To The Moon

Historical data for the ChiNext Index and the STAR 50 Index indicates that after extreme monthly drops exceeding 15%, there is a high probability of recovery in the following month or two, with rebounds often larger than expected. The declines recorded by both indices in July have already entered historically extreme territory. Based on historical patterns, the August-September period presents a relatively high-probability window for a potential recovery.

Historical Review of Extreme Declines

The three largest single-month declines in the history of the ChiNext Index are: 1. January 2016, a drop of -26.53%; 2. July 2026, a drop of -23%; 3. August 2015, a drop of -21.38%. The three largest single-month declines for the STAR 50 Index are: 1. July 2026, a drop of -25.9%; 2. January 2024, a drop of -19.62%; 3. March 2026, a drop of -15.57%. It is important to note that historical performance does not predict future results.

Comparing This Crash to Historical Events

Similarities between the current downturn and past instances include that the major declines all occurred after a period of significant rally, with valuations at elevated levels, and were accompanied by external shocks. The current scenario is similarly based on the rapid rise of the AI sector from April to June, which pushed valuations to extremes. The trigger was the market's concern over overcapacity, sparked by Meta's sell-off of computing power, and the subsequent sharp decline in the SOX (Philadelphia Semiconductor Index). A key difference is that the 2015/2016 events were systemic liquidity crises caused by structural shocks and deleveraging, while the 2024/2025 downturns were driven by the macroeconomic environment and a semiconductor cycle downturn. The current situation is more about short-term industry supply-demand fluctuations and valuation digestion. The long-term AI trend, supported by cloud capital expenditure and computing power demand, has not reversed. If upcoming earnings reports confirm that computing power demand remains strong, the logic for this recovery could be clearer than in most historical cases.

Extracted Historical Patterns

Pattern 1: High Probability of a Monthly Gain In the four data sets analyzed, the probability of the index rising in the month following a crash was 75%. In the two instances for the STAR 50, the subsequent month's gains both exceeded 17%. Even when the following month continued to decline, as in January 2016, a strong rebound occurred in the third month.

Pattern 2: Maximum Strength in the Third Month After the two extreme crashes for the ChiNext Index, the average gain in the second month (T+2) was as high as +18%. This suggests that after extreme sentiment clears, the medium-term recovery momentum is very strong.

Pattern 3: Two-Phase Recovery Path The post-crash trend typically unfolds in two phases. The first phase, within one month, may involve continued consolidation or a mild recovery. The second phase, within two months, is likely to see a significant rebound.

Pattern 4: Recovery Slope Depends on the Nature of the Decline A liquidity or sentiment-driven sell-off, like in August 2015 or January 2024, often leads to a rapid recovery in the following month. A structural shock, such as the January 2016 circuit breaker, results in further consolidation in the second month before a strong recovery in the third. The current situation, where the industry logic remains intact and valuations are being digested, falls between these two types, suggesting a relatively higher probability of a recovery window in August-September.

Implications for the Current Market

In July, the ChiNext Index fell by approximately -23%, its second-largest drop on record. The STAR 50 Index fell by approximately -25.9%, its largest ever. Both have entered extreme historical territory. Based on historical patterns, the probability of an August-September recovery window is relatively high. As the current decline is more about short-term supply-demand expectations and valuation digestion in the AI sector, and the long-term industry trend has not reversed, the recovery logic is clearer than in most historical cases.

Key Verification Points Ahead

Several key upcoming events will be crucial to monitor. First, quarterly reports from the four major cloud service providers—Google, Microsoft, Meta, and Amazon—will be released, with a focus on capital expenditure. Second, the US July non-farm payroll and unemployment rate data, due on August 7, will influence interest rate cut expectations and liquidity. Third, Nvidia's earnings report on August 26 will be key for assessing its Beta acceleration. From a positioning perspective, the AI computing power supply chain, with optical modules as a key component, remains a core theme for the ChiNext Index. After the deep correction in July, valuation pressure has been significantly reduced. Based on historical patterns and sentiment clearing, August to September is theoretically a potential window for the most significant recovery. For medium-to-long-term investors, phased buying opportunities may be considered.

This market analysis is for reference only. Recent market volatility may be significant, and short-term gains or losses do not predict future performance. Investors must make rational investment decisions based on their own financial situation and risk tolerance, paying close attention to position and risk management.