This year's return is 44%. Ping An Fund's Lin Qingyuan: The trend of AI remains unchanged, focusing on the layout of AI power

In a recent live broadcast on CCTV Finance, Lin Qingyuan, the fund manager of Ping An Fund, visited the studio to provide a forward-looking analysis and judgment on the future investment trends in the AI field. The investment representative summarized the key points as follows: Large models have entered a stage of implementation, and AI is truly transforming into productivity. For optical modules and PCBs, we believe these industry trends are still continuing. However, the bottleneck may be shifting from computing power to electricity. The conflict between the exponential growth of computing power and the linear growth of AI electricity supply is becoming increasingly prominent. This situation is just beginning, and orders are already scheduled until 2030, making it the most scarce in all chains. (Many stocks related to power equipment exports) To be honest, stock prices may have slightly outpaced the fundamentals. Many companies may have only 30% to 50% of their overseas revenue, and valuations may have already exceeded a hundred times compared to last year. However, the growth rate of performance can digest the valuation, and noticeable changes may be seen by mid-year. Additionally, A-shares are often influenced by overseas technology mapping. As overseas giants continue to rise, it naturally raises domestic valuations. This year is still focused on the layout of AI electricity. Although many have risen, with the entry of companies with a market value of hundreds of billions, the track has widened, and valuations are not that high. Lin Qingyuan, fund manager of Ping An Fund, has 14 years of experience in the securities industry and previously served as the fund manager of the equity investment department at Rongtong Fund Management Co., Ltd. He joined Ping An Fund in February 2023, managing a fund size of 3.1 billion yuan. His representative work, Ping An Dingyue Mixed (LOF), has achieved a return rate of 187.09% since he took office on November 13, 2023, showing outstanding performance. Another product he manages, Ping An Xin'an Mixed, also performed remarkably: with a return rate of over 52% in the past three months, over 70% in the past six months, over 107% in the past year, and over 44% year-to-date, it ranks 6th and 7th among more than 2,000 similar products. In the quarterly report of Ping An Xin'an Mixed, Lin Qingyuan made a forward-looking judgment on the trends in the AI industry, pointing out that the current investment logic in the industry is undergoing a profound paradigm shift—"the end of computing power is electricity, and the second half of AI is energy." He analyzed that in the past year, the market's main line focused on computing power and chips themselves, but with the exponential growth of AI computing power demand, the linear upgrade of electricity infrastructure is forming structural contradictions. This mismatch is particularly prominent in overseas markets, especially in North America. The average service life of the North American power grid has reached 50 years, and there is a significant gap in electricity supply. Against the backdrop of hollowing out of local manufacturing and aging infrastructure, it is difficult to cope with the explosive energy consumption demands brought by AI data centers The bottleneck of electricity is becoming a key shortcoming that restricts the development of AI. Based on the above judgment, Lin Qingyuan has focused on the A-share market in the fourth quarter, particularly in the power equipment and new energy infrastructure sectors that possess global competitiveness, to capture structural opportunities brought about by changes on the energy side. During this exchange, Lin Qingyuan's views remained unchanged. He reiterated that the conflict between the exponential growth of computing power and the linear growth of AI power supply is becoming increasingly prominent. This year's investment theme will still revolve around "AI + electricity." He admitted that technology investment is the most challenging, but the logic of energy investment is the clearest. The following are the highlights organized by the investment workbook representative (WeChat ID: touzizuoyeben), shared with everyone: The industrial trend of AI continues, large models are entering the implementation stage, and AI is transforming into productivity Host: You mentioned in the program that you previously focused mainly on the line of AI plus computing power, such as optical modules, PCBs, etc. The fund's returns were quite good last year. However, in the second half of last year, you wrote a report on AI electricity and seemed to have made a PPT that mentioned companies like Yingliu Co., Ltd. These stocks have indeed performed very strongly from last year to this year. So, from this structural differentiation, how do you assess the current landscape of technology stocks? Lin Qingyuan: Overall, I believe the industrial trend of AI continues. To be frank, AI has been developing for three years. In the past two years, the market may have been questioning whether it was a breakthrough from 0 to 1, but now I believe that the stage from 0 to 1 has passed, and we are in the process of moving from 1 to 10. This can be viewed from two dimensions: First, last year, some optical module companies in the A-share market had profits exceeding 10 billion, which