Trump Should Buy The KOSPI Crash
What We're Buying Here
In the post below, we suggest that President Trump buys Korea's memory crash, and we offer reasons why. We're not buying more memory stocks until they clear our screens, though (maybe some will, after today's price action).
In the meantime, we are taking a flyer on a beaten-down AI bottleneck stock that's a 2x previous winner for us. If you want a heads up when we place that trade today, you can subscribe to our trading Substack/occasional email list below.
And if you're worried about market risk here (understandably so), you can hedged with the Portfolio Armor website or iPhone app.
A Chinese Debut Deepens Korea’s Rout
On Monday, Chinese memory maker CXMT made a strong stock-market debut. On Tuesday, Samsung Electronics fell 13.4%, SK Hynix fell 14.7%, and South Korea’s benchmark KOSPI fell 10.8%.
Those moves reflected real concerns. Investors are worried about Chinese competition, domestic deep ultraviolet lithography equipment, less compute-intensive Chinese AI models, and increasingly circular financing within the American AI ecosystem. CXMT’s debut showed that China can still mobilize enormous amounts of capital behind a strategic industry.
Leverage accelerated the decline. South Korean regulators had already moved to tighten single-stock leverage rules, including tripling the minimum cash deposit for investors trading leveraged single-stock products. Once Samsung and SK Hynix started falling, forced selling fed on itself.
The result looks like more than a Korean stock-market accident. It threatens to become an American strategic problem.
The Earnings Didn’t Break
The operating results released during this selloff look strikingly different from the price action.
Samsung’s full second-quarter results, released after the U.S. close Wednesday, showed 171.5 trillion won in revenue and 89.5 trillion won in operating profit. Revenue rose 130% year over year, while operating profit increased 1,814%. The company’s operating margin reached 52.2%.
The concentration makes the source of those profits clear. Samsung’s semiconductor unit generated 89.2 trillion won in operating profit on 127.5 trillion won in sales, a 70% margin. Memory revenue rose 471% year over year as the company achieved record bit sales in both DRAM and NAND. Samsung expects demand for server DRAM, enterprise solid-state drives, and HBM to keep the market undersupplied during the second half, supported by continued AI infrastructure spending.
SK Hynix also reported record results, began mass shipments of its latest high-bandwidth memory (HBM), and said it had secured long-term agreements with roughly ten customers.
CXMT presents a genuine competitive threat, particularly in commodity memory. Its rise could eventually affect industry capacity and pricing. Analysts still believe it remains years behind Samsung and SK Hynix in frontier HBM, however. That technological gap is one of the most important pieces of daylight between China and the allied AI supply chain.
One widely circulated forecast projects combined 2029 operating income of $945 billion for Samsung, SK Hynix, and Micron Technology: $420 billion for Samsung, $365 billion for SK Hynix, and $160 billion for Micron.
$MU $SKHY $DRAM Oh no. Terrible news. The three memory makers are only projected to earn almost $1 trillion in operating income by 2029. Devastating. pic.twitter.com/ykIV1o2fLi
— Trade Whisperer (@TradexWhisperer) July 29, 2026
Memory is notoriously cyclical, so we’d treat those figures as an aggressive scenario rather than a promise. Even a large haircut would leave an extraordinary profit pool.
The Reflexive Damage Crosses Borders
Samsung and SK Hynix dominate the Korean market and serve as global valuation anchors for memory stocks. When their shares collapse, investors don’t isolate the damage in Seoul. They mark down American competitors such as Micron and Sandisk, reduce exposure to semiconductor funds, and sell other AI infrastructure names caught in the same momentum and factor trades.
That pressure can then migrate into more speculative American technology companies. Lower public valuations raise the cost of equity, weaken acquisition currencies, narrow the initial public offering window, and make financing committees more cautious. Private-market marks update slowly; financing decisions don’t.
The final links in that chain are harder to measure in real time, but the direction is clear. Lower public comparables make it harder for venture-backed companies to raise money on attractive terms. That matters most in capital-intensive fields such as semiconductors, robotics, power systems, and advanced manufacturing, where promising companies can’t simply code their way around a funding drought.
We’ve argued that physical capacity increasingly matters more than flattering accounting aggregates. China is financing memory, robotics, energy, shipbuilding, and manufacturing capacity at tremendous scale. A leveraged market accident that weakens the capital base of America’s most important Asian technology allies works in China’s favor.
The selling crossed the Pacific before any production line changed.
Why Aren’t They Buying Their Own Shares?
Samsung and SK Hynix have strong balance sheets. Samsung ended the quarter with 190 trillion won in cash and 167.6 trillion won in net cash. That makes them different from Intel, which needed capital when the U.S. government acquired its stake.
Both Korean companies have returned capital to shareholders. Samsung announced a 10 trillion won repurchase program in 2024 and spent another 5.63 trillion won acquiring treasury stock during the second quarter. SK Hynix has discussed additional buybacks as its cash position has improved.
They could announce larger ones. A forcible repurchase would demonstrate management’s confidence in the earnings outlook, absorb forced selling, and exploit valuations that appear cheap relative to current profits.
The competing use of that cash is strategically important, though. Samsung and SK Hynix need to spend heavily on HBM, advanced packaging, leading-edge manufacturing, and research to preserve their lead over CXMT. SK Hynix’s recent Nasdaq listing also included newly issued shares, underscoring its preference for financing expansion rather than shrinking its equity base.
Keeping several generations of technological daylight between Korea’s memory champions and CXMT may create more long-term value than supporting this month’s share price. That tradeoff creates an opening for Washington.
The Intel Precedent, With a Difference
We defended the government’s Intel investment in a post at the time.
A non-controlling equity stake in a strategically important semiconductor company was industrial policy with taxpayer upside, not central planning.
Intel needed the money. Samsung and SK Hynix don’t.
President Trump should direct Treasury Secretary Scott Bessent and the Commerce Department to negotiate a capital-for-capacity agreement.
The United States would invest through newly issued preferred equity, warrants, or project-level securities. Samsung and SK Hynix would commit the proceeds to measurable increases in American HBM production, advanced packaging, research, and supplier capacity. Disbursements would occur as construction and production milestones were met.
There’s already a foundation. Samsung has planned more than $40 billion of investment in its Texas semiconductor complex. SK Hynix has a planned $3.87 billion HBM project in Indiana. A new agreement could accelerate those projects and expand their ambitions.
The securities should include dilution protection, a defined exit mechanism, clawbacks for missed commitments, and enough equity participation to reward taxpayers if the memory supercycle continues. South Korea could contribute matching incentives or guarantees, aligning both governments with the same capacity targets.
Buy Capacity, Demand Performance
The strategic asset here is the allied ability to manufacture the memory that makes advanced AI possible. This week’s collapse may eventually look like an extraordinary investment opportunity. A capital-for-capacity agreement would convert that dislocation into new production.
The deal would give each participant something valuable. Samsung and SK Hynix would receive help de-risking larger American investments. The United States would gain production, research, supply-chain resilience, and taxpayer equity upside. South Korea would strengthen its national champions without asking them to sacrifice the investment needed to stay ahead of China.
President Trump likes deals that combine industrial policy with a visible return. This one fits his model. Washington should buy capacity, demand performance, and participate in the upside when Korean memory stocks recover.