From the “AI bull market” to “leveraged volatility”: why South Korea’s stock market is more crazy than Bitcoin

Author: 137Labs

For a long time, Korea’s main stock benchmark, the KOSPI, has been seen as the world economy’s “canary.”

South Korea is highly dependent on exports, and companies such as Samsung Electronics, SK hynix, Hyundai Motor, and LG Chem are deeply embedded in the global supply chains for semiconductors, automobiles, batteries, and consumer electronics. Therefore, when global manufacturing orders rise and demand for electronics rebounds, South Korea’s exports and stock market often improve first; when global trade weakens, the Korean market typically feels the chill earlier as well.

But once the artificial intelligence investment cycle began, KOSPI’s role has been fundamentally changing.

It is no longer just a traditional cyclical index reflecting overall global economic conditions; it increasingly resembles a “super AI chip fund” dominated by Samsung Electronics and SK hynix. When capital, index weights, leveraged products, and retail sentiment all concentrate in a handful of stocks, the Korean stock market also shifts from being a barometer of the global economy to becoming one of the most aggressive AI trading testbeds in the world.

Once exceeding 60% in annualized volatility, frequently triggering pauses in programmatic trading, semiconductor leaders posting double-digit daily gains and losses, and a rapidly expanding single-stock leveraged ETF—all of this reveals the essence of the rally:

What the Korean stock market is facing is not merely a normal valuation adjustment, but a structural shock formed by the mutually reinforcing effects of highly concentrated indices, AI capital expenditure expectations, leveraged product mechanics, and society’s speculative psychology.

1. Why KOSPI has turned from an “economic barometer” into an AI index

The core logic behind this round of the Korean stock market rally is the demand for storage chips driven by AI infrastructure buildout.

Large AI model training and inference require massive high-speed data transmission. Beyond computing chips provided by companies such as NVIDIA, high-bandwidth memory (HBM) has also become an indispensable key component for AI servers. SK hynix and Samsung Electronics are among the world’s most important manufacturers of storage chips.

The market has therefore begun to redefine these two companies.

Previously, the storage chip industry was usually viewed as a typical cyclical sector: prices rise when demand improves, suppliers expand capacity leading to oversupply, and then the industry enters a phase of price cuts and inventory adjustments. Company valuations were often constrained by the industry cycle.

The AI boom has changed investors’ imagination. The market has started to believe that high-bandwidth memory may no longer be just an ordinary cyclical product, but rather a long-term scarce resource for AI infrastructure. Samsung Electronics and SK hynix have thus gradually transformed from traditional semiconductor cyclical stocks into core beneficiaries of global AI capital expenditures.

This narrative has driven both companies’ market caps to grow rapidly. By late May 2026, the combined market cap of Samsung Electronics and SK hynix at one point accounted for about half of the Korean stock market, and KOSPI’s gain for the year had once been close to 95%, after having already surged significantly in the prior year.

When the index weights of these two stocks reach such a level, KOSPI is no longer a basket of relatively diversified Korean companies—it becomes more like a concentrated investment product dominated by just two storage-chip firms.

This means:

If Samsung and SK hynix rise, KOSPI could quickly hit new highs; if either of the two falls, even if other sectors such as banks, automobiles, telecom, and consumer goods remain stable, they would still struggle to support the entire index.

Some international investors may even indirectly increase their South Korea AI-chip exposure through emerging-market index funds without realizing it. MSCI classifies Korea as an emerging market, while FTSE classifies Korea as a developed market. Therefore, the two “emerging market ETFs” tracking different indices may produce significant return differences depending on whether they include Korea.

Passive investing is not truly “neutral” in this context. Index construction rules, country classification, and weight changes can all turn a seemingly diversified fund into a highly concentrated AI thematic trade.

2. What truly amplifies the rally isn’t only AI, but also leveraged products

If only Samsung Electronics and SK hynix were rising, the Korean stock market might merely be a market with elevated valuation and high concentration.

What truly causes volatility to surge sharply is the large-scale participation of products such as single-stock leveraged ETFs.

A normal ETF typically tracks a basket of stocks or an index. A single-stock leveraged ETF is different: it may use just one stock—such as Samsung Electronics or SK hynix—as the underlying, aiming for double the stock’s daily percentage gain/loss.

For example, if the target stock rises by 5% that day, a 2x long ETF would theoretically rise by about 10%; if the stock falls by 5%, the ETF could also fall by about 10%.

These products are especially likely to attract retail investors who want larger gains with less principal, but they carry three risks that are often underestimated.

