Key Takeaways from the Closed-Door Meeting

In late July, Jiang Cheng of Zhongtai Asset Management participated in a small-scale, closed-door exchange for professional investors. The event was divided into two parts: the first part featured Jiang Cheng reviewing the second-quarter portfolio performance and market changes; the second part focused on addressing investor questions regarding industries, valuations, portfolio management, error correction mechanisms, AI, and team management. To capture the core points efficiently, the author distilled the meeting content into "ten key themes," presenting Jiang Cheng's investment philosophy, operational discipline, industry judgments, risk control logic, and personal development in a single, comprehensive overview, which can serve as a guide to understanding his core thinking. Below are the key takeaways organized by the author based on the meeting minutes.

Second-Quarter Market and Portfolio Operations: Contrarian Rebalancing Amid Extreme Divergence

The market experienced extreme divergence in the second quarter, with the divergence in June being exceptionally rare over the past 20 years. Popular sectors had strong fundamentals but did not meet value investing criteria. The product's maximum drawdown was approximately 12.6%, with no major negative impacts on the long-term core value of the portfolio holdings. The operational discipline adhered to contrarian rebalancing, increasing positions during declines and reducing them during rises, with current positions at a high level for recent quarters. The buy-low, sell-high rebalancing operations contributed about 1 percentage point positively to net value over the year.

Underlying Investment Framework: A Stable Framework, Open Knowledge, and Long-Term Cash Flow as the Core

Stock selection does not differentiate between growth, mature, or declining industries, but instead focuses on the present value of a company's distributable cash flow over its entire lifecycle. Long-term uncertainty is not artificially simplified, with margin of safety used to cover unknowns. If the endgame is unclear or the margin of safety is insufficient, the investment is abandoned. The investment framework and personal capabilities are distinguished: the framework is self-consistent and stable, but research and investment knowledge always have limitations, requiring continuous reflection and error correction.

Market Volatility Mindset: No Anxiety Over Style or Rankings, Only Verification of Holdings' Fundamentals

Market information is divided into two types: optimistic narratives are "fairy tales," while damage to a company's competitive advantage and long-term cash flow is a "ghost story." Declines without negative news represent an increase in potential returns. The core source of anxiety is not net value drawdown or peer performance, but whether one's own judgment of a company's long-term value is wrong. There is no prediction of style rotation or chasing market hotspots, and investment drift is rejected.

Valuation and Position Error Correction Mechanism: No Mechanical Stop-Loss, Adjust Positions Based on Fundamental Evidence

Valuation has no fixed 3- or 5-year horizon, covering the entire lifecycle of a company, using a margin of safety to hedge against long-term uncertainty. There is no uniform stop-loss line based on time or decline. A decline in ROE, profit, or gross margin is only a surface phenomenon; substantial damage to competitive advantage and the present value of long-term cash flow is the signal for reducing or clearing positions. There is no single quantitative indicator to distinguish between short-term cyclical fluctuations and long-term value destruction; this must be assessed comprehensively for each company. There have been past cases of misjudgment due to holding positions too long, and the research and investment response speed is optimized through high-frequency review.

AI and Technology Sectors: Moats Exist, but a Sufficient Margin of Safety is Generally Lacking

The AI chain has real moats, with wafer manufacturing having the highest barriers. The endgame for large models and downstream applications is highly uncertain. Although the AI computing power chain is booming, current industry profit margins are high, and there is significant downside risk from supply-demand reversal. Pessimistic scenarios cannot be underpinned, making it difficult to include in the portfolio. The technology sector is only worth observing when it has deeply corrected and a stable margin of safety can be calculated. Batch buying will not occur just because the sector has declined. AI can only be used to assist in verifying investment logic, not to replicate trading decisions.

Cyclical, Resource, and Traditional Manufacturing: Distinguishing Between Cyclical Trading and Long-Term Value Investing

For cyclical products like livestock and coal, production capacity reduction only brings temporary profit recovery. If the long-term competitive landscape does not improve, they are not suitable for heavy positions. Research and investment capabilities in non-ferrous metals are being strengthened, but there is no current sector-wide bullish conclusion. Targets that simultaneously meet supply constraints and a stable long-term landscape are scarce. For traditional industries like container shipping, water utilities, insurance, tire molds, special steel for machinery, home appliances, and analog chips, some companies have competitive advantages, but there are generally shortcomings in valuation and the certainty of long-term cash flow, so they are not yet in the core holdings pool. Even industries with long-term demand contraction are not entirely uninvestable, as long as the endgame cash flow can be estimated and warrants attention.

Real Estate, Consumer Staples, Utilities, and Dividend Assets: A Rational View of Valuation and Long-Term Sustainability

In real estate, a total industry volume reversal is not expected, and focus is only on leading companies that can continuously collect payments, have manageable existing risks, and are suitable for a low industry scale. For high-end liquor, short-term channel reforms cannot raise long-term return expectations, and long-term demand and competitive landscape remain uncertain. For utilities and dividend assets, static high dividend yields and low valuations do not necessarily represent high cost-effectiveness. The opportunity cost and stability of long-term cash flow must be compared. The long-term returns of A-share and Hong Kong stock dividend indices are similar. Dividend products are not yet included in the personal pension scope. For gold, with no stable cash flow, intrinsic value cannot be calculated, and it is only suitable for analysis within a monetary and safe-haven framework.

Macroeconomic Conditions and Capital Flows: Investment Decisions Are Not Based on Crowded Trades or Fund Flows

Institutional crowding and capital flows are continuously observed, but market capital flow behavior is only used as a reference, not as a basis for buying or selling. Predicting market positions or the timing of style rotations is rejected. Macroeconomic scenario predictions beyond one's circle of competence are directly avoided.

Research Team and Management: Expanding Research Capabilities, Balancing Management and Investment Thinking

The research team is being expanded, with continuous recruitment of equity research and investment talent. Fund manager roles are assigned based on internal mature talent through division of responsibilities, with no core personnel departing. Management tasks encroach on uninterrupted reading time, but fragmented thinking can ensure the priority of investment decisions. Experience in corporate management roles completes the perspective on corporate governance research. Industry coverage is being broadened, with research on traditional cyclical industries like non-ferrous metals and machinery being strengthened, and understanding of the AI industry chain is continuously being iterated.

Learning Methodology and Cognitive View: First-Principles Thinking, Continuous Iteration of Cognition

When learning new things, the approach is to break down what, how, and why, clarifying the underlying definitions of concepts to avoid cognitive divergence caused by vague market concepts. A mindset of open learning is maintained, and new knowledge updates will not immediately change trades, but will optimize the investment framework over the long term. The recommended reading is Rare Earth, using scientific hypothesis testing as an analogy for investment: continuously verify and correct one's own judgments, and acknowledge the inherent limitations of one's own cognition.