indicates that they have indeed realized their performance. Second, AI is indeed transforming into productivity. For example, during the holiday period, one could see that the token numbers of those AI large model companies in the Hong Kong stock market are rapidly increasing. Therefore, I believe that large models have entered an implementation stage, and AI is truly transforming into productivity. But why is there a transformation this year? Because the demand for AI computing power continues, such as optical modules and PCBs, we believe these industrial trends are still ongoing. However, the bottleneck may be shifting from computing power to electricity. The issue of electricity bottlenecks is becoming increasingly prominent. The electricity gap may not be felt significantly domestically because our electricity infrastructure is the strongest in the world. However, overseas, this problem is very prominent. For many years, the development of manufacturing overseas has been slow, and electricity demand has not grown much. The sudden arrival of AI has broken the original supply-demand balance. Electricity shortage has become a global topic Host: So, you just mentioned that when we talk about artificial intelligence and AI now, we have crossed the stage from 0 to 1, and now we are in the stage from 1 to N. In this process, a large amount of new demand is emerging. Recently, there have also been some interesting statements, such as Musk mentioning that the ultimate possibility of US-China AI competition may be electricity Some people believe that the breakthrough point or winning hand for AI investment lies in the focus on power infrastructure. What do you think of these views? Lin Qingyuan: I completely agree. Elon Musk makes this statement from an overseas perspective because he realizes that the bottlenecks abroad are very prominent. There is actually a saying that makes a lot of sense, which is the principle of first principles: the end of computing power is electricity. Everyone can think about it, whether it is tokens or the final output results, they are essentially information transformed from electricity processed by chips. All sources, the upstream, are actually energy, all are electricity. Host: So this expectation of electricity shortage seems to have become a global topic now. Lin Qingyuan: It’s not an expectation, but a reality that is actually happening. Looking at overseas tech giants, for example, the CEO of Microsoft has said that we are not lacking chips, but electricity. Without powered data centers, servers cannot operate normally. This is the dilemma they face. Moreover, their power grids are aging very seriously. Power Industry Chain: Generation, Transmission and Distribution, Transformers, Servers Host: Next, let’s break down the power industry chain related to AI and see which specific links have opportunities, as there must be a sequence. First, is it at the generation end? Lin Qingyuan: Let’s sort out the entire chain. First is generation, then transmission and distribution, and transformers are key. Transformers are like rivers branching out, reducing high voltage electricity step by step to the low voltage needed by chips. When it comes to data centers, there are also transformers in the cabinets. Finally, there are servers. Currently, the most critical link is generation. Abroad, for example in North America, it mainly relies on natural gas, so the gas turbine segment is the most constrained. The top three gas turbine manufacturers in the world have all seen their stock prices increase more than tenfold in the past three years. People may pay more attention to NVIDIA, but the growth of these traditional power generation giants is actually not inferior. Host: This might reflect in the domestic market as well. Lin Qingyuan: Yes. Host: Recently, I have also been reviewing and found that investments over the past two years may have been more concentrated on large models and AI computing power infrastructure construction, neglecting the power infrastructure segment. Lin Qingyuan: And this segment happens to be China’s strongest area. The development of AI may change the price-to-earnings ratios of software companies by dozens or even hundreds of times, while traditional industries remain at a dozen times Host: Yes, it has the most advantages. Previously, the demand for these companies was mainly domestic, and their price-to-earnings ratios were only a dozen or twenty times. Now, the valuation suppression may be being lifted. Lin Qingyuan: When people look at tech companies, they always feel that software companies should naturally have price-to-earnings ratios of dozens or even hundreds of times, while traditional industries can only have a dozen times. However, with the development of AI, this situation may be changing. Recently, overseas software companies have actually been declining because the substitution effect of large models on them is very obvious. Host: For example, the pressure from AI programming. Lin Qingyuan: Yes, it’s the impact brought by AI programming. Host: While equipment companies have instead risen Lin Qingyuan: From a global perspective, the valuation of traditional industries is actually increasing, which is quite interesting. Host: You just mentioned gas turbines, which may remind everyone of traditional energy sources like coal or natural gas. What type of energy do you think will meet the power demands of data centers in the future? Lin Qingyuan: