A：We focus on investment opportunities in the growth of AI core technology through three roundtable forums, with the first forum delving into investment strategies to seize the opportunities in the AI era. The forum will be hosted by Mr. Zong Jianshu, Chief Analyst of Computer at Changjiang Securities, and will invite several fund managers from China Asset Management Co., Ltd. who are known for their expertise in technology growth and have conducted in-depth research and insights into the AI industry chain.

A：There is skepticism in the market regarding the authenticity of the demand for AI and the prospects for commercial implementation. However, the AI industry continues to develop and bring investment opportunities despite the twists and turns. In particular, computing power infrastructure has been the most certain investment theme in recent years. In the rapidly changing AI industry, how to stay ahead of the curve becomes a key test of investment judgment, as the curve of computing power growth changes, models develop, and innovative scenarios emerge on the application side.

A：From a modeling perspective, there is still significant room to drive the demand for computing power during the post-train and test phases, especially as companies like OpenAI continue to invest and see revenue growth, and with TSMC's optimistic expectations for capital expenditure on computing power in the coming years, indicating that global demand for computing power will continue to rise.

A：Although some people believe that equal computing power can alleviate the problem of insufficient computing power in China, there is still controversy over skin law's role in improving model capabilities. However, engineering optimization and continuous investment in computing power are not contradictory. In the short term, strong capital expenditures from cloud vendors, including TSMC, provide certainty for computing power demand. In the long term, the industry chain may still need to proceed cautiously.

A：The core issue in the next stage is how to solve structural problems in the direction of computing power and make more precise investment layouts. Research has found clear trends in model training paradigms and inference, such as multimodal and long-memory capabilities that will drive the value inflation of transmission links (such as optical communications) and storage requirements. In addition, there may be a shift in the supply-demand balance point in upstream electronic bulk commodities such as increased CPU demand and similar storage chips, which are worth focusing on as investment directions in the future.

A：The entire market focus is shifting from the training stage to the inference stage, and further to the demand for agented (agent-based AI) and physical AI. It is anticipated that there will be a rise in training demand by 2024, followed by inference demand in 2025, agented demand in 2026, and physical AI training demand in the next three to four years. These demands are constantly evolving, with the computational infrastructure required for agented AI and physical AI being much greater compared to the second stage of inference demand.

A：Although the index contains second and third-tier companies, the position of industry leading companies is relatively stable. In the short term, investing in leading companies may underperform the index, but in the long term, excellent leading companies will outperform second and third-tier companies. Therefore, sticking to fundamental research, finding and holding industry leading companies for the long term, has the potential to outperform passive indexes.

A：Active equity funds construct relatively fewer selected portfolios through careful selection, mainly investing in targets of core supply chain companies where excess returns have been concentrated in the past few years. Even in the face of the rapid changes in the technology industry, active fund managers can capture excess returns and alpha by thoroughly studying changes in industry logic and narrative reversals.

A：This concern stems from the extremely high barrier to entry in the industry, requiring a large amount of capital investment, which only a few players can bear and participate in the competition. The current environment is different from the Internet technology bubble period, and the risks and returns around AI investment are now the key considerations for both businesses and investors. In addition, the cross-shareholding and massive investment between major companies are aimed at seizing computing power resources and technological iteration advantages, rather than simply pursuing leveraged investments.

A：I believe that the incremental demand for AI will have a significant impact on the non-ferrous and other energy power metal industries. From the perspective of supply and demand, next year, over half of the incremental demand will be driven by AI, including the usage related to data centers and upstream energy infrastructure. Currently, related assets have shown good investment returns in the past two years, and with the growth of AI demand, the correlation with AI-related stocks is gradually increasing. In the future, new investment opportunities can still be found in this major trend.

A：There is indeed a mismatch issue because the main investors in the power grid are mostly the government or private entities, while the expansion of computing power in data centers is driven by internet companies. There are differences in understanding and judgment of power demand between these two parties, which may become more apparent in the future as the demand for AI continues to increase.

A：The demand for copper is growing rapidly, but the demand for aluminum related to data center cooling is also increasing rapidly, and the aluminum production process consumes a lot of electricity. When facing power shortages, balancing the power demand of data centers and aluminum electrolysis production may create a new supply and demand challenge, especially in the coming years, the importance of this issue is becoming increasingly prominent.

