Crazy! SK hynix ADR starts trading for three days, premium jumps to 51%; Barclays calls a target price of $330: the memory shortage has only just begun

SK Hynix’s U.S. stock market listing lasted only three days. The ADR premium surged from 3% at the IPO to 51%. It was then hit hard on the 13th by a South Korea sell-off wave, and on the 14th Barclays immediately called out a target price of $330.
(Background recap: SK Hynix’s ADR jumped 13% on day one, while the Korean home shares collapsed 15%! Will volatility keep getting bigger?)
(Additional background: IBM crashed 24% pre-market! The CEO warned that Q2 missed expectations, and customers pulled all their budgets to rush memory procurement.)

Table of contents

Toggle

• How did the 51% premium come about?
• Bubble or shortage?
• The math behind the memory shortage

For the same stock, on the 13th it was hit hard again by South Korea’s historic sell-off wave, falling 9.3%; yet the very next day, on the 14th, SK Hynix’s U.S. ADR closed up 27% to $193.92. After only three days of trading since the listing, the ADR premium versus SK Hynix’s local Korean shares had already ballooned from 3% at the IPO price to 51%.

How did the 51% premium come about?

In plain terms, an ADR is a certificate that lets U.S. investors buy shares in a foreign company using dollars without having to open overseas brokerage accounts specifically. The reason a premium appears comes down to structure: one SK Hynix ADR corresponds to only one-tenth of the ordinary shares listed in Seoul, and the ordinary shares cannot be converted back into ADRs—supply is effectively locked up, while dollar funds can only enter through this channel.

The market had already known there would be a premium, but no one expected it to run out of control to this extent.

Last weekend, the IPO size SK Hynix completed was $26.5 billion—one of the biggest deals in South Korea’s capital markets history. At the time, the premium at pricing was only about 3%, which was fairly restrained.

One possible catalyst is options: ADR options for this stock began trading yesterday on a U.S. options exchange, effectively opening a door for the world’s largest derivatives market, with capital and leverage rushing in. Once liquidity loosened, volatility was magnified many times over, instantly inflating a premium curve that was already on the high side to 51%.

Bubble or shortage?

The problem is that the day before, SK Hynix had just suffered a major sell-off, which also triggered a trading halt in the South Korean market; then, riding on the time difference into U.S. trading hours, the move kept hitting and pushed down SK Hynix’s ADR by 9.3% during the U.S. session. That also caused many conservative funds to re-check their positions.

To retail traders in the market, the day before they sold “the AI bubble,” and the next day they were hit with “a memory shortage.” These two contradictory narratives really leave investors feeling like they’re getting whiplash.

The math behind the memory shortage

Backing the “memory shortage” view is Barclays. It only officially added SK Hynix to its research coverage on Tuesday, and the moment it did, it issued a fairly bullish “overweight” rating, with a target price of $330. Analyst Gales (Simon Coles) believes that the persistent memory shortage that won’t go away across the tech industry will give this company leverage to raise selling prices—growing revenue by raising prices rather than relying purely on shipment volumes.

In a report to clients, Gales wrote:

“We believe that in the short term there is still some room for gross margin to rise; however, the biggest difference versus the Bloomberg consensus forecast is that, driven by rising prices for high-bandwidth memory (HBM) and SK Hynix’s strong market position, the revenue outlook for 2027 will be materially higher.”

What supports this argument is a supply-demand imbalance. Barclays’ global DRAM model estimates that in 2027, bit supply will only grow by about 20% year over year, while bit demand will accelerate to around 35%. This gap will “persist for several years,” and the shortage will only truly worsen in 2027; in 2028, it will only ease modestly.