Last Friday, NVIDIA, the "king of kings" in the US stock market, saw its stock price collapse, with a single-day drop exceeding 10%, shocking Wall Street. At the same time, over the past month, NVIDIA's stock price has fallen by 20% in total, and the long-term downward trend has become very obvious.

Now the question arises: Why did NVIDIA's stock price suddenly collapse last week? The truth is somewhat humorous; NVIDIA's stock price collapse was actually affected by the sharp drop in stock prices of Advanced Micro Devices (AMD).

Last Friday, AMD's stock was heavily sold by investors, with a single-day drop of 23%. The trigger was a significant drop in AMD's revenue and profits in the first quarter of this year. Investors obtained this result through various channels in advance and decided to sell stocks before AMD officially announced the result (before April 30) to avoid being trapped.

AMD is one of NVIDIA's largest financial supporters. AMD's main business is to integrate NVIDIA's GPUs, NPUs, and storage chips from other companies into an AI server, which is sold to internet companies such as Meta and Open AI for use.

Advertisement

Therefore, the significant drop in AMD's revenue and profits will definitely affect NVIDIA's income; in the first quarter of this year, NVIDIA's revenue and profits will also definitely not meet expectations; throughout 2024, NVIDIA's income will not be as good as many people outside expect.

As for the truth behind the significant drop in AMD's revenue, it is related to the inability of AI in the United States to land and commercialize.

In the past two years, the United States has experienced a technological revolution every three months, blowing up the AI myth; but when the carnival is over, Americans have discovered a cruel fact, that is: the so-called AI technology in the United States cannot combine with the real industry to create products that truly promote the progress of social productivity (like the iPhone at that time) and generate positive revenue industries.Open AI's chatGPT and sora have been hyped as omnipotent electronic deities. The metaverse and Boston Dynamics' robots were also touted as miraculous. But can you imagine that the most advanced drones in the United States still frequently experience signal loss and are unable to return home automatically?

Can you imagine that as China's automated factories and unmanned ports emerge and become widespread like mushrooms after rain, American factories' production lines still require a large workforce to hammer and tinker, which is neither efficient nor reliable in terms of quality.

Can you imagine that the United States, claimed to have the most developed financial industry, is filled with extremely inefficient cash transactions and ATM settlements? Due to the backwardness of electronic payment technology and strong barriers between banks, the American financial and banking sectors cannot even achieve basic real-time payments and settlements.

Recharging a phone bill can take up to 15 minutes, and the slow efficiency is maddening.

Given that American AI is so powerful, can it be integrated with the manufacturing industry, or even the financial industry, to improve social production and transaction efficiency? Whether it is unwilling or unable (most likely unable), this is a fatal flaw of American AI.

If it cannot be integrated with the physical industry on a large scale to increase productivity and transaction efficiency, it cannot generate positive returns. Without positive returns, American AI companies like Open AI are just wasting money; such a money-burning model is destined to be unsustainable.

3

Since the second half of last year, many people have reported that chatGPT (an intelligent chatbot) has become increasingly foolish, with lower and lower accuracy in responses and slower reactions than when it was first released. The reason is that the number of visitors to chatGPT has increased, and the computing power is not enough; however, Open AI does not have sufficient funds to purchase enough AI servers to keep up with the surge in demand for hardware computing power.The so-called paid access to ChatGPT currently cannot bear its astronomical cost expenditures (expensive AI chips + huge power consumption); however, it is also impossible to get people to pay a higher price to purchase ChatGPT. ChatGPT simply cannot provide highly professional practical services and is not worth that much money.

Spending a little money to chat with an intelligent robot for fun is fine; but don't expect me to spend a lot of money on it. Americans are willing to pay up to $15,000 for Tesla's intelligent driving services; yet ChatGPT costs only $20, and many people already consider it too expensive.

PS: See, the United States has a demand for the combination of AI and physical entities, such as Tesla's FUD autonomous driving. It's just that the scale of this practical AI in the United States is too small to make waves.

You should know that even the paid version of ChatGPT doesn't make money, let alone completely free and open-source AI services like Musk's Grok, Google Bard, and Llama, which are even more money-sucking black holes. Therefore, after a period of turmoil and a lot of money spent, these American AI businesses are running out of steam and can't afford to buy a large number of AI servers.

Open AI has even been forced to develop its own AI chips because the super powerful AI servers and Nvidia AI chips are just too expensive. Open AI also has to continuously seek funding from overseas financial groups like Japan's SoftBank and Saudi Arabia's G42. Open AI's cash flow is extremely tight now, and there is a risk of bankruptcy and suspension at any time.

3

It is only natural that the ethereal American AI companies are out of money and chips, and that the revenue and profits of super powerful computers like AMD and Nvidia have plummeted. In the long run, if American AI technology cannot be combined with the physical industry on a large scale to create a business model that can generate a large amount of cash flow, then it is also highly likely that Nvidia's revenue and profits will not meet market expectations, and the stock price will continue to fall for a long time.

The recent collapse of AMD and Nvidia's stock prices indicates that the American AI myth can no longer be sustained and has temporarily reached a dead end.

Of course, Nvidia and AMD can sell AI chips and AI servers to us in China; we in China have a large number of AI physical applications, such as intelligent driving, smart ports, smart factories, smart cities, and so on, with a scale of several trillion dollars.The issue at hand is this: we don't need mid-to-low-end AI chips, as we are capable of manufacturing them ourselves; however, when it comes to high-end AI chips, the United States is overly particular and unwilling to sell, leaving us with no alternative.

In conclusion, I would like to say: no matter how great a technology may be, it must be applied in practical ways; the great steam engine revolution of the past was also aimed at addressing the problem of pumping water out of coal mines to enhance societal productivity. Unfortunately, the current state of AI in the United States has gone astray. It does not aim to solve real-world problems, nor does it wish to integrate with the down-to-earth manufacturing industry. It merely seeks to maintain a lofty position, creating an electronic deity that is articulate, knowledgeable, and omnipresent.

While such a deity may be useful in providing answers and guidance, it ultimately serves as nothing more than a bowl of chicken soup – consuming too much of it does not benefit real life. Spending trillions of dollars to create an electronic deity is an immense waste and sin in terms of resources.