July marked a high-stress test for global technology stocks worldwide.
"Who would have thought the South Korean stock market would have such a huge impact now?" This remark from an industry insider captures the true situation in the capital markets.
A party built on leverage ended in a stampede, quickly crossing borders and triggering a sharp sell-off in US semiconductor stocks and a significant pullback in A-share computing power sectors. Technology stocks in China, the US, and South Korea experienced a rare co-movement, even creating a new cross-border risk model.
Huang Liang, Deputy Director of Market Support and Management at China Merchants Fund, defines this as a "trading-driven cross-border risk transmission model." In his view, this risk is confined to the semiconductor technology sub-sector, with the source being rule changes and regulatory adjustments in secondary market derivatives trading. The repair cycle is expected to be relatively short-term rather than long-term.
After a round of risk release, the A-share market has recently rebounded. However, industry insiders believe it is still necessary to monitor core variables like the progress of leverage liquidation and trading concentration to determine if the adjustment is completely over. The existence of this new type of cross-border risk also presents new thinking for future risk control and pricing logic.
How the South Korean Stock Storm 'Spread' Globally
This year, the South Korean stock market has experienced an extreme roller-coaster ride. Its main index surged over 120% in the first half of the year, but after hitting a peak in late June, it saw a maximum retracement of over 30% in just one month, triggering multiple circuit breakers during that period. Among these, the two tech giants Samsung Electronics (005930) and SK Hynix (000660) both saw maximum retracements of over 33%, becoming the core trigger points for this adjustment.
This sharp correction, which began in South Korea, did not stay within the regional market. It quickly spread outward, dragging down global tech stocks collectively, and the A-share computing power sector also weakened significantly. Wind data shows that as of July 22, the Philadelphia Semiconductor Index (SOX.GI) and the CSI Computing Power Index (931688) both experienced maximum retracements of over 22% in the past month.
In the industry's view, this round of global tech adjustment is not an indiscriminate unwind of AI trades. Instead, it shows a clear gradient transmission characteristic of "memory - upstream computing power - mid-to-downstream." The South Korean market and the global memory sector are the epicenters of this adjustment, with the core risk source pointing directly to the deleveraging process in the South Korean market.
The destructive power of this storm is rooted in the special two-tier structure of the South Korean market. The first is a "top-heavy" structure at the individual stock level. The combined market capitalisation of the two memory giants, Samsung Electronics and SK Hynix, exceeds half of the total market capitalisation of South Korea's main board, meaning the performance of these individual stocks can dictate the entire market's direction.
The more critical factor is the "triple-nested" leverage on the capital side. Xia Fanjie, a senior strategy analyst at CSC Financial, summarises the leverage structure of South Korean retail investors as "triple nesting." This involves borrowing money from outside the market, using margin trading within the market, and then buying 2x or 3x leveraged ETFs. In other words, investors borrow funds through credit loans to enter the market, further increase leverage through brokerage margin trading, and then concentrate the borrowed funds into buying 2x and 3x leveraged ETFs for SK Hynix and Samsung Electronics.
"When the three layers of leverage multiply, passive selling during a price decline is transmitted and self-reinforces layer by layer, forming the microstructural root of this round of violent fluctuations in the South Korean market," Xia Fanjie explained.
Within this structure, the relevant single-stock leveraged ETFs act as key "accelerators" and "transmitters," and this is a core feature distinguishing this risk from traditional market movements.
Huang Liang told Caixin that, on one hand, these products convert local volatility in South Korea into a benchmark for global industry sentiment pricing, creating a channel for cross-market risk transmission. On the other hand, the daily rebalancing rules of leveraged products trigger passive, cascading liquidations during downturns, amplifying individual stock declines and spreading them across the entire sector, making them a core tool fueling the market stampede.
Under this mechanism, volatility in South Korea's tech leaders is transmitted to the A-share market through three clear pathways: fundamental expectations along the industry chain, northbound cross-border capital flows, and Asia-Pacific market sentiment.
"The initial extreme and rapid single-day drops were more driven by panic than being completely divorced from fundamental industry support. Emotional shocks only cause short-term impulsive movements, but leveraged financing amplifies the magnitude of sustained corrections," he said.
Rare Three-Region 'Co-Movement' Creates New Cross-Border Risk
How did leveraged volatility in a single market evolve into a collective adjustment for the global tech sector?