In the long term, we definitely need to rely on new energy sources, such as green energy. But distant water cannot quench a near thirst. The electricity consumption scenario of data centers is different from traditional electricity usage; it requires stable power supply 24 hours a day, with high demands for base load energy. From an overseas perspective, natural gas is the most suitable, which is why gas turbines are currently in such short supply. Moreover, once gas turbines are produced, the deployment time is short, and tech companies cannot wait. In China, there are mainly two types of companies in the gas turbine sector: one type manufactures complete units, such as Dongfang Electric; the other type produces components, such as blades and castings, which have high technical barriers. Host: So the entire industry chain from gas turbines is being uplifted. These companies may not have seen much growth for five to ten years. Lin Qingyuan: In the past, people thought these companies were just "blacksmiths" and couldn't command high valuations. Host: A ten to fifteen times valuation would be the ceiling, and this state may have lasted for eight to ten years. But this time, many companies have increased their market value from two to three billion to a hundred billion or even higher in just one year. What do you think about this change in investment logic? Lin Qingyuan: I think there is a rational basis for it. For example, overseas gas turbine manufacturers now have market values close to 2 trillion RMB, increasing tenfold. The main reason is scarcity; the bottleneck is too obvious. Additionally, the profit margins in traditional manufacturing may only be three to five percent, but if a certain link becomes scarce, the profit margin elasticity can be very large. For instance, if the net profit margin increases from three percent to six percent, it effectively doubles the profit. Another point is that customers will have a high acceptance of prices. The tech giants are planning to invest 700 billion USD this year, and power generation equipment accounts for a small proportion of that, but it is very urgent. The Biggest Investment Opportunity Comes from Transformers Host: Besides the power generation side, what do you think about the transmission and distribution and grid renovation segments? Lin Qingyuan: China is very strong in this area, with a nationwide grid. However, overseas, the demand chain brought by AI may be shorter. Data centers will build their own power stations, and with transformers directly connected to the machine rooms, this may become the future standard. So the biggest investment opportunity may lie in transformers. Some Companies' Stock Prices Are Ahead of Fundamentals Host: So recently, people may be paying more attention to stocks related to transformers and substations. However, many stocks related to power equipment exports have already risen significantly, with some even tripling or quintupling. What do you think about the current valuations in this sector? Lin Qingyuan: To be honest, stock prices may be slightly ahead of fundamentals. Many companies may have overseas revenue accounting for only 30% to 50%, and their valuations may already be over a hundred times compared to last year. However, the growth rate of performance can digest the valuation, and we may see significant changes by mid-year. Additionally, A-shares are often influenced by overseas tech mapping. As overseas giants continue to rise, it naturally raises domestic valuations Host: So the valuation of technology stocks is going down, while the valuation of traditional industries is recovering upwards. Lin Qingyuan: Yes, this is a very interesting change. In recent years, I have a profound feeling that the speed of technological development is accelerating. What was once considered a deep moat may quickly be iterated away. Therefore, it is essential to keep up with the changes of the times. AI Power May Become the Biggest Bottleneck Host: Just like this switch from computing power to electricity, many people didn't expect it. High technology suddenly shifted to the most basic field of electricity. Lin Qingyuan: There is a very interesting example. There was a company that used to make stainless steel pipes, and its price-to-earnings ratio was only ten times. But it also supplied steel pipes for gas turbines overseas, and its price-to-earnings ratio rose from ten times to twenty times, with the stock price doubling first. Host: The switch from computing power to electricity still accompanies different industrial cycles. However, in any major cycle, there will be phase risks. What do you think about the capital expenditure of tech giants this year? For example, $700 billion, will the actual expenditure end up being less? Lin Qingyuan: Currently, the progress of overseas tech giants is still in line with expectations. At least this year, this money will still be invested. From a risk perspective, AI power may become the biggest bottleneck. Even if procurement in other areas is delayed, the electricity issue must be resolved first. Therefore, even if capital expenditure encounters obstacles, the bottleneck of electricity still needs to be addressed. Host: So will your investment portfolio reference overseas mapping? Lin Qingyuan: Yes, global technology is in a state of competition and cooperation. For example, optical modules also supply Nvidia and Google, so looking at the stock prices of the final customers