This is not merely a structural bull market, but a quintessential case of "extreme market concentration." The primary risks to this concentrated rally hinge on two key variables.
The current market landscape is characterized by a stark divergence. On one side, sectors like semiconductors and optical communications within the AI supply chain are surging, repeatedly hitting new highs. On the other, traditional sectors continue to weaken, nearing a state of liquidity drought. Since April, the STAR 50 Index has risen 50% to a new peak, yet 65% of all stocks in the market have still not recovered their losses from earlier periods (as of May 25, 2026).
This phenomenon represents a classic "extreme consensus" trade. When capital abandons the search for diversified opportunities and instead piles frantically into a select few sectors, it prompts a critical reassessment: How much further can this overcrowded vessel sail? And if it turns abruptly, could it trigger a stampede and a capsize?
No one can predict short-term movements with precision. This analysis revisits five historical cycles of market concentration and subsequent unwinding in the A-share market to provide readers with a framework for reference.
**1. What Drives the Dramatic Swings?** To understand market concentration, one must first grasp the underlying concept of "shareholding structure." Stock price movements are not solely determined by fundamentals but are significantly influenced by "who holds the shares." The same company will exhibit vastly different price action if its shares are held by long-term institutional investors versus dispersed among short-term retail traders.
When capital begins to concentrate, shareholdings rapidly shift from a dispersed state to one of extreme concentration. A large portion of the float becomes locked up by long-term funds like mutual funds, which trade infrequently. This reduces the actual tradable supply in the market. Consequently, even modest incremental buying can quickly drive up prices—this is the reality behind the major upswings in concentrated sectors: "marginal pricing."
Conversely, when prices rise to a certain level, long-term holders begin taking profits, and shareholdings start dispersing again. Retail and momentum investors tend to flee at the first sign of trouble, making the shareholding structure extremely fragile. Once selling pressure emerges, prices enter an almost unstoppable downtrend.
In essence, the formation of a consensus trade is the process of shareholdings moving from dispersion to extreme concentration. Its unwinding is the stampede that occurs as holdings move from extreme concentration back to dispersion.
**2. The Long Lens of History: The Fate and Cycles of Five Episodes** History often rhymes. Analysis of five typical episodes of market concentration in A-shares reveals a recurring script of dramatic rises and falls.
1. **2007: The Cyclical Frenzy of the "Five Golden Flowers"** Accelerated urbanization and RMB appreciation converged, leading capital to concentrate on the cyclical "Five Golden Flowers" sectors: finance, real estate, and resources. The concentration ratio surged from below 30% to 45.7%. The cycle ended with the outbreak of the global financial crisis and the collapse of external demand, sending these cyclical stocks plummeting from their peaks. 2. **2010–2012: The First Consumer Sector Concentration** From 2010 to 2012, weak investment contrasted with strong consumption, supported by policies encouraging consumer spending. Capital concentrated for the first time on consumer staples like baijiu. However, the European debt crisis erupted, compounded by black swan events like the plasticizer scandal, shattering the defensive myth of consumer stocks and leading to a severe valuation compression in baijiu shares. 3. **2013–2015: The "Dream Valuation" of Mobile Internet** From 2013 to 2015, the issuance of 4G licenses and the rise of asset-light models fueled a concentration in mobile internet stocks. M&A activities spawned countless "dream valuation" myths, pushing the concentration ratio near 47%. The bubble burst abruptly as monetary policy tightened and financial deleveraging began, resulting in a market-wide plunge that became a nightmare for a generation of investors. 4. **2017–2020: The Era of "Core Assets" and the "Maotai Index"** Against a backdrop of foreign capital inflows and deleveraging from 2017 to 2020, capital concentrated on "core assets" represented by the "Maotai Index." The Maotai Index and the baijiu index significantly outperformed the broader market. However, as trade frictions escalated and economic downward pressure increased, the perceived certainty of these assets no longer granted immunity, and they too suffered valuation declines. 5. **2021–2022: The New Energy Surge of the "Ning Portfolio"** From 2021 to 2022, the "dual carbon" goals emerged, and the penetration rates of electric vehicles and photovoltaics rose rapidly. The new energy-focused "Ning Portfolio" rallied against the trend, with trading concentration soaring from 34.6% to 50%, a historical extreme. However, aggressive Fed rate hikes coupled with industry overcapacity plunged these sector stocks into a prolonged period of valuation digestion.
**3. Is This Time Different?** The current concentration around AI computing power and optical communication bears similarities to past episodes but also carries unique imprints of the current industrial cycle.
