The A-share market has shown resilience at its bottom, with policy, capital, domestic demand, and sentiment floors becoming clearer. Since July, state-owned capital platforms have increased holdings, stock ETFs have seen continuous inflows, listed companies have boosted buybacks, and major shareholders have reduced their selling, providing strong support for the market. The July Politburo meeting set a positive tone, and as fiscal funds accelerate in the second half of the year, policy effects are expected to further transmit to the economy and corporate earnings.
Overseas AI cycle shows flaws, liquidity shocks amplify volatility
Overseas tech adjustments are triggered by crowded trades, deleveraging, and tightening liquidity expectations, but the deeper cause is a loosening of the AI industry's fundamental cycle. In South Korea's stock market, excess returns from leveraged funds have significantly declined, leading to a short-term rebound, but leveraged positions remain high, and deleveraging may not be complete, constraining the rebound's strength and sustainability. The U.S. stock market may also see a recovery driven by oversold conditions and easing policy expectations, but the upward revision in capital expenditure partly stems from rising storage and hardware costs, increasing pressure on cloud vendors' free cash flow. This has loosened the original positive cycle of "revenue growth, cash flow improvement, and capital expenditure increases." The sustainability of the subsequent rebound depends on market confidence in improving capital returns from AI applications.
Increased overseas volatility does not alter the direction of internal rebalancing in A-shares
Since July, the distribution of individual stock returns in A-shares has been gradually shifting rightward, while the return distribution of actively managed equity funds has shifted leftward. Underperforming sectors like coal, oil and petrochemicals, banking, and food and beverages have shown significant recovery. Market opportunities are spreading to more carbon-based and silicon-based downstream industries. The share of trading volume for silicon-based upstream companies remains high, and it will take time to complete rebalancing in asset allocation.
Maintain the strategy of "rebalancing allocation and actively positioning for a counterattack"
Continue to focus on blue-chip value directions with strong fundamentals, such as new energy, non-ferrous metals, and non-bank financials. Within technology, pay more attention to AI applications, with a focus on Hong Kong-listed tech stocks, domestically listed computer companies, and media sectors. Finally, according to the rebalancing principle, if allocation to silicon-based upstream is low, consider domestically listed "true tech" leaders that have undergone sufficient adjustments, have reasonable valuations, and have clear competitive moats.
Funding and policy factors converge, A-share market bottom may have been found
Over the past two weeks, we have published reports "Embrace Golden Allocation Opportunities, Actively Deploy for Counterattack" and "Re-evaluation of Carbon-Based Blue Chips, Flourishing." The carbon-based blue chips have led an A-share rebound. Under high volatility in global equity markets, A-shares' dual-creation sectors continued to adjust, but the overall market remained resilient, supported by carbon-based blue chips. We maintain an optimistic view on the A-share index's future trend, with policy, capital, domestic demand, and sentiment floors becoming clearer. Since July, state-owned capital platforms have increased holdings, stock ETFs have seen continuous inflows, listed companies have boosted buybacks, and major shareholders have reduced their selling, all providing strong support for the market. The July Politburo meeting set a positive tone, emphasizing "enhancing the resilience and confidence of the capital market." With fiscal funds expected to accelerate in the second half of the year, policy effects will further transmit to the economy and corporate earnings, potentially confirming the market bottom.
On the funding front, policy capital is becoming a key support for A-shares. On July 19, state-owned capital platforms like Guoxin and Chengtong announced plans to increase holdings in Chinese stock assets, sending a clear signal of market stability. Historically, after such announcements by entities like Huijin, Guoxin, and Chengtong, the Shanghai Composite Index has often stabilized and rebounded. As important tools for stable capital entering equity markets, ETFs have seen significant inflows since July, playing a notable supporting role. Beyond policy capital, listed companies and industrial capital are also turning more positive. On one hand, the number of repurchase proposals in A-shares increased significantly in July. Historically, a concentrated increase in repurchase proposals often coincides with market bottoms, serving as an important signal of industrial capital's attitude. On the other hand, at current price levels, major shareholders' willingness to sell has weakened, with net selling sizes shrinking significantly, reducing industrial capital's drag on the market.
On the policy front, the July Politburo meeting set a positive tone, and fiscal funds are expected to accelerate in the second half of the year. The policy focus may shift from stabilizing expectations to promoting the implementation of existing policies, with previously allocated fiscal resources gradually converting into physical work volumes. As fiscal spending increases, policy effects will gradually transmit to the economy and corporate earnings, further solidifying the domestic demand and corporate earnings bottom.
Overseas AI cycle shows flaws, liquidity shocks amplify volatility
Trading structure is crowded, deleveraging and liquidity tightening are triggers for overseas tech adjustments
On June 22, the South Korean Financial Supervisory Service criticized leveraged ETFs tracking Samsung and SK Hynix, leading to a peak and subsequent decline in the South Korean stock market, followed by a deleveraging process. Based on the 2015 A-share experience, deleveraging typically requires two stages. In the first stage, excess returns from leveraged funds are cleared. When leveraged funds' excess returns fall to zero, it often corresponds to a rebound window after the first wave of panic subsides. In 2015, after the first round of A-share declines, leveraged funds' excess returns fell from highs. This means the first wave of profit-taking, stop-losses, and passive liquidations were concentratedly released, making it easier for short-covering funds and policy stabilization capital to drive a technical rebound. In 2026, the South Korean stock market showed similar characteristics: leveraged funds' excess returns fell from near 150% to around zero during this adjustment, creating conditions for short-covering and risk appetite repair. However, this signal only indicates that a short-term rebound has micro-transactional support, not that the market's mid-term bottom has been confirmed.
