The market closed for three days in a row, and finally stabilized at 3,300.

Although the turnover fell back to 1.6 percent, individual stocks rose and fell on average, and the market returned to a state where a partial structural market started.

With the gradual disclosure of financial reports from US technology stock companies, for example, the AI application company AppLovin's performance greatly exceeded expectations, and the stock price surged 46%. The AI application concept reflected in it stood out in A-shares.

What is particularly worth mentioning is the Hong Kong stock exchange technology. Since November 7, the stock price has risen sharply by 277%.

Is AI back?

01

AI+ application mapping

The market fluctuated in a narrow range after opening low in early trading, and there was no significant fluctuation at the close. In terms of sectors, the Internet, precious metals, and power generation equipment led the way, while sectors such as education, telecommunications, and diversified finance registered the highest declines.

On the subject, the concept of controlled nuclear fusion led the way. Dongfang Seiko had 6 consecutive boards, Jiusheng Electric, Hongxun Technology, Hailu Heavy Industries, Yongding Co., Ltd., and Rongfa Nuclear Power rose and stopped.

According to the news, a test platform for a quasi-cyclic symmetric simulator independently developed and manufactured by China recently passed the test. The simulator is a controlled nuclear fusion device, and nuclear fusion reaction is an important way to solve the ultimate human energy problem, and has great commercial and strategic value.

This issue is no longer limited to discussions on topics in laboratories. The AI industry has exploded in recent years, and behind it, huge and stable energy infrastructure support is needed. Nuclear power stocks, like hardware such as chips and optical communication, have entered the scope of investment on the AI circuit.

Hardware and cloud service companies such as Nvidia and Microsoft were the first to eat the first wave of crabs, and their performance and profit growth also began to grow first. With the continuous penetration of AI applications, companies using AI tools reaped improvements in quality and efficiency, and companies providing these tools began to expand their business. The three-quarter report of AppLovin, a leading US AI advertising and marketing company, is an example.

Q3 revenue was 1.2 billion dollars, +39% year over year, with a net profit margin of 36%. Among them, the revenue of software platforms, including the AppDiscovery advertising engine driven by the AI recommendation engine Axon, increased to 0.835 billion US dollars, +66% year over year, verifying the efficiency logic of AI advertising commercialization, and the stock price has increased by 716% since the beginning of the year.

The business model for AI applications has been certified. Revenue covering token costs is a prerequisite for the industry to be willing to try to invest, and it is also a prerequisite for AI applications to prosper. The emergence of applications is expected to bring new demand for hardware, so energy and inference costs are expected to be drastically reduced.

This virtuous cycle not only stimulates the continuous emergence of new bullish stocks in US technology stocks, but also maps A-shares.

In the midst of this three-day turbulent market, AI applications have emerged in a relatively bright structured market. Among them, what can be directly mirrored with AppLovin is the marketing circuit in the media sector, big advertising platforms such as Tencent, Byte, and small media service providers, such as Tiandi Online (the acquisition of Shanghai JiaTou, known as Little AppLovin), and Huiliang Technology (its Miteraral platform specializes in overseas marketing).

However, the overall profitability of listed companies in the A-share marketing sector is low, which effectively verifies the increase in performance and profit improvement brought about by AI marketing. The adjusted EBITDA profit margin is as high as 60%, far exceeding the level of peers.

Expectations are almost fermenting in one application scenario. The logic also applies to other application scenarios, so the mapping of various application branches of AI is rising, but some are just hyping up related concepts, and there is no actual business promotion.

AI marketing: Tiandi Online, EasyDianxia, Blue Cursor, Leo Co., Ltd., Gravity Media, Zhejiang Internet, etc.

Al Agent: Kunlun Wanwei, Wanxing Technology, Jinshan Office, Pan-Micro Network, etc.

AI search: Kunlun World Wide, 360, etc.

AI audio: Kunlun World Wide, Chinese Online, Shengtian Network, Tom Cat, etc.

AI video: Mango Supermedia, Shanghai Film, Huace Film and Television, TV, etc.

Al e-commerce: worth buying, rebate technology, focus technology, Yuanwang technology, etc.

AI education: Doushen Education, Century Tianhong, iFLYTEK, etc.

AI games: Palm Fun Technology, Kaiying Network, Giant Network, 37 Mutual Entertainment, etc.

AI hardware: Doctor glasses, Guoguang Electric, Rambler, etc.

AI toys: Aofei Entertainment, Shifeng Culture, Guangbo Co., Ltd., Yuanlong Yatu, Yao Ji Technology, Tom Cat, Xinghui Entertainment, etc.

It just so happened that the US stock AI had a major event this week. Nvidia's earnings report for the third quarter yesterday exceeded expectations, but in the end, the year-on-year revenue growth slowed down. The median guidance for the next fiscal quarter was also slightly conservative, falling 5% after the market.

Are expectations too high and adjustments are needed, or is the slowdown in revenue causing market concerns?

But soon after Open AI will launch another major product, does that mean that the market paradigm around hardware investment will shift to the emerging application side?

02

Investing in AI: Computing Power or Application?

