"Why don't you buy twice as much long position on SK Hynix?"

This month in Korea, if you are neither an SK Hynix employee nor own SK Hynix stock, then you are probably a "unfortunate soul."

After the quarterly report revealed huge profits, the investment banks, not shy about making a fuss, not only actively revised SK Hynix’s profit expectations for this year but also raised employees’ expectations for year-end bonuses. They used the principle of allocating 10% of operating profit each year as a bonus pool to calculate this year's per capita year-end dividend, which amounted to several million RMB. They also conveniently threw Samsung’s capitalists into the fire of injustice and unrighteousness.

Since then, anything related to SK Hynix IP has received enthusiastic attention.

SK Hynix’s work uniforms have become the preferred pass in Korea’s matchmaking market; the real estate agents in the headquarters city of Icheon have enjoyed a dreamy quarterly, with property prices and transaction volumes rising across multiple areas along SK Hynix’s commuter bus routes; the semi-Korean Chinese semiconductor ETF, which is on the edge, has also been pulled up to a 30% premium, with occasional temporary suspensions.

Even the Hong Kong stock market, long criticized for lacking technological content, has shown some signs of life.

As of May 13, 2026, the Southbound East Asia SK Hynix Daily 2x Leverage ETF (07709.HK) listed on the Hong Kong Stock Exchange, with assets approaching HKD 60 billion, surpassing Tesla’s long-dominant 2x long ETF (TSLL.NASDAQ) in the US stock market, and ranking first globally in the scale of single-stock leveraged derivatives.

No matter how niche an investment product is, once the market surges to this level, any online wave, even just updates from tech digital bloggers, can often be met with enthusiastic comments from netizens—"Why don’t you buy 2x long SK Hynix?"

Deadly leverage

On October 16, 2025, when the 2x long SK Hynix ETF just listed on the Hong Kong Stock Exchange, its issuance scale was less than HKD 5 billion. If calculated based on the closing price on May 13, 2026, in just 7 months, this leveraged ETF’s net asset value increased by 1011.58%, and its scale soared more than 13 times.

On the same day, the cloud-based technology company Yunji Technology, claiming to be the “first hotel robot stock,” had its stock price already climbing steeply, with a market value less than four times its IPO.

You might say this is the terrifying efficiency of 2x leverage. The underlying stock of SK Hynix listed in Korea from October 17 last year to May 13 this year only gained a “total” increase of 324.49%. Under the support of a unilateral bullish wave, this leveraged ETF’s deviation even exceeded the theoretical double return by 362%, delivering an excess return. In the face of such violent money printing, it seems more appropriate to describe it as “triple leverage.”

But if we look at the past 7 months in a broader perspective, this kind of on-paper excess is actually temporary.

Just two months ago, the Strait of Hormuz fell into Schrödinger’s blockade, and global markets plunged into panic amid sudden oil and gas disruptions. Amid the swings and rapid changes of the situation, the market did not experience a traditional unilateral decline but instead fell into a state of mental schizophrenia during this atypical geopolitical conflict.

During the day, traders still adhered to the risk-averse logic of “three wars breaking out, supply chains breaking,” but at night, due to ambiguous statements from White House spokespeople, they could quickly switch to “conflict de-escalation, returning to the tech mainline,” engaging in a frantic short squeeze. The ambiguity and uncertainty of this evolution path, amplified by social media’s rapid dissemination, led to intense selling of tech stocks or frantic bottom-fishing during dips.

Although common sense tells us that the war will eventually end, and the daily token consumption in the AI industry continues to accelerate, when market volatility becomes too intense, the twists and turns of the process cannot be ignored.

More people felt the volatility loss of this leveraged ETF product during this period.

From the actual trading data between March and April 2026, SK Hynix’s stock price experienced sharp fluctuations and downward movements. Declines are problematic enough, but violent rebounds exceeding 10% multiple times made things worse.

For the 2x long SK Hynix ETF that rebalances daily, a unilateral decline might be tolerable, but high-volatility oscillations are the real meat grinder. During the most painful times, the ETF fell more than 50% beyond the underlying stock.

Without considering other trading and management fees, the daily rebalancing mechanism means that in a rising market, yesterday’s profits automatically become today’s “principal,” and with double leverage added on top, it yields even more excess positive returns. Conversely, during a sharp decline, because the daily calculation base shrinks, the actual loss will be less than the theoretical double.

However, once the market enters a “rise-fall alternation” volatile phase, the leverage ETF reveals its ferocious side.

The 2x long SK Hynix ETF repeatedly experiences “long and short kills”—after a big rise yesterday and adjusting positions, a big drop today deducts more blood, then rebalances, and when it rebounds tomorrow, it suffers again from the damaged base.

The repeated friction caused by rise and fall swings leads to net value drawdowns far exceeding twice the decline of the underlying stock, creating significant negative volatility drag, eroding investors’ principal.