Jindalai, a macro strategy researcher at Golden Eagle Fund's equity research department, told Caixin that both China and South Korea provide manufacturing support for the US AI industry chain and benefit from hardware procurement like semiconductor equipment and optical modules driven by AI capital expenditure from US hyperscale cloud vendors. Therefore, changes in the stock prices of South Korea's tech leaders reflect, to some extent, market concerns about the sustainability of AI industry chain capital expenditure.
A source from a major institution's equity department added that Samsung and SK Hynix, as the duopoly in the global memory chip market, effectively act as a "pricing anchor" for AI computing power hardware expectations. Their stock price volatility reflects marginal changes in global memory demand (expectations). These changes in industry expectations are directly transmitted along the industry chain to A-share sectors related to memory chips, optical modules, and PCBs.
It is this deep binding at the industry chain level that has led to the "synchronised co-movement" of the technology sectors in China, the US, and South Korea, which industry insiders consider "relatively rare in history."
"In the past, cross-market tech pullbacks were mostly initiated by weakening demand in the US market, and there was a time lag between adjustments in different markets. This time, the tightening of supply-side policies in South Korea triggered the risk, with the US and A-share markets quickly following the decline synchronously," Huang Liang analysed. The core pricing factor is the supply-demand expectation for the global semiconductor industry chain, which is the link with a stronger transmission effect in this decline.
In his view, compared to previous tech cycles, this round of adjustment has three core differences: First, the trigger was active deleveraging by South Korea's domestic regulators, not macro-cycle factors. Second, the risk originated from high-leverage ETFs concentrated on a single heavyweight stock, not broad-based index instruments. Third, the extreme crowding in the global AI track amplified the speed and magnitude of the synchronous sell-off.
The aforementioned fund manager further explained that in the past, whether during the tech bubble or the mobile internet era, technology penetration and capacity expansion usually involved a "time lag" and "regional barriers." However, this time, the consistency across the three regions' tech sectors is very strong. In fact, the current global tech cycle is almost entirely built on the single grand narrative of "AI computing power infrastructure." Once the trend reverses, it is very easy to trigger programmatic synchronous selling and concentrated profit-taking, forming an extreme co-movement in the downturn.
Based on this unique volatility logic, industry interviewees define this market movement as a "trading-driven cross-border risk transmission model," which is fundamentally different from traditional financial crises.
Huang Liang explained that traditional crises typically impact the real economy, consumption, and financial systems, with longer repair cycles. In contrast, this round of risk is confined to the semiconductor technology sub-sector and has not affected bank credit or the real economy. The risk source is rule changes and regulatory adjustments in secondary market derivatives trading. Therefore, the repair cycle is expected to be relatively short-term rather than long-term.
However, there are dissenting voices within the industry regarding the actual impact of the risk transmission. Jindalai believes that the financial risk in South Korea does not have a substantial liquidity path to significantly affect the A-share market.
She argues that, on one hand, capital flow changes in South Korea are unlikely to materially impact the A-share market. Compared to the South Korean market, the A-share market has not accumulated excessive leveraged capital during its recent rally. On the other hand, South Korea and the A-share market have different investor structures and industry market capitalisation distributions, meaning their sentiment digestion processes may not necessarily be synchronised.
Is the 'Danger Alert' Lifted?
During this round of global co-movement adjustments, the A-share market experienced a maximum retracement of over 10% in the past month. However, with stabilising signals from regulators, state-owned capital operating platforms, and financial institutions, coupled with positive factors like increased ETF trading volume, the A-share market has recently rebounded. As of July 23, the Shanghai Composite Index has risen nearly 3% over the past four days, while the STAR 50 Index has risen 4.33% over the same period.
Amid this short-term rebound, the core debate in the market is whether the "danger alert" for the tech stock adjustment has been completely lifted. Institutions generally believe that short-term risks may have eased, but a full stabilisation still needs verification.
The Invesco Great Wall investment and research team believes that negative factors, such as the leveraged ETF issue in the South Korean stock market, may gradually be resolved in the future. The recent market pullback may have already fully released its potential. Compared to global markets and other asset classes, A-share valuations still hold strong appeal. The AI industry chain has seen a significant correction. Once the short-term capital flow negative feedback loop is completed, it may present a re-entry opportunity.
"Currently, the negative narrative for AI requires a positive signal from the earnings reports of North American cloud vendors to reverse, and the risk from leveraged ETFs in the South Korean stock market is also being released," the team believes. Current market valuations imply overly pessimistic expectations. In a low-interest-rate environment, market valuations need to consider cross-asset comparisons. A-share valuations remain attractive compared to global and other assets, so further pessimism is unwarranted.