In the second stage, leveraged fund positions are cleared. Before leveraged fund exposures are fully de-leveraged, market rebounds are highly fragile. From January to June 2015, A-share leveraged funds grew by about 120%. After the first round of declines, leveraged funds' scale only fell by about 43%. Although leveraged funds' excess returns had largely been lost, the financing scale was still significantly higher than pre-rally levels. Residual leverage later turned into selling pressure after the market rebound stalled, leading to a second round of deleveraging around September. From January to June 2026, South Korean stock leveraged funds expanded by about 203%, significantly higher than A-shares' leverage expansion in 2015. As of July 30, 2026, the leverage fund scale had fallen by only about 53% from its June peak. Its residual exposure and potential selling pressure may still be significantly higher than the first rebound phase of A-shares in 2015. During the South Korean stock rebound, residual leverage may re-accumulate stop-loss and liquidation pressure, posing a risk of double-bottoming.
Overall, excess returns from leveraged funds in South Korean stocks have been cleared, with concentrated profit-taking, stop-losses, and passive liquidations being partially released, creating conditions for a market rebound. However, leveraged fund positions in South Korea have not yet been cleared. The market can only form a mid-term bottom condition after leverage exposures are fully de-leveraged and residual selling pressure is fully absorbed.
Looking at US stocks, due to larger earlier declines and a marginal easing of Fed policy expectations, the Philadelphia Semiconductor Index sharply rebounded after a deep decline last week. However, we believe economic uncertainty is high, with increasing internal divergence within the Fed. Subsequent US employment, inflation, and economic growth data may still fluctuate, and macro disturbances could appear at any time, making it difficult for the market to sustainably and linearly trade a single easing logic. Whether it's the Philadelphia Semiconductor Index or South Korean memory stocks, the trigger for the June peak and decline was crowded trades combined with deleveraging and liquidity tightening expectations, but the deeper reason is flaws in the AI industry's fundamentals. Therefore, the completion of deleveraging or marginal easing of liquidity is expected to bring a rebound, but it will be difficult for the market to return to the pre-June uptrend before fundamental concerns are alleviated.
The AI cycle is unsustainable, and fundamental loosening is the core reason for the decline
Recently, some investors have been discussing whether the current AI cycle adjustment resembles 1998 or 2000, as these two hypotheses lead to very different outcomes. In terms of magnitude and speed of gains, the current Philadelphia Semiconductor Index is closer to the state around the peak of the 2000 tech bubble, while the Nasdaq Index's overall performance is relatively more moderate, closer to the period around 1998. Of course, whether it's 1998 or 2000, both have short-term rebound potential, but whether it can evolve into a new trend depends on whether the AI industry logic can re-close. The market needs to see new application breakthroughs and confirm that the revenue, profits, and cash flow generated by AI businesses can cover high capital intensity investments. So far, the AI cycle may be unsustainable, and it's becoming increasingly difficult for the industry logic to re-form a closed loop.
In the past period, the AI cycle was as follows: revenue exceeding expectations drove cash flow improvement, cash flow improvement supported capital expenditure increases, capital expenditure increases further strengthened upstream computing power demand and industry growth expectations, ultimately pushing the market to raise valuations for related companies. However, based on the current earnings season, this cycle has loosened. On one hand, the expansion of AI capital expenditure has been too fast, with upstream cloud vendors being "unable to support it alone." Currently, some cloud vendors face cash flow pressure. Although capital expenditure guidance remains high, the upward revision is relatively limited. Some new investments may be more used to cover rising storage and hardware costs rather than proactively increasing AI investment intensity. On the other hand, the current AI industry chain's investments, financing, and procurement contracts are highly interdependent, creating a fragile cycle financing structure. When the financing environment is smooth, this model can accelerate industry expansion. But once financing costs rise, model companies' ability to pay decreases, or AI application returns disappoint, the cash flow pressure across the industry chain may also be amplified synchronously. Therefore, from a trading logic perspective, after the sharp drop in overseas tech stocks, there is indeed rebound momentum, but the rebound's space and sustainability are still constrained by fundamentals. Anthropic ARR, OpenAI IPO, Microsoft CAPEX, Google FCF, and SK Hynix operating profit all fell short of expectations. We believe the subsequent market trend is unlikely to continue the strong unilateral performance of the first half of the year.