But now, as the market returns to the AI mainline, with frantic capital again pouring in, it brings a happy scenario of one-sided surges.

As SK Hynix’s market value hits new highs and hundreds of billions of leverage ETF products ignite frenzied trading, the market inevitably returns to that perennial question: does this industry revolution truly have no cycle?

Silicon-based cyclical stocks

One must admit that, from the listing time point, the 2x long SK Hynix ETF is essentially stacked with the blessings of “fate, luck, and feng shui.”

For a long time before, storage was not the absolute focus of the secondary market’s bullish AI trend. After all, since humans boarded the information age train in the 1990s, storage has often been a place of bloody slaughter after a fiery boom, with cycles far more terrifying than dreams of growth.

Memory chips (especially traditional DRAM and NAND) are highly standardized commodities. Aside from different brand labels, the physical performance of memory modules from various manufacturers is almost indistinguishable, making them the “pork stocks” of the silicon sector. The entire industry has long been trapped in a brutal cyclical cycle:

Shortage and price increase → Giants expand production madly → Overcapacity → Price collapse → Losses and cutbacks → Shortage again.

Each upward cycle is often labeled a “super cycle” under overly optimistic expectations. Each downward cycle leaves behind a trail of bones in fierce price wars and billions in losses.

After experiencing the most brutal semiconductor winter from 2022 to 2023, the surviving memory giants—Micron, Samsung, SK Hynix—mutually cut capital expenditures, no longer maliciously expanding to kill competitors but avoiding self-destruction.

Image source: IC Insights

Then came the AI narrative, repeating the cycle of shortage and price increase, directly installing a printing press for money.

Especially since the second half of last year, the focus of AI industry competition shifted from “training” to “inference,” with infrastructure demand moving from “computing power” to “storage capacity,” supply bottlenecks shifting from bandwidth to capacity, and the widespread shortage of storage becoming the hottest trading narrative.

Today, anyone still mentioning “isn’t the end of AI supposed to be electricity?” is probably missing the boat.

After Q3 2025, news about the AI industry almost always involves shortages of storage chips—sometimes giants announce HBM orders are booked into 2027 and beyond; other times, DDR5 supply is also tight, and prices are rising across the board, regardless of high or low-end.

As a supplier of NVIDIA’s HBM, SK Hynix has gained a significant first-mover advantage and market share, and the 2x long SK Hynix ETF, launched at the right time, almost seamlessly caught the wave of memory prices soaring to more expensive than gold, with a single box capable of buying a Shanghai house.

So, can jumping on the AI bandwagon free us from the cyclical pull? It’s not about drawing conclusions now but about exploring where the change will happen.

SK Hynix’s dominant position under the barrier of HBM yield rates saw its Q1 2026 gross margin reach a historic peak of about 79%, even surpassing NVIDIA’s profitability during the same period.

Human nature tells us that extreme excess profits will inevitably attract a flood of capacity expansion. The tacit understanding among memory giants to “cut production” in the face of super profits is untrustworthy.

Therefore, whether Samsung or Micron’s yield rates will break through at some future point to diminish the scarcity narrative of HBM, and the divergence between bulls and bears causing sector volatility, remains a variable to watch closely.

Besides supply-side changes, demand-side disputes have not disappeared despite the accelerated adoption of agents and increased token consumption.

Ultimately, SK Hynix’s frenzy is built on NVIDIA’s madness; and NVIDIA’s madness is built on the annual hundreds of billions of dollars in AI capital expenditure by downstream giants.

The marginal change in Capex remains the greatest driver of all anxieties and pride in the secondary market regarding AI.

Epilogue

Buying or not buying the 2x long SK Hynix ETF will both serve as a subtle footnote when we look back on this period of history.

In this era, the debate between bulls and bears often boils down to two things: belief in the AI industry, and concerns about macro geopolitical risks.

People habitually open history books, trying to find parallels in the internet frenzy of the millennium or earlier macro upheavals. But each technological revolution unfolds differently, and this time, the “unlike before” lies in the unprecedented speed of industry upheaval.

AI is reshaping global productivity and social relations at an unprecedented acceleration. This extreme “speed” breaks the long, slow process of technological cycle penetration and fermentation. It hardly gives the market time to digest valuations gradually, nor does it allow “old-timers” a chance to enjoy rotation from liquidity flooding.

Whether it’s industry giants or secondary market funds, they are forced to make quick decisions and price within a very short window. As a result, the unit of stock price increase becomes how many times; seasoned AI practitioners have already accepted that in this era, six months is already considered long-term.

However, the storm over the Strait of Hormuz still places this round of technological revolution under the commonality of all past cycles: industry determines the ultimate outcome and returns, while macro factors influence the path and volatility—what causes the huge negative deviation of the 2x long SK Hynix ETF is not the interruption of AI progress, but the extreme swings in global macro expectations during that month.

And the real-world vulnerabilities are not only the 33 kilometers at the narrowest point of the Strait of Hormuz.