news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

On July 22, at the 47th Korea Chamber of Commerce and Industry Jeju Forum, the Chairman of the Korea Chamber of Commerce and Industry and Chairman of SK GroupChoi Tae-wonTo warn participants that if the company cannot passAITechnology makes money,NvidiaThe business model may "collapse".

Choi Tae-won said at the forum: "If competitors such as AMD and Arm can provide high-quality chips at a lower cost, Nvidia's business model may face severe challenges and even risk collapse." He also compared the current artificial intelligence boom to the California Gold Rush in the mid-19th century, and predicted that Nvidia is expected to continue to maintain its peak market value in the next three years, just as prosperous as the pickaxe and jeans manufacturers during the gold rush.

However, Cui Taiyuan also reminded everyone: "The gold rush will end one day. When gold is no longer easy to obtain, the mining tools will lose their market. Similarly, if the artificial intelligence industry cannot continue to make profits, this craze may quickly recede and repeat the mistakes of the gold rush."

A few weeks ago, Choi Tae-won traveled to the United States and had in-depth exchanges with senior executives of technology giants such as Amazon, Intel, Microsoft and Open AI. In April this year, he also met with Nvidia CEOJen-Hsun HuangChey Tae-won pointed out: "Large American technology companies have shown strong demand for efficient artificial intelligence data centers built with our semiconductor and energy solutions. Although it is impossible for us to provide all the components required for AI data centers, we will use our own technological and material advantages to strive to build more efficient artificial intelligence data centers."

Nvidia aims to reach 500 billion U.S. dollars in data center by 2023GPUNvidia became the world's most valuable company last month due to its record sales. Although its stock price has declined after the market adjustment, it still ranks among the top three in the world in terms of market value. Given that the strong momentum of artificial intelligence development has not slowed down, we have reason to believe that Nvidia will continue to maintain its industry leadership in the foreseeable future.

However, it is worth noting that the cost of training a new generation of large-scale AI language models is increasing at an alarming rate. Dario Amodei, CEO of Anthropic, revealed that the cost of model training currently in progress has reached billions of dollars, and it is expected that models costing hundreds of billions of dollars will appear by 2025. The surge in the cost of AI development has attracted the attention of financial institutions such as Goldman Sachs, who have begun to question whether the excessive hype and huge investment in AI can bring corresponding returns.

Although artificial intelligence has become an indispensable part of modern society, its future development still depends on whether companies can find a way to make money from this technology. If a sustainable profit model cannot be found, the competition in artificial intelligence may turn into a bubble and eventually burst. In this case, investment in hardware designed to support artificial intelligence processes may drop significantly, and Nvidia's business may be affected.

However, industry insiders do not think that the "green team" (i.e. technology developers) will back off. Looking back at Nvidia's history, when its founder Huang Renxun started his business, artificial intelligence had not yet emerged. Nvidia has always had a solid foundation in the gaming industry as its core market, and its chips continue to demonstrate value in a variety of applications.

However, the main challenge facing Nvidia comes from its competitors. Although AMD has made some GPUs with good performance, it still lacks in super sampling technology. FSR 3.0 technology is still considered to be behind DLSS 3.5, but AMD is catching up quickly. At the same time, Intel is also making progress in the GPU market. Its Arc GPU and XeSS super sampling technology are changing the market competition landscape.

As long as Nvidia can continue to deliver unparalleled performance, individual and institutional customers will continue to buy its products. But its competitors are not slowing down. As AI giants such as Microsoft, Amazon, Google, and OpenAI increase their investment in independent hardware research and development, Nvidia's leading position in AI acceleration may face unprecedented challenges. (Xiaoxiao)