Increased overseas volatility does not alter the direction of internal rebalancing in A-shares
Some investors worry that against the backdrop of increased volatility in US and South Korean stocks, A-shares will also come under pressure. However, in fact, we believe A-shares are transitioning from a bull market for a few companies to a bull market for more companies. The first half of the year saw a distinct structural market in A-shares. The distribution of listed company returns was skewed to the left, with significant gains for a large number of companies, and market gains were mainly driven by a few silicon-based upstream companies. Correspondingly, actively managed equity funds achieved significant excess returns by concentrating on core themes, with their return distribution significantly better than most individual stocks. Since July, this structure has begun to reverse. The distribution of listed company stock price movements has gradually shifted rightward, significantly improving the wealth effect at the individual stock level. Meanwhile, the return distribution of actively managed equity funds has shifted leftward, with the excess returns from concentration holdings in the first half being given back. The industry performance also shows a clear mirror image. Industries that lagged in the first half, such as coal, oil and petrochemicals, banking, food and beverages, and beauty and personal care, have seen significant recovery in July. Meanwhile, strong sectors in the first half, like electronics and communications, have experienced significant adjustments. We believe the key to obtaining excess returns in the future is no longer to cling to the single strong theme of the first half, but to seize the direction of diffusion of the wealth effect, gradually rebalancing toward value blue-chip directions.
From a price and valuation perspective, A-share rebalancing has made significant progress, with extreme market pricing divergences being alleviated. Specifically, the standard deviation of secondary industry returns has rapidly fallen and is approaching historically low levels, indicating that return divergences between industries have significantly narrowed. Meanwhile, the relative performance gap between TMT and the overall A-share market has also rapidly declined from historical highs, with the tech sector's accumulated relative gains being given back, while the relative performance of value and traditional industries is gradually recovering. Although price and valuation rebalancing has been relatively clear, rebalancing at the trading and holding levels remains insufficient. Currently, both the share of trading volume from the top 5% of stocks and the share of trading volume from the TMT sector are at high levels, indicating that market trading activity remains concentrated, and the rebalancing of capital trading attention has a long way to go. Overall, we believe rebalancing is far from over. The share of trading volume for silicon-based upstream companies remains high, and it will take time to complete rebalancing in asset allocation. Even if the tech sector experiences a periodic rebound, it does not mean the market will return to the single concentrated state of the first half. A more likely scenario is that market opportunities spread to more carbon-based and silicon-based downstream industries, forming a more diversified market structure among technology, cyclical, and consumer sectors.
Continue to maintain the strategy of "rebalancing allocation and actively positioning for a counterattack"
Currently, the lower limit of the A-share index is gradually becoming clearer, and the wealth effect is starting to spread to more stocks. However, overseas market volatility, uncertainty in AI industry chain capital expenditure and cash flow, and the problem of crowded trades still exist. We believe the core issue for the market in the future is how to actively seek assets with offensive capabilities in the next phase while defending the lower limit of the portfolio. In terms of allocation, we continue to maintain the approach of "rebalancing allocation and actively positioning for a counterattack." First, continue to focus on blue-chip value directions with strong fundamentals, such as new energy, non-ferrous metals, and non-bank financials. Second, within technology, pay more attention to AI applications, with a focus on Hong Kong-listed tech stocks and A-share silicon-based downstream directions. Against the backdrop of adjustments in A-share silicon-based upstream, some mainland active funds are seeking lower-valued and more competitively strong Hong Kong-listed tech assets through southbound channels, thereby driving relative returns for the Hang Seng Tech Index. Within A-shares, compared to the silicon-based upstream, which is highly concentrated in capital expenditure, the silicon-based downstream is closer to the application and commercialization stage. If AI applications gradually land, downstream companies are more likely to directly convert technological progress into revenue and profits. Finally, pay attention to domestically listed "true tech" leaders that have undergone sufficient adjustments, have reasonable valuations, and have clear competitive moats. Rebalancing does not mean zero allocation to technology. Portfolios can moderately restore attention to high-quality tech assets, but should not return to the highly concentrated state of the first half. In the future, a batch of tech companies with strong industrial competitiveness plan to enter the capital market, putting existing listed companies under more direct competitive pressure. Only those companies with technological barriers, product capabilities, customer bases, and profit models that can withstand competition from new entrants are more likely to obtain relatively sustained valuation premiums.
Risk warnings
1) Increased overseas risk disturbance: Escalation of global geopolitical conflicts, regional friction, and trade tensions could boost global risk aversion, triggering cross-border capital flows and risk asset volatility. Meanwhile, uncertainty in the monetary policy paths of major overseas economies could lead to repeated interest rate expectations, suppressing the valuations of growth sectors and creating temporary external disturbances for A-shares. 2) Sustainability of AI industry chain capital expenditure falls short of expectations: Global AI computing power and data center investments are highly dependent on the capital expenditure pace of overseas tech giants. If corporate earnings come under pressure and investment return cycles lengthen, it could lead to a slowdown in AI capital expenditure growth, thereby weakening demand in upstream computing power, hardware, and materials sectors, affecting the growth certainty of the AI infrastructure chain. 3) Domestic policy and economic recovery fall short of expectations: The current domestic economy is still in a stabilization and recovery phase, with weak endogenous momentum. Property sales and private investment recovery are slow. If the pace of economic recovery slows, it will further suppress market earnings expectations and risk appetite.
Comments