Samsung's profit is set to jump 18-fold and SK Hynix wants $28bn from Wall Street — the AI memory boom is now carrying whole economies

Finance News 3 min 59 sources

The two companies that make most of the world's advanced memory chips are riding a wave of AI demand so large it is reshaping their profits, their listing plans, and South Korea's public finances.

Key takeaways

Samsung's quarterly profit is set to jump 18-fold and SK Hynix is seeking about $28 billion from a US listing, both riding a surge in AI memory-chip demand.
South Korea is building a national "future fund" from its chip windfall — a country treating a private boom as public money, which only holds while the boom does.
The gains are narrow: a few AI-linked names are carrying markets, so pension and index funds rise with them and would fall with them too.

The clearest money story this week isn’t a rate cut or a market crash. It’s how much of the world’s prosperity now rests on a single, narrow source of demand: a handful of companies buying memory chips to build artificial intelligence.

The AI memory wave, in numbers

Samsung, the South Korean electronics giant, is likely to report an eighteen-fold jump in operating profit for the last quarter, driven by surging demand for the memory chips that AI systems need [12]. That is not a typo — profit is expected at roughly eighteen times what it was a year earlier.

Its main rival, SK Hynix, is going further. The company plans to launch a listing on a US stock exchange to raise about $28 billion, one report put the figure near $29 billion, explicitly to get closer to the American investors pouring money into AI [14][13][1]. A listing is when a company sells shares to the public to raise cash. SK Hynix already trades in Seoul; adding New York is a bid to sit where the AI money is.

Memory chips are the part of a computer that stores and moves data. The advanced kind — called high-bandwidth memory — is what lets AI systems handle enormous amounts of information at once. Two companies, Samsung and SK Hynix, make most of it. When AI demand surges, their fortunes surge with it.

The wave reaches beyond them. Foxconn, the Taiwanese firm that assembles much of the world’s electronics, said its second-quarter revenue jumped 40% from a year earlier [9]. Across Asia, shares edged higher on Monday as investors positioned for a run of earnings reports [3], after a strong week on Wall Street [0].

When a boom becomes a nation’s budget

The most telling detail sits in a government plan, not a market. South Korea said it will create a “future fund” from the windfall its chip industry is generating, aiming to spread the gains — to spur growth and tackle inequality [41].

That is a country treating a private industry’s boom as public money. It works only while the boom lasts. South Korea’s economy leans heavily on semiconductors; when memory demand is strong, exports, profits, and now a national fund all rise together. The same concentration that makes the good years very good makes the country exposed if AI spending cools.

For anyone with savings in a pension or index fund, the practical edge is quieter: a growing share of the market’s recent gains has come from AI-linked companies. The strong week on Wall Street [0] was not evenly spread. When a few names carry the index, your fund rises with them — and would fall with them too.

The rest of the ledger

Away from chips, the day’s other moves were smaller but worth marking.

The US dollar sat near a two-week low as bets on further interest-rate hikes faded, with Japan’s yen still under pressure [4]. When investors expect fewer rate rises, a currency tends to weaken, because higher rates are part of what makes holding it attractive.

Oil slipped after OPEC+, the group of major oil-producing nations, agreed to raise output targets again [32][30]. More supply, softer prices — the same dynamic that has kept oil under pressure for weeks.

Dealmaking stayed busy. EasyJet, the UK budget airline, agreed in principle to a takeover by investment firm Castlelake at £6.90 a share, valuing it at about £5.5 billion [35][29]. Lockheed Martin is leading the race to buy naval-defence firm Ultra Maritime for about $3.5 billion [8]. And Uber paused its European food-delivery expansion as it pursues a takeover of Delivery Hero [16].

02 · Lesson · why it matters

The tide that lifts you is not yours to steer

When your good fortune is powered by a demand you don't control, the size of the boom and the size of the risk are the same number.