A：This year, the most anticipated project is the large model trained by the NVIDIA GB series, which is expected to be released between the first and second quarters. Model technology is experiencing accelerated changes, from large models to native multimodal models. The standards for large models will become more complex and diverse, and evaluating model capabilities will not only be based on competitive rankings, but also on the ability to solve specific problems, cost efficiency, real-time online learning, and post-train training for adapting to different tasks. In terms of domestic models, significant progress is expected to be made this year in areas such as long-text reasoning and full-modal pre-training.

A：The future basic model ecosystem will be relatively stable, with fewer model manufacturers on the top ten list. However, each manufacturer will not engage in completely homogeneous competition; each will have its own advantages and characteristics. For small and medium-sized enterprises and second-tier internet companies, they can choose to use existing large model capabilities as the base model, and develop their own vertical applications and small models on top of it to reduce costs and increase efficiency. Overall, the large model ecosystem will present a diversified competitive landscape, no longer just homogeneous competition.

A：The applications of AI have gradually unfolded in many scenarios, especially at the consumer end, where large model software installed on smartphones will become important personal assistants. Starting in 2026, many tedious applications will be replaced by AI, such as code generation in the programming field, which has already been largely completed by AI. In addition, there is a significant amount of manual cost in processes such as American credit card billing and loan assessment, and these financial industry scenarios are expected to be rapidly replaced by AI, thereby increasing labor productivity.

A：Although the development of AI technology may squeeze out other IT expenses, in the long run, companies with top large models have higher investment value, such as Alibaba. However, for those companies that mainly rely on AI investment, their business models may be affected, so I personally do not hold a very optimistic view of such companies.

A：Looking at the AI computing power infrastructure related fields, it is expected to continue to outperform market expectations in the first half of the year. In the second half of the year, with the release of new production capacity, supply and demand balance may change. At the same time, edge devices such as robots are also a long-term focus. Although short-term certainty is not high, they are expected to become a high certainty and large investment direction in the long term.

A：Recently, the A-share market has been experiencing a lively Spring season trend. The underlying driving force behind this is the scarcity of supply and demand far exceeding market expectations, especially in the AI computing power infrastructure sector. This includes aspects such as storage, optical chips, and copper foil production capacity, all of which are extremely tight. This has led many companies to maintain high-speed growth in EPS or even exceed expectations. However, in the second half of the year, with the release of new production capacity, the supply and demand situation may improve, which could have a certain impact on market valuations.

A：Now the macro environment has undergone significant changes compared to the past. In the past, it was an investment-driven economy, with a large proportion of real estate and fixed asset expenditures, leading to significant cyclical fluctuations. However, since 2021, the proportion of investment-driven industries in the macro economy has gradually decreased, reducing their impact on the economy and weakening economic cyclical fluctuations. Now, more attention is focused on changes in industrial structure and the impact of monetary cycles, similar to the U.S. market, where market fluctuations depend more on monetary policy cycles.

A：In the current context of abundant liquidity and the transformation of old and new sources of energy in the industrial structure, it is more important to look for structural opportunities while considering the reduction of macroeconomic risks, such as the diminishing negative impact of real estate and debt on the economy. Therefore, market risk preferences are increasing, and investors can focus more on directions with larger long-term growth potential.

A：I think short-term performance may largely be due to luck, while long-term performance can truly show one's abilities. The reason why the funds I manage show strong explosiveness is because the industries I deeply cultivate, such as semiconductors and military industries, happen to benefit from industrial policies and market hotspots, such as the commercial aerospace sector. Grasping hotspots depends on long-term cultivation and understanding of the industry, rather than blindly chasing short-term hotspots.

A：During the bustling spring market, it is important to stick to one's competence circle, focus on the long-term areas one has been cultivating, seize investment opportunities that appear in one's own field of expertise, rather than actively chasing all the hotspots. For example, when industries such as manufacturing, machinery, or power equipment experience a market trend, make sure to accurately grasp and seize these opportunities.

A：I am very optimistic about the off-shore manufacturing sector in 2026, it is a long-term trend. From a timing perspective, some companies have been laying out their off-shore strategy many years ago and now their performance is starting to explode. China's off-shore manufacturing is a gradual process, starting from simple products to complex products, which will create more investment opportunities. The two directions I am most optimistic about next year are: first, capital goods in the European and American industrial chains, especially as the European economy has greater margin elasticity; second, machinery and vehicles in the South American region, due to strong market demand and better than expected economic performance.

