SDIC Lin Rongxiong: The clumping of AI technology is not over yet; it’s still too early to “burn the cold stove.”

Full Summary

Core Conclusion: Over the past period, we have repeatedly emphasized that all trends are essentially herd behavior, and the final frenzy is the most frenzied. Considering that the current macro "gray rhino" is not clear and AI capital expenditures are hard to falsify, the current AI tech herd behavior is unlikely to end easily. Its core main theme positioning has not been shaken by high stock price fluctuations. So far, we tend to believe that the current AI tech is more about experiencing a phased left-side rotation from high to low, rather than the final frenzy. Looking at tech, don't dwell on the first peak, prepare for the second peak, persist until the peak, and hold onto the big light and buy the light ring. Based on the industrial wave, the first peak of the final M-top for stocks is likely not yet here.

At the same time, we have always frankly admitted that whether it's the current AI tech, the "New Ning Combination," or the ChiNext Index, as they continued to rise in Q2, the A-share high-to-low rotation index reached a high of 60% by mid-June, hitting the upper edge of the normal range, and the endogenous momentum for rebalancing is continuously strengthening.

Objectively speaking, from the perspective of a phased left-side rotation from high to low, not all low-position sectors have equal recovery elasticity: 1. Consumption, real estate, and Hong Kong stocks (tech) have not seen clear policy or logical signals. Even if there is a rebound, how much upside space there is remains uncertain. Focus on agricultural products that have reversal expectations. 2. In fact, apart from brokerages and high dividends, no other sectors are currently capable of absorbing tech funds. Among low-position cyclical sectors, brokerages have a significant advantage. Among high dividends, focus on coal. 3. Within growth, non-AI growth will take on some of the capital flowing out of high-position AI tech. Focus on overseas new energy and robotics with industry catalysts. In fact, what really needs to be distinguished is: whether the low position comes from short-term style suppression or from medium to long-term downward revisions in fundamentals; whether the cheap valuation is due to mispricing or a reasonable discount after fundamental weakening. Summary assessment: Currently, it is not the time for a broad reversal of low-position sectors. It's too early to systematically "warm up cold stoves," and it's quite difficult to overwhelmingly "scavenge cold stoves."

Objectively speaking, from the perspective of a phased left-side rotation from high to low, not all low-position sectors have equal recovery elasticity: 1. Consumption, real estate, and Hong Kong stocks (tech) cannot rule out that a batch of companies already have clear absolute return space, but there are no clear policy or logical signals. Although overall valuations are low, even if there is a rebound, how much upside space there is remains uncertain, and the sustainability of beta excess is questionable. 2. In fact, apart from brokerages and high dividends, no other sectors are currently capable of absorbing tech funds. Among low-position cyclical sectors, brokerages have a beta advantage. Chemicals and overseas sectors (innovative drugs, new energy, engineering machinery) have a batch of stocks with better-than-expected interim reports. Among high dividends, coal remains the best option. In fact, what really needs to be distinguished is: whether the low position comes from short-term style suppression or from medium to long-term downward revisions in fundamentals; whether the cheap valuation is due to mispricing or a reasonable discount after fundamental weakening. This means that currently, it is not the time for a broad reversal of low-position sectors. It's too early to systematically "warm up cold stoves," and it's quite difficult to overwhelmingly "scavenge cold stoves."

1. After the recent adjustment, dividend assets have regained relatively clear relative value. Focus on coal. Its pricing anchor mainly falls on the premium of dividend yield over risk-free interest rates, not just the dividend yield level. Currently, the dividend yield of the CSI Dividend Index has rebounded to about 5.6%, and the spread over the 10-year government bond yield has widened back to around 400BP, indicating that the attractiveness of dividend assets relative to bonds has clearly recovered. Considering that domestic interest rates remain low and market risk appetite is unstable, dividend assets are more suitable as a source of stable returns and volatility buffers in a portfolio. If overseas tightening concerns re-emerge later, their defensive attributes and hedging value will be more easily repriced.

2. Cyclical resources are currently in a position with elasticity but still need macro verification. Focus on chemical stocks with better-than-expected interim reports. Over the past period, the year-on-year rebound in PPI has provided price recovery clues for steel, chemicals, non-ferrous metals, and some upstream manufacturing sectors. Some sub-sectors (chemical fibers, ordinary steel, energy metals, etc.) have also shown marginal improvement in fundamentals. If factors such as "anti-involution," capacity constraints, and overseas price resonance continue to ferment, pro-cyclical assets may see a linkage between earnings expectations and valuation recovery. The key to pricing this line still lies in the unclear overseas macro situation. If price recovery cannot be transmitted to corporate profits, the low position can only bring phased trading elasticity. Gold has basically clearly seen its historical peak.

3. Although consumption is also at a low position, the recovery pace may be slow. Focus on agricultural products with reversal expectations. Currently, the credit impulse is downward, reflecting weak domestic demand. The constraints faced by the consumption sector are more related to residents' income expectations, willingness to consume, and insufficient credit expansion. Low valuations can explain the current position but cannot directly lead to a strong recovery. Whether consumption can regain capital attention in the future still depends on whether residents' expectations improve and whether data such as retail sales, income, and credit show more sustained signs of stabilization. Before that, consumption is more likely to have structural opportunities, and the basis for a broad beta rally is not solid enough.