A record profit that no one at Samsung caused

Samsung is about to report profit eighteen times higher than a year ago. Its factories didn’t get eighteen times better. Its managers didn’t have an eighteen-times brighter year. What changed sits outside the company entirely: a small number of firms decided to spend enormous sums building artificial intelligence, and AI runs on memory chips, and Samsung makes memory chips.

The profit is real. The money will land in real accounts. But it arrived on a tide, and the tide was made somewhere else.

This is worth naming clearly, because a boom this large tempts everyone to tell a story about skill — our chips, our strategy, our moment. Some of that is true. Most of it is timing.

The pattern: borrowed momentum

Here is the force underneath. Some of what looks like your success is really the shadow of someone else’s decision to buy.

A restaurant on a suddenly-fashionable street is packed every night. The owner works hard, and also the street chose him. A landlord’s rent doubles because a big employer opened nearby. A supplier’s order book triples because one customer scaled up fast. In each case the good fortune is genuine, and in each case it is powered by a demand the person didn’t create and can’t switch back on if it stops.

The trap isn’t enjoying the boom. The trap is mistaking borrowed momentum for the durable kind — and then building as if it were yours to steer.

The clean signal that momentum is borrowed: you can name the outside decision that started it, and you have no vote in whether it continues.

The size of the wave is the size of the drop

There’s an uncomfortable symmetry here. The same concentration that makes the boom enormous makes the exposure enormous. They are not two facts. They are one fact seen twice.

If a few AI buyers drive most of memory’s demand, then memory profits soar when they spend — and sag when they pause. An eighteen-fold rise is only possible because the demand is that narrow and that intense. Narrow and intense is exactly the shape that can reverse. You don’t get the giant upside without carrying the matching downside; the market is not offering one without the other.

The mistake is reading the giant number as safety. Big and stable are different things. A boom driven by one source is big because it is not spread — and unspread is the opposite of safe.

Who’s holding this without knowing

Now the part that reaches past the companies.

South Korea is building a national fund out of its chip windfall — public money raised on a private boom. That fund’s health is now tied to whether a handful of foreign firms keep buying memory. A nurse or a bus driver in Seoul, who has never bought a chip, is now quietly downstream of decisions made in a few American boardrooms.

And it reaches further, into pockets that feel nothing to do with chips. If you hold a pension or an index fund, you already own a slice of this. Markets have leaned hard on a few AI-linked names; when they rise, your fund rises, and it feels like the whole economy is doing well. It isn’t the whole economy. It’s a narrow current, and you’re standing in it whether or not you chose to.

That’s the reach of borrowed momentum: it pulls in people who never made the bet and can’t see that they’re holding it.

The arrangement that decided this before today

None of this was an accident of the week. It sits on a structure built over decades.

The world let memory-chip making concentrate into a couple of firms in a couple of countries, because concentration is efficient — it’s cheaper to make these chips at giant scale in few places. That efficiency is real, and it lowered prices for everyone for years. The same arrangement that delivered cheap chips also means a whole industry, and a national economy, now rise and fall on one wave of demand. Efficient and fragile were bought together, in the same decision, and the bill for the second half comes due only when the wave turns.

Nobody chose “be dangerously exposed to AI spending.” They chose “make chips as cheaply as possible,” and the exposure came attached — quietly, for free, until it isn’t.

On the whole

A record profit is easy to read as proof that you did something right. Sometimes it’s just proof that a tide came in. The honest question, when good fortune is this large, is not “how did we earn this” but “whose decision powers it, and what happens to us when they decide differently.”

That question doesn’t have a comfortable answer, and it isn’t meant to. It’s meant to make the number smaller in your mind — to hold the boom a little more loosely, knowing how much of it was never yours to steer, and how many people are standing in the same current without ever having seen it.

03 · Lab · your turn

Build on the Wave

Rehearse committing to a boom you don't control, and feel that the size of the upside and the size of the risk are one choice.

04 · Hope · carry this

A boom this large is proof of how fast people can build something the whole world suddenly wants. Our appetite for what comes next has never yet run out — and it's the thing that keeps pulling us forward.