can also reflect on domestic companies. Host: Right, mutually confirming and strengthening your judgment. Lin Qingyuan: Moreover, they will announce backlogged orders. For example, the top three global manufacturers have orders lined up until 2030, and they have already paid deposits. So this segment is very tight. The Contradiction Between Exponential Growth of Computing Power and Linear Growth of AI Power Supply is Becoming More Pronounced Host: Looking back at your investment in tech stocks over the past few years, how would you divide it into stages? Lin Qingyuan: Starting from large language models, the first stage is the corpus; the second stage is optical modules and PCBs; the third stage is AI power. Host: Did you buy AI power stocks for the first time based on logical deduction or was it triggered by an incidental event? Lin Qingyuan: It might be more like the latter. But this is also related to my research approach. As mentioned earlier, I tend to track technological trends over the long term. In fact, even before Musk made this point, about two or three years ago, Siemens Energy's stock price had already started to rise. Siemens Energy is a well-known overseas leader in gas turbines. However, at that time, this bottleneck issue was not yet prominent. The key is that, taking artificial intelligence as an example, we all know its growth model is exponential. However, the growth of resources in the real world is often linear. This gap between exponential and linear growth will inevitably widen. This is precisely the core contradiction currently facing the overseas market: the conflict between the exponential growth of computing power and the linear growth of AI power supply is becoming increasingly pronounced Host: In this situation, do you think it's just the beginning? Lin Qingyuan: Yes, and the orders are scheduled until 2030, which can be said to be the most scarce in all chains. Host: For these companies in A-shares that supply parts to overseas giants, are their orders also scheduled until 2030? Lin Qingyuan: The orders from the manufacturers of complete machines are indeed scheduled until 2030. This reflects the advantages of Chinese manufacturing. Overseas, there is a lack of skilled workers, and these experienced craftsmen have become the most scarce resource. Host: A gas turbine is such a huge item; it takes several months to build one, and it cannot be mass-produced. Lin Qingyuan: Correct. Host: So the contradictions this year will be very prominent. Lin Qingyuan: Yes. This Year Still Focused on AI Power Layout Host: So can we buy power equipment companies this year? Lin Qingyuan: A combination can be made. On the power generation side, buy gas turbine complete machines, then buy some blades and castings, and for transformers, buy the leading companies, along with some non-ferrous metals. Host: What does non-ferrous metals have to do with this? Lin Qingyuan: The blades of gas turbines must withstand thousands of degrees of high temperature, so they must be mixed with non-ferrous metals to enhance heat resistance. The domestic industry is indeed constrained in this aspect. Host: We used to think that the rise in non-ferrous metals was due to geopolitical factors, but now it seems there is actual demand as well? Lin Qingyuan: Both. For some metals, the demand driven by AI is even higher. Host: So we can't just look at large varieties like copper and aluminum; we need to look at tungsten and tantalum as well. Lin Qingyuan: Correct. Host: For example, Dongfang Tantalum has performed well last year. Lin Qingyuan: Tantalum capacitors are also used in servers. The market capacity was small before, and the valuation was reasonable, but now with new demand emerging, the space has opened up. Host: Can you give a few examples of non-ferrous metals? Lin Qingyuan: Tungsten has risen sharply, tantalum is also continuously optimistic, and there's germanium. Host: This year your return rate is already 40%. Is there a high probability of adjusting your portfolio? Lin Qingyuan: This year is still focused on AI power layout. Although many have risen, with the entry of companies with a market value of hundreds of billions, the track has widened, and the valuation is not that high. Technology Investment is the Most Difficult, but Energy Investment is the Clearest Host: What advice do you have for small and medium investors? Lin Qingyuan: Technology investment is the most difficult, but energy investment is the clearest. For example, buy an ETF for power grid equipment. Host: Cross-industry, involving machinery and non-ferrous metals. Lin Qingyuan: Yes, I have created an AI power index myself. Host: The direction remains unchanged this year? Lin Qingyuan: Yes, AI plus non-ferrous metals. Host: Are you still paying attention to semiconductors and robotics? Lin Qingyuan: I am also paying attention to technology, but AI power has the strongest certainty. Source: Investment Workbook Pro, Author: Wang Li For more insights from industry leaders, please follow↓↓↓ ![](https://wpimg-wscn.awtmt.com/e60cb70c-8975-4774-9ad5-94191391987c.png? imageView2/2/w/640) Risk Warning and Disclaimer The market has risks, and investment should be cautious. This article does not constitute personal investment advice and does not take into account the specific investment goals, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article align with their specific circumstances. Investment based on this is at one's own risk