1. Leverage amplifies not only gains, but also the speed of losses

Investors often only see “double the upside,” but ignore “double the downside.”

When the underlying stock moves sharply up and down in succession, a leveraged ETF’s net asset value can shrink quickly. Even if the stock later returns to its original price, the leveraged ETF may not recover its starting value.

Assume a stock first drops by 10%, then rises by 11.1%, roughly returning to the starting point.

A 2x long ETF could first fall by 20%, then rise by about 22.2% at the lower net value, and still end up below the initial level. This is volatility drag. The more violent the market swings and the longer the holding period, the more pronounced the drag tends to be.

So a goal of 2x daily returns does not mean long-term returns will equal two times the stock’s cumulative price gains.

2. The ETF’s daily rebalancing can end up hitting the underlying in the opposite direction

To maintain its target risk exposure, a leveraged ETF typically needs to adjust its derivatives or stock positions before the market closes each day.

When the underlying stock rises, the fund may need to keep adding long exposure; when the stock falls, it may be forced to cut positions.

This creates a momentum-trading mechanism:

Buy when it goes up; sell when it goes down.

In a normal market, the impact of this kind of operation might be limited. But when single-stock leveraged ETFs are huge in scale and the underlying is concentrated in a small set of leading companies, rebalancing trades can end up amplifying price swings in return.

From the perspective of market microstructure, this kind of daily rebalancing has characteristics similar to the “short Gamma” effect commonly seen in derivatives markets: when the market rises, related products must buy more; when the market falls, they must sell more—turning what should provide a buffer into a volatility accelerator. In Korea, dozens of trading pauses occurred during 2026, far more than in the previous year.

3. The same risk gets packaged into different products

Investors may hold at the same time:

• Shares of Samsung Electronics or SK hynix;
• Index funds that track KOSPI;
• Emerging-market ETFs that include Korean stocks;
• Semiconductor sector ETFs;
• Single-stock 2x leveraged ETFs;
• Stocks purchased via margin loans using securities financing and lending.

From an account interface, these seem like multiple different products; but from the underlying risk, they may all concentrate on the same variable:

Whether global AI data center capital expenditure can keep growing.

Once the market starts doubting this variable, multiple products can fall together. In the end, “diversification” may just be betting on the same theme under different names.

3. Why Korean retail investors are so willing to take on leverage

Simply attributing this rally to “retail greed” cannot explain the whole issue.

Behind Korean investors’ high risk appetite are deeper economic and social reasons.

Korea’s real estate prices—especially housing prices in Seoul—create a relatively high wealth barrier for younger people. At the same time, competition for jobs, slower income growth, education spending, and household burdens reduce the appeal of accumulating long-term wealth through wages.

When traditional upward mobility paths become narrow, stocks, crypto assets, and leveraged products are easier for some young people to see as tools that can rapidly change their wealth situation.

In recent years, the scale of margin trading in Korea has kept expanding. Some younger investors have relied heavily on margin trading to participate in the stock market. Even after experiencing large investment losses, they still tend to borrow more and add leverage. At one point, Korean investors’ debt exceeded 60 trillion won, reflecting that high-leverage speculation is no longer just the behavior of a few people.

This phenomenon can be summarized as a form of “risk-taking driven by wealth anxiety”:

When people believe normal savings cannot keep up with house prices and asset prices, the opportunity cost of taking risks declines.

Without leverage, they may believe they could never catch up; with leverage, even if they fail, it at least offers the imagination of a quick comeback.

This also helps explain why the effect of risk warnings is often limited. For ordinary investors, regulators see “high-risk products,” while participants may see “the last ticket to board.”

More notably, leveraged trading itself can create social mimicry.

At the beginning of a bull market, a small number of investors earned high returns through chip stocks. After these stories spread via social media, investment groups, and short videos, they attracted more people to enter. Latercomers, to replicate early investors’ gains, often must accept higher prices and higher leverage.

As a result, the market gradually forms a dangerous feedback loop:

Rising stock prices create wealth stories

Wealth stories attract new capital

New capital pushes stock prices higher

Higher prices force latecomers to use more leverage

Leveraged capital further amplifies the rally

During the rising phase, the loop looks almost perfect; but once prices reverse, it flips quickly.

4. Why even great performance can’t save the stock price

The most confusing point about this semiconductor-stock correction is that company fundamentals did not deteriorate immediately.

Demand for AI servers remains strong, and profit expectations for chip companies are still high. Some market forecasts show that semiconductor companies could contribute a large portion of quarterly profit growth in the S&P 500, and in some cases year-over-year profit growth for certain companies could even reach triple digits.

