ConsultingAI

Between two Samsung lines, the eyes post: the shortage is now a payout story

Short answer

We could not find a quoted post or an image attached to this one — on the account mirrors it is exactly two emojis, no text. What it sits between is more informative than what it says. Jukan posted it during Samsung’s Q3 call, immediately after “the memory shortage will worsen in 2027” and next to the preferred-share buyback remarks. The subject that repriced that morning was not demand. It was the return plan: KRW 90–110 trillion of payouts announced on top of a market that is sold out through 2028.

👀👀 @jukan05 on X

What the post is, and what it is not

Jukan’s timeline that hour is a live transcription with commentary: the shortage worsening in 2027, the preferred-share buyback comment, the eyes post, then the 2028 line and the 10+10 LTA disclosure. A wordless post in the middle of that run is normally a reaction — to a print, a chart or a number that surprised him. We checked the post through public mirrors and through three research rounds aimed at finding the quoted post, its replies and any image, and found no attachment or quoted status. So the honest version is: unknown subject, known context. We are writing about the context, not inventing the subject.

The number that got repriced

The company’s own release is more precise than the headline. Samsung estimates KRW 90–110 trillion of total shareholder returns for 2026, of which KRW 35 trillion is already planned or completed, leaving KRW 55–75 trillion to be returned through a “substantial share repurchase and cancellation program” plus cash dividends. The third-quarter cash dividend alone is KRW 4,601 per common and per preferred share, KRW 30 trillion in total. The release also acknowledges “the growing market interest in the price gap between common and preferred shares” and says the company will actively consider adjusting the preferred share in its buyback and cancellation programme, with the final amount and execution set at a board meeting in late January 2027 (Samsung Electronics, Q3 2026 results). A September note had already flagged the mechanics: preferreds trading about 26% below the common, and a programme reported at KRW 110 trillion, about $81.8 billion (@FirstSquawk, 8 Sep 2026).

Consensus, as FnGuide had it in early October, put Samsung at KRW 108.7 trillion of third-quarter operating profit and SK hynix at KRW 77.2 trillion, both near eight times a year earlier, with buyback-type buyers absorbing about KRW 31.3 trillion of KOSPI in September (@Ricci_nov). The shares slid anyway, because the programme was delayed in structure, not in size: a record payout with the allocation postponed to January gives the market a number without a mechanism.

Why the payout and the shortage are the same story

A memory maker can only return cash at this scale if it is not funding the next two fabs out of its own pocket alone. That is what the contracts changed. With 50%+ of 2027–28 capacity sold and deposits in hand, capex is partly pre-financed by the customer, which is precisely what lets Samsung raise capex to KRW 84.4 trillion and still hand back up to KRW 110 trillion. The buyback is not a signal that management thinks the cycle is ending. It is a signal that management thinks the cycle has been de-risked enough to be borrowed against.

The peers are doing the same arithmetic. SK hynix announced a KRW 100 trillion return plan alongside its record quarter, with more than 50% of its 2027–28 capacity under long-term agreements (@quantum_macro). Two suppliers, two record payout programmes, one cause: the cash arrives before the depreciation does.

The bear reading, and why it is weaker than it sounds

The obvious objection is that capital discipline at a cyclical peak is what tops look like, and that money spent on preferred-share cancellation is money not spent on wafer capacity. It is a fair objection; it is also measurable. The capacity is being built anyway — KRW 52.8 trillion of Samsung’s semiconductor capex was already spent through Q3 — and the bits arrive on a construction calendar, not a shareholder calendar. What would make the bear case right is a demand break, and the shortage is currently being defended by deposits rather than by forecasts.

The cleaner criticism is about structure, not direction: a programme whose allocation is decided in January 2027, after full-year earnings, is a programme the company can still shape. That is exactly why the stock slid, and exactly why January is now a real catalyst rather than a formality.

What to watch in January

Three things. First, how much of the KRW 90–110 trillion is buyback and cancellation versus dividend, because the first two change the share count and the third does not. Second, whether the preferred-share buyback becomes a commitment; that is the part that closes the 26% discount. Third, whether the payout is funded from cash flow or from the deposits customers paid for 2027–28 capacity — that distinction tells you how comfortable the company really is with the next two years of capex.

What we could not verify

The subject of the eyes post. The final size and allocation of the return plan, which Samsung has postponed to January 2027 and which is reported in a range. The exact preferred-share discount, which moves daily.

Sources

FAQ

Was the eyes post about the buyback?

We do not know. It followed the preferred-share comment on the timeline, but there is no quoted post or text to confirm it. Treat it as a reaction to the call, with the subject unconfirmed.

Does a record buyback mean Samsung expects prices to fall?

Not necessarily. It means the company believes its contracted cash flows can carry both the capex plan and the payout. That is a statement about visibility, not about price direction.

Why buy preferred shares rather than common?

Preferreds trade at a discount to common with no extra votes. Buying them in adds per-share value without changing control, which is why the market watches that option closely.

What does this mean for buyers of memory?

That the supplier’s balance sheet is no longer the constraint. The constraint is wafers, panels and time, which is a harder thing to negotiate against.

Building something like this?

I help teams ship agents, RAG and model evaluations that hold up in production.

Work with me →
nextSamsung’s HBM4 sold at “target price” is the memory bill for Rubinkeep scrolling