A：There are mainly two ways to avoid this kind of risk. The first is to choose companies with sufficient global presence, such as in the technology sector, where giants like Apple usually help solve related issues if a problem arises. The second is to look for investment opportunities in regions with lower geopolitical risks, such as Africa, Southeast Asia, and parts of Europe.

A：Companies should try to find businesses with diversified global layouts, rather than just focusing on a single market and sector. This way, even if one market experiences fluctuations, the overall risk will be relatively manageable. For example, some companies in the engineering machinery industry have their business operations diversified and low exposure to individual countries, so the impact of risks in a single country is minimal.

A：Many Chinese manufacturing companies that were originally considered traditional and domestic-focused have gradually diversified their business structures and moved towards globalization in the process of going international. During this process, they not only expanded their EPS growth, but also experienced changes in their valuation systems. For example, a leading heavy truck company saw its valuation increase from single-digit PE ratios to around 10 to 20 times due to opportunities in the field of artificial intelligence.

A：The electric power equipment industry is a comparative advantage industry in China, possessing world-leading grid technology and related technical expertise. With the growth of electricity demand, especially driven by AI and increasing energy storage needs, more electric power equipment is needed as a buffer zone between the grid and AIIDC, providing the electric power equipment industry with opportunities for certain growth. As for the non-ferrous industry, its prices are influenced by various factors including interest rates, demand, supply, geopolitics, etc. In the current environment, it is important to pay attention to factors such as US interest rate cuts, supply and demand relationships, and technological advancements affecting metals like copper.

A：Gold is one of the most difficult commodities to predict globally, making it difficult to accurately judge its future trends. Although financial attributes such as the US dollar credit crisis have brought widespread attention to gold, the main issue with holding gold is that it does not bear interest and does not have continuous growth value. However, from an alpha perspective, some gold-listed companies with resource expansion capabilities and mining acquisition advantages, such as trillion-level leading non-ferrous metal companies, may outperform the price of gold itself due to continuous integration of mineral resources and overseas expansion.

A：From the perspective of industrial trends, semiconductors can be divided into two directions: cyclic and growth. The cyclical direction mainly focuses on the scarce global memory sector. Based on an understanding of the progress of models, especially from the perspective of technical difficulty, the current development of AI has created a very strong demand for memory. For example, new technologies such as HBM continue to emerge to meet the cost and efficiency requirements, which will make the memory cycle market stronger and longer-lasting. In addition, derivative opportunities include the application of new technologies in domestic computing power semiconductor equipment and storage.

A：The level of institutionalization in the Hong Kong stock market is relatively high, making it a global offshore allocation market. However, there is limited stable local capital, leading to significant fluctuations in Hong Kong stock valuations. When the market is relatively cheap, it may be suitable to hold onto Hong Kong stocks for the long term. Currently, favorable factors include an increase in Hong Kong IPO companies providing more investment targets, as well as opportunities to invest in high-dividend state-owned enterprises and tech giants. In particular, tech giants can better reflect the domestic demand economic situation.

A：In 2026, the main source of income for fixed income products comes from the equity portion. The interpretation of the stock and bond markets this year is worth paying attention to. Overall, the stock market does not pose a significant risk and can maintain some expectations. However, it is necessary to pay special attention to the volatility and risk of overseas markets. For fixed income products, asset scarcity should be emphasized in allocation, while controlling positions and seizing the right timing to allocate unique assets in order to achieve a better risk-return ratio.

A：I believe that, considering the prices of assets running at high levels in the long term, if further upside is to be supported, there will be a high demand for monetary and fiscal policy space in the United States. Therefore, my views on this are relatively conservative.

A：This year I tend to pay more attention to overseas market fluctuations, especially in the stock market. As for the bond market, I believe that current yields are trending upward, mainly due to the revision of overly pessimistic expectations for the Chinese economy and capital markets. Given the current weak economic growth, which requires a low interest rate environment for support, my view on the bond market this year tends towards neutral, and I suggest balancing stocks and bonds when appropriate opportunities arise. I am relatively optimistic about the bond market this year. Considering the country's active participation in forward-looking investments in the global competition and stabilizing the economy at the grassroots level through fiscal and monetary hedging. With the current high general deficit rate and significant policy resource allocation, the economy may continue to maintain stability in the future. The limited upside potential for bond yields, coupled with scarce loan resources, makes bond yields closely linked to loan yields. Therefore, bond yields are not likely to rise easily, and with downward pressure on financial institutions' borrowing costs, bonds become a relatively low-risk but potentially low-yield asset class.

