Shenwan Hongyuan Group: A rebound is expected in September, but the rebuilding of a major wave market will take longer.

Shenwan Hongyuan Group's research report states that the global stock market remains sensitive to disruptions related to U.S. Treasury bonds. The underlying mid-term issues behind rising U.S. bond yields objectively exist and continue to suppress the global stock market in the medium term. The primary contradiction in technology pricing has returned to the matching of industrial logic and valuation, and the performance of computing power inflation is bound to differentiate in the future. The growth rate of 2027 is likely to exceed that of 2026, a trend that may show differentiation and strengthening in Q4 of 2026. The bank maintains its judgment that the rebound in September will continue, with a structural focus on finding directions in the oversold rebound that can restore losses to a greater extent. Within the AI chain, attention is paid to non-institutional heavy positions, while domestically produced computing power chains and small AI stocks provide "mid-term protection for the short term"; medium-term focus remains on storage, high-end CCL, PCB, and capacitors in the computing power inflation segment. The main points of Shenwan Hongyuan Group are as follows: 1. The global stock market remains sensitive to disruptions related to U.S. Treasury bonds: There are objective mid-term problems behind high U.S. Treasury bond yields (weak U.S. fiscal discipline, global reserve asset diversion, diversion of debt financing in the AI chain, instability of the yen). Current U.S. policies (including potential interest rate hikes) lean towards short-term responses rather than long-term mechanisms. This continues to act as a mid-term suppression factor for the global stock market. The U.S. non-farm payroll data exceeded expectations, raising the probability of a rate hike in September, which keeps the stock market under pressure. The issues behind high U.S. Treasury yields remain mid-term suppression factors for the global stock market. There are four objective mid-term problems we can summarize: 1. Weak U.S. fiscal discipline, with the proportion of short-term debt financing for Treasury bonds increasing and interest payments rising. 2. The major global central banks are overly diversified in their reserve asset allocations, constantly facing a loss of funds during the rolling issuance of U.S. Treasury bonds. 3. Changes in economic structure have increased the demand for debt financing in the AI chain, which also diverts global allocation funds. 4. The fragility of the Japanese economy and the yen cycle has become apparent, leading to selling pressure on U.S. Treasury bonds even within the existing intervention framework. Given these objective mid-term problems, the current policy mix in the U.S. is tilted towards short-term responses rather than long-term mechanics. U.S. Treasury actions aimed at reversal cannot be long-term deduced under weak fiscal discipline. Furthermore, the Federal Reserve's monetary policy framework in the AI era is still exploratory, with the parameters leaning towards traditional views, resulting in short-term interest rate hike risks. Under this combination, the global stock market remains sensitive to disruptions related to U.S. Treasury bonds, and in the absence of medium- and long-term mechanisms, the pattern of repeated disruptions is expected to continue for a while. 2. Technology has downplayed micro-structural discussions and returned to matching industrial logic and valuation: In the short term, the supply-demand balance for funds and market volatility has declined in sync, addressing the contradiction arising from rapid increases in the AI chain propelled by financial circulation in Q2 2026. The main contradiction in technology pricing has returned to the alignment of industrial logic and valuation. The future performance of computing power inflation is bound to differentiate, pending the initial stability of static performance valuations (post communications Q3 2026 and electronics Q4 2026), where the growth rate for 2027 may see differentiation and strengthening starting in Q4 2026. Additionally, we revisit the micro-structural issues in technology. First, it is unrealistic to expect a significant decline in public fund allocation ratios in the short term. Setting rising conditions based on this premise is overly strict. Based on 162 heavily weighted technology stocks in public funds, we break down micro-structural issues. By the end of Q2 2026, the proportion of revenues from the technology sector in the entire A-share market was 3%, profits accounted for 4%, and the market capitalization proportion was 16%, while the equity public fund configuration ratio was at 54%. Although 54% seems extreme, breaking down the ratios shows a significant rise in the sales net profit margin in Q2 2026, reflecting an industry foundation; however, static PE valuations remain at historically high levels, with the electronic sector rising 91% in Q2 2026, ranking fourth in historical industry increases. The rapid increase in market capitalization due to funds' positive circulation addresses the issue of excessive speed in valuation, which was also evident in the concentrated digestion from July to August; interestingly, the allocation coefficient actually marginally declined in Q2 2026. Historically, a reduction in public fund allocation ratios typically results from weakening fundamental trends or a decline in asset allocation environments. Thus, expecting significant declines in public fund allocation ratios in the short term is unrealistic. Recent factors leading to potential declines in allocation ratios mainly involve non-sector investment managers shifting focus to other areas for excess returns. Moreover, during the rising phase in May-June, public equity sector funds were marginally influenced by funds, and the underlying increase (approximately 210 billion) against the stock scale (approximately 790 billion) was not significant. Pricing power largely relies on concentrated shareholding and high turnover rates among investors. During the adjustment phase, the resilience of sector fund positions is evident, and the reallocation, reduction, or redemption actions of