However, even after companies report solid earnings, the stock price may still fall.

The reason is that stock trading has never been simply about whether results are “good,” but rather:

Whether the actual outcome exceeds the extremely high expectations that the market has already priced in.

When a stock rises by several times in a short period, the market price may already embed multiple optimistic assumptions:

• AI data center spending will grow rapidly over the long term;
• HBM will continue to be in persistent shortage;
• Chip prices will remain high;
• Company profit margins will keep expanding;
• Competitors cannot quickly add supply;
• Cloud computing companies will not cut capital expenditures;
• Technology routes will not undergo major changes.

As long as any of these assumptions weakens, the valuation can be marked down.

This is also why the market can show a pattern of “earnings rising while stock prices fall.” Investors are not denying current profits; they are reassessing the durability of future profits.

The semiconductor industry is especially prone to this issue.

Because chip production requires massive capital investment in advance. When demand is booming, each company increases equipment and capacity; but when new capacity actually comes online, market demand may already have slowed. As supply-demand shifts from shortage to oversupply, product prices and profit margins can drop quickly.

AI demand may extend this industry upcycle, but it may not be able to fully eliminate the cyclical nature of semiconductors.

In other words, AI can change the level of demand, but it doesn’t necessarily cancel the supply response.

5. From bull market to bear market—why the switch happens so fast

A highly concentrated market has a typical feature: both rallies and selloffs become faster.

During the up phase, Samsung Electronics and SK hynix keep increasing their KOSPI weights. Passive index funds that need to track the index have to buy more of the related stocks; active funds fear lagging their benchmarks and also add allocation; retail investors, seeing the rally, then chase the move via leveraged ETFs and margin accounts.

This creates multi-layer capital resonance.

But when the market turns, nearly all participants face the same question: who will take the other side?

In June 2026, after Korean regulators issued warnings about leveraged ETF risks, KOSPI fell by nearly 10% in a single day. Afterwards, the index retreated more than 20% from its June peak at one point and entered a technical bear market zone.

Price volatility of SK hynix after it listed in the United States further intensified instability in the local Korean market. The company’s share price saw a record single-day drop and, together with Samsung Electronics, dragged KOSPI down by about 9%, triggering 20-minute trading pauses. Storage-related stocks in the U.S., such as Micron, SanDisk, and Western Digital, also fell in sync.

This shows that the turbulence in Korea’s stock market is no longer a closed domestic event.

When a Korean chip company simultaneously has Seoul-listed shares, U.S. depository receipts, options, and leveraged ETFs, new arbitrage and risk transmission channels form across markets. A decline in the Seoul market can affect U.S. semiconductor stocks; changes in sentiment in the U.S. after-hours can then hit the Korean market the next day.

Especially when the supply of securities, conversion mechanisms, and shorting conditions are not fully symmetric between the two regions, prices can temporarily detach from fundamentals.

At the peak of market sentiment, SK hynix’s U.S. depository receipts at one point traded at a premium of as much as about 51% relative to the Seoul-listed shares. Normally, arbitrage trading would quickly eliminate obvious price gaps between the same company’s securities listed in different markets, but with limited security supply, difficulty in shorting, and restrictions on cross-border conversion, the gap cannot converge in time.

This premium itself is a signal of market overheating.

It means investors are not just buying the company’s future cash flows, but also a scarce trading instrument within a specific market.

6. Why regulators are forced to make an abrupt turn

Korean regulators face a typical dilemma:

They want to develop the capital market and increase trading activity, yet they must prevent financial innovation from evolving into systemic risk.

In the past, Korea did not allow domestic funds to directly launch single-stock ETFs. In early 2026, Korea’s financial regulators still stated that under the then-diversified investment rules, ETFs had to hold at least a certain number of underlying assets and single-underlying weight was limited, so the country could not issue a traditional single-stock ETF.

Then policies were loosened, and related high-leverage products expanded rapidly. But after the market experienced severe volatility, regulators quickly changed course.

Korean financial regulators subsequently announced restrictions or pauses on listing new single-stock leveraged ETFs, raised the minimum margin requirements for some investors, and strengthened risk training.

The regulator even made a rare admission that the rollout process for these products was too rushed. The head of Korea’s financial supervisory agency compared certain securities firms to intermediaries providing advice in a “gambling scenario,” worrying that platforms offering these products could earn stable profits, while ordinary participants bear most of the losses.

But whether these measures can truly reduce volatility remains in question.