A：In the current supply and demand structure, there is a shortage of short-term bonds and an excess of long-term bonds. However, from the perspective of overall asset supply, investments in risk-free interest rate products should prioritize yield, focusing on purchasing those with higher yields. For complex products such as credit bonds, choices should be made based on a combination of risk tolerance and credit ratings.

A：From a historical perspective, the performance of the equity market is often related to the tightening of policies. This year, both domestic and foreign policies are generally loose, which is favorable for the performance of the equity market. We should pay more attention to the level of EPS (profitability) and, based on past experiences with valuation and performance, while maintaining an optimistic attitude, focus on investment opportunities at the EPS level.

A：The core source of excess contribution in the past two years comes from the non-ferrous metal industry. Our layout in the non-ferrous sector began three years ago, focusing on its long-term strategic value. In different market stages, we try to avoid tactical errors as much as possible, although sometimes we may miss short-term opportunities, but overall we are overweighted on the strategic level.

A：In investing, we need to deeply consider the strategic background of the current era and look for industries or sectors with long-term excess returns. For example, the non-ferrous industry has shown excess returns for three consecutive years, and its market performance has been sustained for a long time, reflecting underlying macro trend changes.

A：Since the outbreak of the epidemic, the world has entered a period dominated by fiscal policies, which is different from the previous era dominated by currencies. The increase in physical demand, especially in industries such as non-ferrous metals, has had a positive impact. In addition, the rise of resource populism has led to a fragile future supply, and the restructuring of demand at the industry level (such as energy transformation, AI industrialization, etc.) has replaced the demand for real estate and infrastructure, all of which have had a positive impact on the non-ferrous industry.

A：Currently, the valuation of convertible bonds in the market is relatively high, but in essence, it is highly correlated with stocks and is a reflection of equity assets. If the stock market is doing well, convertible bonds will also perform well, and usually have a lower volatility coefficient in the short term market. Convertible bond assets themselves have the characteristic of a high Sharpe ratio, meaning a high annual return rate and low drawdown, due to the strong demand for convertible bonds, high conversion probability, and the ease of achieving deterministic returns in high activity levels. The current high valuation of convertible bonds is based on their asset characteristics and the potential need for inflation hedging and technology theme asset volatility increase in the future, which overall seems reasonable.

A：I think the current overvaluation of convertible bonds is a reasonable phenomenon, it reflects the trend of declining risk-free interest rates in the entire society and the scarcity of assets. Due to the strong liquidity expectations, everyone is pricing call options more actively, so although the current valuation is expensive, the market has not yet reached an extreme state, and investment opportunities still exist.

A：This year, we are optimistic about the two major industries of AI and non-ferrous metals, and we will focus on and extend our efforts in these areas. For example, in the field of AI computing power, we will focus on new technologies and domestic supply chains. In terms of applications, we are bullish on internet giants with large-scale model layouts and the robot industry. In the second half of the year, there may be opportunities after the slowdown in storage price increases in the fruit chain. In addition, in the AI field, we are also optimistic about semiconductor equipment, chip companies, gas turbines, and power equipment related to rising storage prices. In the non-ferrous metals industry, in addition to gold, copper, aluminum, and other varieties, we will also focus on small metals with disrupted supply chains and new technology fields such as solid-state batteries that show clear industrialization trends. At the same time, we are also focusing on non-bank, consumer-related companies, and companies related to the export chain that have potential for growth. The coal industry has a relatively high investment success rate during certain periods.

A：Industry rotation mainly revolves around three dimensions: firstly, industry outlook and expectations, which are the core factors determining the success rate of investments; secondly, valuation levels and corresponding chip structures, which will affect the investment odds; thirdly, the macro policy direction at the national level, as international and overseas policies also have a significant impact on equity investments. We will focus on tracking industry outlook forecasts, combine them with an industry outlook tracking framework for strict monitoring, and pay attention to opportunities for industry bottom reversals.

