Since listing American depositary receipts in New York, SK Hynix’s shares have climbed sharply, with Barron’s describing American money as the Korean memory maker’s new growth engine. This week came the inevitable accompaniment to a hot stock: GraniteShares launched the first U.S. exchange-traded funds on SK Hynix, a two-times-long fund and a two-times-short fund, each with a fee of 0.50 percent.
The timing was awkward. The same day, a report from Bank of America cited by Chinese financial media warned that SK Hynix’s actual new production capacity by 2028 would amount to only about one-sixth of its original plan. The company, in other words, is being priced as the supplier of the AI era’s most wanted component, while the bank’s analysis suggests the capacity that would justify that price is not coming as fast as promised.
The gap between the two stories is the story. Investors have bid up SK Hynix on the strength of high-bandwidth memory, the specialized chips that sit beside AI accelerators and feed them data, and on the company’s position as the dominant supplier to Nvidia. American funds, newly able to buy the stock through depositary receipts, have added fuel. Meanwhile, the physical reality of building memory fabs, which take years and cost tens of billions of dollars each, moves at a slower pace than any rally.
Bank of America’s reading cuts to the supply question. Memory makers have a long history of promising capacity and then delivering less, in part because building fabs is hard and in part because managing supply is how they manage prices. A shortage of memory is profitable for the makers; a glut destroys their margins. Analysts said the gap between plan and delivery could reflect either execution problems or deliberate restraint, and the two have very different implications for the stock.
The new funds add a new element. Two-times-long and two-times-short funds reset their exposure daily, which makes them suitable for traders and unsuitable for most long-term investors, and their launch tends to coincide with the noisiest phase of a stock’s cycle. Their arrival does not move the underlying company, but it measures sentiment: new derivative products appear when demand for a name has become a crowd phenomenon.
HBM is the industry’s fastest-growing corner. The chips stack memory cells vertically and sit directly beside AI accelerators, and Nvidia, the dominant buyer, has contracted for years of supply in advance. SK Hynix was first to mass-produce the highest-bandwidth versions and has kept the lead through successive generations, which is why its shares trade on AI sentiment rather than on the broader memory cycle. That also makes the stock hostage to one customer’s fortunes, a concentration its new American holders may not fully price in.
The underlying business remains strong by most measures. SK Hynix has ridden the HBM boom to record profits, its technology has won the highest-margin contracts in the industry, and its American listing has given it a currency and visibility it never had in Seoul alone. The question was never whether the company is well positioned; it is whether the stock has gotten ahead of what the company can actually deliver.
Memory cycles have a rhythm that investors ignore at their peril. The industry’s history is a series of booms followed by corrections, each one triggered by capacity arriving faster than demand, and each correction wiping out a large share of the prior rally. The current cycle is unusual only in the size of the demand driver: AI has pulled forward years of memory consumption into a few quarters, and the market is paying prices that assume the demand never pauses.
The capacity shortfall, if real, cuts in the opposite direction. If SK Hynix can only add a sixth of its planned capacity, supply stays tight, prices stay high, and the current earnings power lasts longer than bears expect. The divergence between the stock’s price and the company’s delivery schedule, in other words, does not necessarily end in a crash; it could also end in years of scarcity pricing that the bears have underestimated.
The American listing has changed the shareholder base, and that matters for the cycle. U.S. funds tend to hold and trade more aggressively than Korean institutions, and the depositary receipts have brought the stock into indexes and portfolios that never touched it before. When the cycle turns, those investors will exit with the same enthusiasm they brought on entry, and the new funds will amplify the move in both directions.
Analysts said the test will come in the company’s capacity announcements over the next two years. Every fab expansion that comes online on schedule will support the bulls; every delay will feed the bears. In the meantime, the stock trades on a story that is partly about chips and partly about money flow, and the two do not always point the same way.