First, limiting new products does not make existing products disappear. Existing leveraged ETFs still have to rebalance daily, and existing margin accounts won’t immediately de-lever.

Second, raising the entry threshold may reduce some new investors, but it cannot change Samsung Electronics and SK hynix’s high weight in the index.

Third, the stricter domestic restrictions may push investors toward overseas markets to trade similar products. In the U.S., multiple 2x leveraged ETFs tracking SK hynix’s U.S. depository receipts have already emerged.

Therefore, risk may not be eliminated—it may be transferred from domestic Korean exchanges to cross-border accounts, derivatives, and U.S.-listed products.

What regulators truly need to address is not simply whether “leveraged ETFs exist,” but three deeper issues:

1. Whether a highly concentrated index is suitable for supporting large-scale passive capital;

2. Whether leveraged funds’ rebalancing will create a significant impact on underlying stocks;

3. Whether retail borrowing, securities-firm financing, and derivatives risks can be monitored in a unified way.

7. Could a systemic financial crisis occur in Korea’s market?

High volatility does not necessarily equal a financial crisis.

A sharp fall in stock prices first causes investors’ wealth losses. Only if losses further transmit to banks, brokerages, corporate financing, and household consumption could it escalate into systemic risk.

Korea’s current risks mainly exist along three transmission paths.

First path: forced liquidation of margin accounts

When investors borrow money to buy stocks and the stock price falls below the maintenance margin requirement, brokers demand additional funds. If investors cannot meet the margin calls, brokers will force-sell.

Forced selling drives the stock price down further, triggering more account liquidations, thereby forming a loop of “price decline—additional margin calls—forced selling—price decline again.”

Second path: weakening of the household wealth effect

If large numbers of households suffer losses in the stock market, consumption may be reduced, home purchases postponed, or other spending lowered.

Because retail participation in Korea’s stock market is high, the impact of stock declines on households’ psychology and consumption behavior may be more visible than in markets dominated by institutions.

Third path: risk exposures of financial institutions

Brokers not only provide margin financing, but also issue and sell structured products, ETFs, and other derivative instruments. Even if product risk is nominally borne by investors, in extreme markets, liquidity gaps, counterparty risk, and client defaults may still return to the balance sheets of financial institutions.

However, Korea’s market is currently closer to a high-leverage asset price adjustment, rather than a full-scale banking crisis already forming.

To judge whether risk is escalating, what matters is not single-day moves, but the following indicators:

Whether margin balances are steadily rising or rapidly contracting;

Whether brokerages show liquidity stress;

Whether household loan default rates are rising;

Whether corporate financing costs are clearly increasing;

Whether the won experiences disorderly depreciation;

Whether foreign investors keep withdrawing continuously and in large amounts;

Whether semiconductor companies’ capital expenditures and orders deteriorate materially.

If a stock market decline mainly digests overvaluation while the banking system remains stable, it is more likely to be a painful but controllable de-leveraging process.

If stock losses simultaneously trigger margin defaults, liquidity stress at financial institutions, and currency depreciation, only then does the risk character clearly change.

8. Why foreign capital and domestic retail stand on opposite sides of the trade

Korea’s capital flows also show clear divergence: domestic retail keeps buying, while overseas investors tend to reduce positions.

This does not necessarily mean foreign investors are “bearish on Korea’s long-term prospects”; it may also reflect different participants’ risk management approaches.

Overseas institutions typically need to consider:

• the weight of Korean stocks in global portfolios;
• won exchange-rate risk;
• correlation between the semiconductor sector and U.S. tech stocks;
• concentration in a single market and a single industry;
• fund volatility and drawdown limits.

When Samsung, SK hynix, and U.S. semiconductor stocks rise in sync, the global funds’ overall AI exposure may already be too high. Selling Korean stocks is sometimes simply to reduce concentration risk within the portfolio.

Korean retail, by contrast, is more likely to view chip leaders as a combination of national competitiveness and long-term AI trends, based on opportunities in their own local market and personal wealth.

So the same company may carry different meaning to both sides:

To Korean retail investors, it is the most familiar domestic asset with growth potential; to global institutions, it may just be one part of the already overcrowded AI trade.

This difference in perception helps explain why retail may keep catching falling knives while foreign capital keeps selling.

The problem is that when domestic buy pressure relies mainly on margin financing and leveraged products, retail’s ability to absorb shares is not infinite. Once confidence declines or financing conditions tighten, the buy pressure that had seemed stable can quickly turn into sell pressure.

9. What Korea’s experience means for the U.S. market

Korea’s market size and structure differ from the U.S., so you cannot simply assume that what happened in Korea will be replicated on Wall Street.