"The recent market volatility is a sharp fluctuation dominated by trading-side deleveraging triggers, concerns over domestic and foreign liquidity, and amplification by crowded structures. However, it is not the end of the industry or profit logic," Wang Li, a senior macro strategy researcher at Great Wall Fund, told Caixin. In the short term, tolerance for volatility is necessary while tracking market signals and industry delivery progress. The offensive slope should be re-evaluated after crowding and volatility are repaired.
So, when will the adjustment truly end? According to institutional interviewees, to determine if the subsequent market can stabilise sustainably, several core verification variables must be closely monitored.
Wang Li believes it is necessary to track indicators such as whether leverage has been cleared to a stage-appropriate level, whether the volatility premium has returned, whether trading concentration has de-crowded, and whether external disturbances have become less sensitive. Among these, special attention should be paid to the earnings window for US AI chains and cloud vendors, focusing on revenue, capital expenditure guidance, and utilisation rate changes. If the demand-return feedback loop is not disproved, it could help alleviate the repricing pressure on global AI hardware.
"Without new external shocks, the most intense phase of trading-side deleveraging may have passed. However, since the volatility term structure and crowding have not been fully repaired, short-term high volatility may still be difficult to avoid," he said. From a historical comparison and industry tracking perspective, the first wave of decline driven by trading factors may be accompanied by a rhythm of "technical rebound followed by re-differentiation." The subsequent slope will mainly depend on the verification strength of "earnings and orders."
Building a 'Breakwater' for A-shares
Beyond short-term coping strategies, a longer-term question arises: Against the backdrop of a highly interconnected global tech industry chain and free cross-border capital flows, how can irrational external shocks be effectively isolated and a robust local market risk defence line be built? This has become a topic of discussion for the A-share market and institutional investors.
"Against the backdrop of K-shaped global economic divergence, the A-share market, as a deep participant in the AI industry chain, can hardly fully insulate itself from the co-movement volatility of overseas markets. Furthermore, whether the prosperity trend of the AI industry chain continues will determine if the tech main line can return after undergoing sufficient deleveraging," said Jindalai.
From a strategic allocation perspective, she suggests that when the market reaches a stage of consensus bullishness and even overheated trading, a more cost-effective allocation strategy might be style balancing. After trading crowding and leveraged capital have cleared, it is still necessary to monitor industry trend progress. If the trends remain strong, a re-entry into core links with increased AI demand share and supply-demand gaps could be considered.
From a market hedging and long-term development perspective, Huang Liang proposed a more systematic approach. He stated that the A-share market currently relies on various measures such as industry optimisation, support from medium-to-long-term capital, control of on-exchange leverage, and counter-cyclical industrial policies to hedge against external shocks and reduce the impact of irrational overseas market volatility. Fundamentally, stable industrial policies and long-term growth expectations are the core foundation for strengthening market pricing.
"Institutional investors need to comprehensively upgrade their risk control and allocation thinking, such as building a special monitoring system for overseas leveraged derivatives, strictly controlling position limits in single tech tracks, and setting risk control discipline for extreme market scenarios," he argued. In his view, shocks caused by pure market sentiment or programmatic trading are highly reversible. They represent derivative transmissions at the stage of capital trading and will not independently have a significant impact on long-term trends.
Huang Liang suggested that on the allocation side, emphasis should be placed on balanced positions under valuation considerations, valuing the role of high-dividend defensive assets, while continuously tracking growth-oriented domestic demand tracks like equipment and materials with strong substitution potential. Strategically, fixed-income assets should be combined to hedge systemic volatility and mitigate the high volatility characteristics that risk assets face periodically.
The aforementioned fund manager believes that given the current high degree of industry chain binding, it is unrealistic for the A-share market to achieve complete "physical isolation" from overseas shocks.
"The true moat can only be built on 'industrial independence and pricing power'," he argued. In his view, besides existing capital project management and price limit rules, the A-share market needs to cultivate a local IT innovation and self-controllable ecosystem that is less affected by overseas capital expenditure. Simultaneously, it must strengthen the power of domestic long-term institutional capital to avoid being led by the sentiment of cross-market quantitative funds.
Huang Liang suggested that on the allocation side, emphasis should be placed on balanced positions under valuation considerations, valuing the role of high-dividend defensive assets, while continuously tracking growth-oriented domestic demand tracks like equipment and materials with strong substitution potential. Strategically, fixed-income assets should be combined to hedge systemic volatility and mitigate the high volatility characteristics that risk assets face periodically.
Comments