions in North America have brought irreparable losses to NVIDIA.
In this regard, some foreign media bluntly said: NVIDIA's situation is basically solved, advanced GPU chips can no longer be shipped to Chinese companies, with the continuous enhancement of the strength of Chinese companies in the mainland market, Nvidia can only face the fate of being replaced in the future!