4. The low position of real estate more reflects the continued suppression of sector fundamentals and risk appetite. Whether the sector can recover still depends on whether housing prices, sales, and the credit cycle can stabilize. Currently, the process of finding a bottom for housing prices still has fluctuations, residents' home-buying expectations have not significantly reversed, and the profitability and balance sheet repair of the real estate chain lack continuity. In terms of allocation, real estate is more about policy expectations and phased rebound trading, and the certainty should not be overestimated. Only when a more stable positive feedback loop forms among stabilizing housing prices, improving sales, and repairing the financing environment can the sector truly shift from low-position trading to medium-term allocation.

5. The phased recovery of innovative drugs mainly comes from the decline in real interest rates on US Treasuries from highs and recent fund rebalancing. Focus on projects with positive progress at low valuations. As a growth asset with long duration and sensitivity to liquidity, innovative drugs are prone to valuation reactions when overseas interest rates decline or easing expectations rise. However, this type of recovery is currently more phased. Sustainability still depends on two lines: first, whether overseas real interest rates can continue to fall; second, whether industry-level R&D pipelines, commercialization, and overseas cooperation can continue to deliver. If only interest rate conditions improve without fundamental verification, innovative drugs are more likely to show valuation repair rather than a full reversal.

6. The current low valuation of Hong Kong stocks is still suppressed by the US dollar and overseas liquidity in the short term. For Hong Kong-listed internet companies, we are waiting in the medium term for them to make significant positive progress in AI. Against the background of a strong US dollar, the overall beta of Hong Kong stocks is weak. Hang Seng Tech, Hang Seng Index, and Hong Kong high-dividend stocks all find it difficult to completely break free from this constraint. The valuations and dividend yields of Hong Kong stocks have already become attractive, but the main factor behind the current strong US dollar is the weak exchange rates of the Eurozone and Japan, the relatively weaker economy compared to the US, and the rising expectations of Fed rate hikes. Considering that the US economy's driving force comes from AI-related capital expenditures, it may be difficult to reverse the strong dollar trend in the short term.

7. Within growth, AI tech vs. non-AI growth: The deep meaning of the severe divergence between AI and non-AI within growth is that the market's current pricing of "growth" has become very narrow. Focus on overseas new energy and robotics with industry catalysts. AI hardware, computing power, communication equipment, semiconductors, and computer equipment have absorbed the main capital in the growth style, while innovative drugs, new energy, defense, and high-end manufacturing including robotics, even with growth attributes, have not received the same pricing. If high-position AI tech experiences a rebalancing based on crowding, the first to absorb capital may also be non-AI growth, which belongs to the same growth style but has significantly underperformed earlier.

In summary: In fact, we have always frankly admitted that whether it's the current AI tech, the "New Ning Combination," or the ChiNext Index, as they continued to rise in Q2, the A-share high-to-low rotation index reached a high of 60% by mid-June, hitting the upper edge of the normal range, and the endogenous momentum for left-side rotation from high to low is continuously strengthening. In practice, the phased left-side "rotation from high to low" is easily confused with the first peak of the M-top. The key difference is: Although phased rotation from high to low is sometimes accompanied by industry disagreements, in retrospect, it is often unrelated to damage to industry logic. It is more often associated with profit-taking from tech overheating, macro constraints (including regulation), and policy and fundamental catalysts in low-position sectors. Considering that the current macro "gray rhino" is not clear and AI capital expenditures are hard to falsify, the current AI tech herd behavior is unlikely to end easily. Its core main theme positioning has not been shaken by high stock price fluctuations!

In the current environment, our advice is very clear: 3 parts prediction, 7 parts response. The core of the response is not to rush to find low-position sectors, but how to handle high-position tech issues. For handling tech, the core must be to keep an eye on US stocks. Here, we always emphasize three sentences: 1. When the industry wave enters the middle stage, we must deeply understand that "all trends are herd behavior, and the final frenzy is the most frenzied." The pricing of industry trends will go through three stages: trend — herd behavior — frenzy. 2. "The general does not dismount." The essence is to emphasize that before the AI industry trend ends, and before the industry trend encounters the two "roadblocks" of macro "gray rhino" and industry competition structure collapse, even if there is a short-term phased rotation from high to low, the market will still return to the main line of the industry trend later. This is irrefutable and never fails. 3. Facing tech, don't dwell on the first peak, prepare for the second peak, and persist until the peak. So far, we tend to believe that from the perspective of the M-top final ending of the current AI tech based on the industry wave, the first peak is likely not yet here.

Source: Lin Rongxiong Strategy Salon

Risk Warning and Disclaimer

        Market risks exist, and investment requires caution. This article does not constitute personal investment advice, nor does it consider the specific investment objectives, financial situations, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Investment based on this is at your own risk.