The U.S. market is deeper and broader, and the types of financial institutions and companies are more diverse. Two companies alone cannot occupy half of the major index in the way Samsung Electronics and SK hynix did in Korea.

But Korea’s market still offers three important warnings.

1. Leveraged ETFs can turn a local rally into a market-structure problem

U.S. leveraged ETFs have also been growing quickly in assets, with technology and semiconductor products making up a large share. Based on public information, U.S. leveraged ETF assets have reached a record level of about $600k, with technology and semiconductors-related funds making up a significant portion.

When large amounts of capital concentrate on NVIDIA, Tesla, semiconductor indexes, or other popular stocks, daily rebalancing may similarly intensify end-of-day trading and short-term volatility.

2. Index concentration does not equal risk diversification

Although major U.S. indices include hundreds of companies, the market-cap-weighted mechanism causes the technology companies with the biggest gains to take on increasingly higher weights.

Investors may hold S&P 500 funds, Nasdaq funds, technology ETFs, semiconductor ETFs, and AI thematic funds—seemingly diversified, yet in reality they depend heavily on a small number of large technology companies.

Korea simply pushed this concentration risk to a more extreme, and easier-to-observe, level.

3. Valuation risk often appears when fundamentals are at their best

A bubble does not necessarily form around fake companies with no revenue.

The truly dangerous rallies often form around excellent companies, real demand, and high-speed growth. Because fundamentals are indeed strong, investors are more likely to believe that any price is justified.

Samsung Electronics and SK hynix are not “no-profit” concept stocks. They have real technology, production capacity, customers, and cash flows.

But even great companies can trade at excessively high prices. Between a good company and a good investment, there is always valuation.

10. The core of this turbulence isn’t whether AI is real—it’s price and structure

When discussing Korea’s stock market, the easiest trap is to fall into two extremes.

One view says AI is a revolutionary technology, so any drop in chip stocks is a buying opportunity; the other says after a surge in stocks, there must be a bubble, and AI investment will eventually collapse across the board.

Both judgments are too simplistic.

AI demand may grow for a long time, and Samsung Electronics and SK hynix may continue to benefit; but the existence of a long-term industry trend does not mean stocks are worth buying at any price.

Similarly, a large stock price correction does not necessarily mean the AI industry logic has completely broken. It could just mean the market is reverting from an extremely optimistic state to more reasonable expectations.

The real problem exposed by the Korean stock market can be boiled down to four types of concentration:

• Index concentration: two chip companies determine most of the market’s direction;
• Industry concentration: South Korea’s economy and exports rely heavily on semiconductors;
• Product concentration: many ETFs and derivatives point to the same group of stocks;
• Behavior concentration: retail investors, index funds, and momentum/trend capital make similar trades at similar timing.

Any one of these concentrations does not necessarily cause a crisis. But when all four happen at the same time—and are added on top of margin borrowing and daily rebalancing mechanisms—the market loses its buffer layer.

When prices rise, no one wants to sell; when prices fall, no one wants to buy.

Conclusion

The extreme volatility in Korea’s market cannot be simply attributed to an AI bubble, nor can it be entirely blamed on retail investors or leveraged ETFs.

It is the result of multiple forces acting together:

AI infrastructure buildout creates real and strong chip demand; the market-cap rise of Samsung Electronics and SK hynix makes KOSPI highly concentrated; index funds and overseas capital further reinforce the uptrend; single-stock leveraged ETFs and margin trading convert a normal move into mechanical chase-and-sell behavior; and housing pressure and wealth anxiety provide ongoing social foundations for high-risk speculation.

From this perspective, Korea’s stock market is not an isolated abnormal market—it is a magnifying glass.

It has shown in advance the new kinds of financial risks that may emerge in the AI investment era:

When real technological progress, massive capital expenditures, passive investing, derivatives innovation, and personal wealth anxiety combine, even a market centered on the highest-quality companies can form an extremely unstable price structure.

What will determine KOSPI’s direction in the future is not only how much Samsung Electronics and SK hynix can earn in the next quarter, but also three more important questions:

Whether AI capital expenditures can keep delivering; whether leveraged capital can exit in an orderly fashion; and whether regulation can cut the positive feedback between product rebalancing, margin trading, and stock price volatility without choking off market innovation.

If these issues are eased, the adjustment in the Korean market may ultimately become a process of normalizing valuations.

But if investors continue to chase highly concentrated assets with higher leverage, then every rebound may just be fuel for the next round of volatility.