Trium Capital bets SK Hynix’s 48% U.S. share premium will shrink
Tej Johar says U.S. demand for AI exposure has inflated the markup. His pitch depends on prices converging—not on a new breakthrough from the chipmaker.
Loading page…
Tej Johar says U.S. demand for AI exposure has inflated the markup. His pitch depends on prices converging—not on a new breakthrough from the chipmaker.
Listen to this story
At the Sohn London conference, Trium Capital portfolio manager Tej Johar pitched a relative-value trade in SK Hynix: buy its cheaper Seoul- or Frankfurt-traded securities and short the U.S.-listed shares. He argues the premium reflects U.S. demand for AI exposure and expects it to narrow, potentially as institutions switch into cheaper securities. Johar’s $8 billion “alpha opportunity” is his estimate of the trade’s potential, not realized returns; whether the gap closes remains uncertain.
Johar said the U.S. premium had widened from 13% after SK Hynix’s July Nasdaq ADR listing to about 48%.
On November 13, he compared the Seoul- and Frankfurt-traded securities at about $90 with U.S.-traded shares at $132.
Johar cited a roughly zero average ADR premium over the prior 25 years and said similar gaps had closed quickly.
Investors buying SK Hynix’s U.S.-listed shares are paying about 48% more than buyers of comparable securities in Seoul and Frankfurt, according to Trium Capital’s Tej Johar. He is betting that gap will shrink. CNBC reported on November 20 that the portfolio manager had presented a trade buying the cheaper securities and shorting the expensive U.S. shares.
Johar laid out the position at the Sohn London investment conference on Wednesday. He described it as a timely opportunity while AI investing continues to shape global markets, estimating an “$8 billion alpha opportunity”—his assessment of the potential opportunity, not a return the fund has already earned.
The South Korean chipmaker listed on Nasdaq in July through American Depositary Receipts, or ADRs. Johar said the U.S. premium over Seoul-traded shares and Frankfurt-traded global depositary receipts had widened from 13% to about 48% since that listing. His pitch targets that widening difference between markets, rather than a single price for the company.
Johar attributed the divergence to U.S. investors’ strong appetite for AI exposure. Investors elsewhere, he said, had not paid the same markup. That explanation makes the location of the investment central to his argument: enthusiasm for the same chipmaker has translated into different prices across its traded securities.
The structure pairs two opposing positions in SK Hynix securities. “It’s a really simple trade — buy the cheap one, short the expensive one,” Johar said. In hedge-fund language, he described a relative-value arbitrage trade: the intended opportunity is a narrowing price gap between comparable investments, rather than a standalone bet against SK Hynix.
Johar offered a historical comparison to support his confidence. He said the typical ADR premium over the preceding 25 years was about zero, calling SK Hynix an “extra special case.” He also said past gaps of this kind had closed quickly. Those are the historical grounds for his forecast, not evidence that this particular gap has already closed.
His case does not rest solely on history. Johar cited what he called the “shaky” AI trade as one possible force behind a narrowing premium. He also identified a more direct route: large institutional shareholders could move into the cheaper securities. That would make holders’ choices, not just changing sentiment toward AI, part of the proposed correction.
Johar went further, arguing that holders arguably have a fiduciary duty to switch. His challenge to institutions was blunt: “Paying 48% more for the same share is incredibly hard to justify to clients, to risk committees, frankly, to anyone who’s ever been in the investment business.”
Loading discussion...
Join the conversation
Explain what would make that premium acceptable—or unacceptable.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.