Judging from the current two major segments of the AI industry - lower-level infrastructure and upper level applications, the real strong performance is still the underlying infrastructure, especially AI computing power, as can be seen from Nvidia's strong performance growth.

Because Nvidia is the only GPU supplier in the world with an absolute monopoly position, customers - cloud computing vendors, are still snapping up Nvidia's Blackwell products. Therefore, there is no need to worry about Nvidia's fundamentals unless it is reported that there is a huge downside in demand, such as cloud computing vendors drastically cutting capital spending on AI computing power.

However, it is also true that Nvidia's performance declined later. Looking back, Nvidia's stock price has been rising for 2 years, and the increase is as much as 10 times. Although some investment banks are optimistic that EPS will reach 5-6 US dollars in 2025. Based on 30 times PE, Nvidia's target price for next year can reach 150-180 US dollars, but after all, it has risen so much that we can no longer brainlessly do as much as the past two years, and we should be careful. In particular, it has been two times in a row. There is a gap between next season's guidelines and buyers' requirements.

It reminds me of Tesla 3 years ago.

Since the second half of 2019, after Tesla's production capacity bottleneck was broken, Tesla's shipments have surged, and each time its performance has greatly exceeded expectations. Fortunately, it also experienced global water discharge caused by the outbreak of the epidemic. As a result, Tesla's stock price was hit by a huge Davis double. In just two years, it rose more than 10 times, peaking at 414 US dollars.

However, as the high growth in shipment volume came to an end, and high US inflation led to a strong interest rate hike by the Federal Reserve, Tesla's stock price fell all the way in 2022. By the beginning of 2023, the stock price had fallen 76% from its historical high.

Will Nvidia follow Tesla's footsteps?

I'm not sure, but that doesn't mean Nvidia isn't worth the investment.

In fact, Nvidia's competitive environment is far better than Tesla back then. Tesla needs to face the siege of many competitors, but Nvidia doesn't need to, even in the foreseeable future, such as 2-3 years. In terms of the growth and certainty of AI performance, Nvidia is unique and far ahead.

So, Nvidia's core issue is valuation.

If the valuation is very high, the capital will naturally lack interest in continuing to buy, but if the valuation falls far, there will soon be capital to bottom out. If the valuation for the next year is around 20-30 times, I think it's quite reasonable; if it's too high, you should pay attention.

As for the application level, it will undoubtedly be the focus of the future. Because the ultimate goal of infrastructure construction is for application. Although the app is still in its early stages, and it is often criticized that there are no apps that can generate particular profit, the good news is that we are seeing the release of the performance of some app companies one after another, such as AppLovin.

The global technology community agrees that the AI industry is one of the most important technology industries in the future, as well as the development path of infrastructure first and then application.

Therefore, for AI investors, first-hand infrastructure and first-hand application are reasonable portfolios.

03

epilogue

Looking back at the operational level that everyone is most concerned about, first of all, we need to be clear. It is much more rational to invest in AI now than last year. Only by seeing the performance growth generated by specific AI can we go big. By drawing cakes alone, it is difficult to stir up capital nerves as easily as last year.

This determines that when investing in AI now, it's best to pay more attention to win rates rather than simply looking at odds.

From the perspective of pursuing win rate, for a popular company such as Nvidia, a more comfortable trading position must meet the following two conditions:

First, there was a major collective retracement in the US stock market. For example, in July-August, all technology stocks fell. This is because the market cleared itself. The NASDAQ fell 15% in just one month, and Nvidia fell as high as 34%. Like the retracement in April, the NASDAQ fell 8% and Nvidia fell 21%, which is acceptable.

Second, there has been no change in fundamentals. This includes the fundamentals of the US economy as well as the fundamentals of the company. Because economic fundamentals can guarantee an overall rebound in US stocks, corporate fundamentals can guarantee a rebound in individual stocks.

This kind of decline, which has nothing to do with fundamentals, is clearly that the market is creating opportunities, and the probability of a violent backlash later is very high. However, the most active trading companies, such as Nvidia, are often the target of priority capital purchases once the market rebounds.

If you simply apply the above experience, then if Nvidia's stock price can retreat 20-30% in a short period of time, such as 1-2 months, and reach 100-120 US dollars, that is a good buying position.

Don't forget the prerequisite: the fundamentals haven't changed!

However, those A-share computing power concept stocks that have been mapped by Nvidia can also follow the trading strategy described above, provided that the influence of geographical issues is ruled out first.

However, for applied companies, it is possible to lay out ahead of time, especially companies with reasonable valuations and high certainty about future performance, whether it is A-shares or US stocks. Although it is difficult to predict how many times its performance will explode, the stock price may not rise much for quite some time after the purchase, but if you are investing in AI in the medium to long term, there is no problem with the general direction; you can just hand over the rest. This is a simple and easy strategy that saves time, effort, and worry.

Because in the next 10-20 years, there will be endless applications of AI, which will release commercial value and return on investment.

Buffett said that investing doesn't require a high IQ, but it requires a good direction.

Compared to chasing hot spots every day, hard work may not be profitable. It would be better to focus on the general direction of AI, select leading companies, and do more medium- to long-term layouts. (End of full text)