A：In the management of fixed income plus products, we emphasize the transmission of long-term investment methods rather than short-term asset trading. First, we look for assets with low volatility in the long term, emphasize the importance of asset allocation, and follow the asset intercorrelation based on statistical rules. Secondly, we ensure that all types of assets are purchased at a sufficiently cheap price to improve the portfolio's Sharpe Ratio. In specific operations, we will adhere to the effectiveness of the methodology, even if it requires enduring market fluctuations in the short term, we must believe in the method and flexibly respond to various considerations in the daily market.

A：I believe that first of all, we must adhere to the principle of "don't do what you don't understand", and then balance the concentration of market funds and long-term trends. For assets with high short-term fund concentration, although they may attract a lot of attention and potentially bring quick profits, the long-term expected return rate may decrease. Therefore, we need to find a balance and maintain patience while waiting for valuation regression. At the same time, we need to understand the market trends in different areas as much as possible, so that we can be prepared to participate in the market when specific assets experience fluctuations.

A：Investing can be likened to buying candles when the power goes out in a neighborhood. If a neighborhood is about to experience a widespread blackout, the demand for candles will increase dramatically, causing prices to rise in the short term. In this scenario, it is likely a good idea to participate in the candle market if there is a significant amount of capital involved. However, if in an unfamiliar field (like predicting power outages in Chongqing as a Shanghai resident), due to limited marginal knowledge and difficulty in obtaining information, one should not blindly participate. Instead, one should make investment decisions within their own abilities and knowledge.

A：In asset allocation, the first thing to ensure is comprehensiveness, to understand as many types of assets as possible, so that opportunities can be seized when specific events occur in specific assets. However, human energy and time are limited, and it is impossible to make accurate judgments on all sudden events. Therefore, it is necessary to balance long-term value and short-term bubbles, maintain a mentality of weakness, avoid blindly chasing hot spots, such as semiconductor investments, while adhering to long-term investment concepts, such as value investment, to convey a stable asset allocation strategy to holders.

A：Retreat control mainly starts from three aspects: first, asset allocation management of major categories, especially strict control of equity asset positions in low-volatility fixed-income products; second, the selection of equity assets, purchasing companies with core competitiveness at reasonable valuations, and cashing in profits when valuations are too high to reduce the volatility of single equity assets; third, the duration of bond investments also needs to be adjusted according to the macro environment to reduce the retreat contribution of bond assets.

A：The similarity lies in the pursuit of long-term compounding returns, with a consistent underlying logic of avoiding overpriced stocks. The difference lies in adjustments made in equity allocation based on different customer risk preferences and product positioning. For example, cyclical themed funds may be more aggressive with higher volatility, while equity funds focused on overall market value direction may be more stable. In secondary bond funds, the equity portion may be more conservative to match the needs of customers seeking stable income.

A：Our company places a strong emphasis on teamwork, encouraging team members to each bring their own strengths and contribute to each other. Through cross-departmental cooperation, we have created a work atmosphere that is open, gentle, and willing to communicate. This not only helps stabilize performance, but also promotes the personal growth of fund managers. The company encourages teamwork from both the human resources system and the assessment system to help team members grow together and realize their individual value.

A：This year, the risk of fixed income products turning into reduced fixed income products is relatively small, but it has indeed occurred in history. I believe the key lies in adhering to the top-down asset allocation methodology. Through asset allocation, in the current monetary and financial environment, this methodology helps to obtain stable returns, like a "free lunch". To achieve this goal, one must first believe in and adhere to this methodology, and purchase various assets at relatively cheap prices as much as possible during the investment process. At the same time, one must acknowledge that everyone has knowledge blind spots, so it is necessary to broaden one's knowledge base as much as possible, diversify assets within one's area of expertise, and ensure investments are made in high probability positive return situations. In terms of specific operations, flexibility in adjusting strategies under different market conditions is necessary to ensure success in various niche areas and achieve stable absolute return goals. In preventing fixed income products from turning into reduced fixed income products, I believe it mainly relies on two points: first, adhering to a fixed investment methodology, and second, ensuring that the purchase prices of various assets are relatively cheap during the investment process. At the same time, one should strive to be knowledgeable in different investment areas and avoid situations where all investment areas experience negative returns in a specific year. For example, in bonds, convertible bonds, stocks, and other asset categories, one should seek and seize high probability positive return opportunities, effectively managing each investment area to achieve stable absolute return goals.