non-sector public funds have relatively limited impact. From tracking recent funds and supply-demand, electronic and communication holdings are hovering near the breakeven point, indicating narrow fluctuations in fund inflows and outflows. Market volatility has decreased, with transaction volumes declining, which corresponds to a reduction in the volatility of fund supply-demand. The influence of negative financial circulation pressures from July to August has diminished. The main contradiction in technology pricing has reverted to matching industrial logic and valuation. In this context, we infer that the subsequent performance of computing power inflation remains highly likely to diverge. The beta level of AI capital expenditure may see annual growth rates decline from 2027 to 2028, with a retreat pressure later in 2026. On the alpha level of computing power inflation, the fundamentals for computing power inflation in 2026 have universally improved; however, in 2027, the fundamentals may significantly diversify, with supply-demand contradictions easing in some areas, potentially leading to slower price increases. By 2028, the performance growth rate in most computing power chain segments may retreat. Up until the end of June 2026, projecting the industry's trend forward linearly shows nearly no resistance. However, at this stage, it is challenging for the market to support high valuations based on optimistic economic outlooks following 2028. Even so, after digesting the valuation, directions indicating improvements in the economic landscape for 2027 relative to 2026 should resume an upward trend. We highlight that the first objective for digesting valuations is for static valuations to return to historical medians. Barring any further adjustments, after the Q3 2026 communications report and the Q4 2026 electronics report, static valuations are expected to revert to historical medians. This corresponds to Q4 2026 potentially marking the starting point for a differentiated uptrend in the computing power chain. 3. Based on the A-share Q2 report, we discuss the performance outlook: Starting from Q3 2026, the year-on-year growth rate of net profits attributable to the parent for the entire A-share market has stepped down and exhibits volatility. The consensus on cyclical improvements remains difficult to achieve. Cost pressures may be prominently reflected in financial statements for Q3 2026. Impairments in Q2 2026 were above seasonal levels, and impairments for Q4 2026 may also be elevated. In Q2 2027, PPI growth will be low on a high base of year-on-year comparisons, placing pressure on revenue growth and leading to a drop in profit growth. From the A-share Q2 reports, the subsequent outlook for cyclical fundamentals shows that the apparent growth rate for the entire A-share market in the second quarter surpassed expectations; however, three potential issues require attention: 1. Supply clearances are gradual, but weak demand has resulted in revenue growth falling below original expectations. 2. The significant influence of non-recurring losses, with net investment income and fair value changes contributing positively, while exchange rate losses and asset value losses contribute negatively. 3. Cost pressures are not fully reflected in financial statements. The degree of extra improvement in profits (the primary influencing factor being the independent logic on the cost side) significantly exceeds historical reasonable confidence intervals, reflecting market prices that follow the trend, influenced by the cost pricing method of FIFO. Many industries have achieved price increases for finished products, resulting in effective pricing, yet inadequate cost pricing in Q2 2026 led to extra profit improvements. This corresponds to high costs starting to be priced in Q3 2026, alongside PPI rising again, leading to additional declines in profit growth. We infer that starting from Q3 2026, the year-on-year growth rate of net profits attributable to the parent for the entire A-share market will step down and show volatility. Cost pressures may become evident in the financial statements for Q3 2026, significantly pulling back expected profit growth rates. Impairments in Q2 2026 were higher than seasonal expectations, and impairments for Q4 2026 may also be elevated. In Q1 2027, the effects of high costs will marginally diminish, leading to a rebound in profit growth, returning towards the mid-term average. Meanwhile, PPI growth in Q2 2027, on a high comparative base, will suppress revenue growth, leading to a decline in profit growth. Under such performance validations, the upward expectations for cyclical improvements will face substantial resistance. 4. We maintain the judgment that the September rebound will continue, with a structural focus still on finding directions in the oversold rebound that can restore losses to a greater extent. Within the AI chain, we pay more attention to non-institutional heavy positions, while domestically produced computing power chains and small AI stocks provide "mid-term protection for the short term." In the medium term, focus remains on segments of computing power inflation, where performance growth rates in 2027 continue to accelerate, particularly in storage, high-end CCL, PCB, and capacitors. After September, the fluctuations and adjustments of technology are expected to continue. The time for non-technology directions to outperform has lengthened, and assets with high dividends have a prolonged period of absolute returns. In the non-technology sector, CXO and innovative pharmaceuticals have already been proven to enable a positive financial cycle. Precious metals may have short-term fluctuations, but in the medium term, as the pricing of U.S. dollar credit weakens, there remains room for growth. Additionally, opportunities in industrial metals and basic chemicals should be monitored. Continue to seek high-dividend assets based on "CSI 800 index weightsQ2 2026 public fund positions," with a focus on banks, non-bank financials, food and beverage, and utilities. Holding high-dividend assets remains viable while observing subsequent critical validation windows. Risk warning: Overseas economic recession exceeds expectations, and domestic economic recovery falls short of expectations.