1. **Commonalities: The "Old Recipe" Persists** Historical analysis identifies three macro-level commonalities in past concentration episodes, which remain relevant today: * **Divergent Growth Structure:** Currently, global computing power construction is in a high-growth phase while traditional consumption and investment remain weak. This mirrors patterns seen in 2007 ("weak external, strong internal") and 2013 ("strong asset-light, weak asset-heavy"). * **Compelling Industry Narratives:** Past narratives included "urbanization + RMB appreciation," "consumption upgrade," "mobile internet revolution," "core assets," and "new energy revolution + dual carbon." The current narrative is "AI computing power revolution + optical communication new infrastructure." * **Complementary Domestic and External Liquidity:** Consensus trades thrive in a "bull steepening" interest rate environment (falling long-end rates, even looser short-end) and fear "bear flattening." Currently, domestic bond markets exhibit a tendency toward bull steepening, forming a staggered complement with U.S. Treasuries. The overall "liquidity friendliness for consensus trades" remains at a medium-to-high level. From a shareholding perspective, the proportion of total A-share turnover accounted for by the top 5% of stocks has risen from 33.1% in March 2025 to 42% in April 2026, approaching historically high levels. While crowding is significant, it has not yet reached the extreme levels seen in 2007, 2013-2015, or 2021-2022, which exceeded the current level by more than 15 percentage points.
2. **Differences: This Time Might Be "Somewhat Different"** Comparative analysis suggests several points of divergence in the current episode: * **Valuation Digestion Pace:** Historically, when a bull market theme peaked, the median PEG ratio often exceeded 2-3x (e.g., TMT in 2015, core assets in 2021). Currently, the median PEG for the AI computing power industry chain hovers around 1x. This suggests the current AI concentration is not solely propped up by multiple expansion; the earnings delivery capability of core segments far exceeds the "dream valuations" of the past. * **Market Correlation:** Past concentrations were more isolated to the A-share market. This time, the correlation between domestic tech stocks and the Nasdaq has surged to 0.63, placing it in the 91st percentile historically. This indicates the A-share AI narrative is deeply intertwined with the global tech supply chain, with capital expenditure plans of overseas tech giants acting as a barometer for the domestic concentration. * **Diffusion Logic:** Previous concentrations often rigidly focused on a few leading stocks. The current AI rally is spreading across the entire industry chain based on a triple logic: "shortage-driven price hikes → exploration of new demand → capacity crowding." This has created emerging opportunities in segments like liquid cooling and advanced packaging.
3. **Triggers for an Unwinding** Although the current concentration is still supported by earnings and global synchronization, historical patterns warrant attention. Historically, the unwinding of consensus trades is never triggered simply by being "too expensive" or "too crowded." High crowding is a necessary condition, but a reversal in macro fundamentals is the actual trigger. Currently, the primary risks that could loosen this concentration lie in two variables: * If domestic fixed-asset investment accelerates its recovery, a回暖 in pro-cyclical assets could divert funds away from tech stocks. * If the U.S. economy shows signs of marginal weakness, the AI sector's substitution effect on traditional industries could backfire, undermining its own capital expenditure rationale. Additional potential triggers for an unwinding in this cycle, which merit close monitoring, include: 1. Underperformance in application-layer monetization, where commercial revenue fails to cover high computing costs, leading to cuts in capital expenditure. 2. Industry overcapacity or excessive investment in certain segments; monitoring corporate CAPEX levels is crucial. 3. The emergence of new technologies rendering existing capacity stranded assets. 4. A sharp shift in the liquidity environment, breaking the positive feedback loop of capital flows, or a subsequent lack of incremental funds. 5. Other macro changes that could cause a shift in the market's main narrative.
**4. Avoiding a Seat on the Crowded Ship** The greatest danger often lies not in obvious negative news, but in an overwhelmingly crowded consensus. For the average investor, the most perilous move in the face of extreme market concentration is to be lured by short-term, spectacular gains and make a concentrated bet on a single sector at the peak of euphoria. When all market capital is crammed behind one door, the damage from a potential stampede, should an alarm sound, is irreversible. While we cannot precisely predict when a consensus will unravel, we can adhere to certain principles: not mistaking the era's beta for one's own alpha, not believing "this time is different" amidst the clamor, replacing single-sector bets with balanced asset allocation, and seeking equilibrium between the high volatility potential of tech and the certainty offered by lower-volatility assets. What endures across cycles is not any single, extreme sector bet, but the resilience of a well-considered asset allocation and a respectful adherence to financial常识.
Comments