Explaining Korean Stocks and the AI Memory Trade - Rekt Mando

Summary

• Rekt Mando’s South Korea call is a two-engine rerating: shareholder reforms attack the “Korea discount” just as Samsung and SK Hynix sit at the center of AI memory demand. He says the pair comprise more than half of the main index, Korea controls roughly 90-95% of memory, and leading names still trade near 4-5 times forward earnings. His shorthand: “right time, right place to be long.”
• The trade is not simply cheap memory stocks: Korea’s chaebol structure historically trapped earnings instead of returning them to minority shareholders. The new laws, as Mando describes them, effectively push dividends, buybacks, director accountability, and more independent boards. Previously, “the E is kind of a meme” because shareholders could not count on receiving the earnings.
• Energy dependence is the cleanest way to break the Korean thesis. South Korea imports much of its food, energy, and rare earths; a Strait of Hormuz disruption or another bottleneck like the Qatar-helium scare could hit it disproportionately. Mando nevertheless calls Korea “one of the best macro trades out there” and something investors might own for “a long, long time.”
• Memory’s order book is strong but not infinitely bankable: hard commitments extend roughly to 2028, while claims of being booked through 2030 include softer orders. Micron, Samsung, and SK Hynix are adding fabs, and prior semiconductor cycles ended when supply outran demand. “This time does feel slightly different,” Mando says, but SK Hynix remains the pure bet where “you better hope the demand continues.”
• Mando’s core trading edge is brutally selective trend-following: buy assets approaching or clearing all-time highs, then ignore attractive-looking bottoms. His journal forces an explicit upside, downside, and exit; a 20-30% ETH/BTC opportunity failed the filter because it lacked “juice.” “Catching the bottom is a trade you should just be allergic to.”
• His current macro trio is South Korea, uranium through URA, and Google, with supply elasticity separating the themes. Rare-earth supply can respond, memory sits in the middle, and uranium is far more constrained; URA may be emerging from a 20-year range. He believes Google, trading around 18 times forward earnings by his estimate, could become the world’s most valuable company.
• Crypto is becoming part of the same cross-asset hunt for reflexive upside, but the discussion distinguishes breakouts from bounces. Mando does not own Zcash, though it is the crypto asset he would most likely choose; below its roughly $700 all-time high, it remains “still a bounce.” Thread Guy’s blunt reaction to a possible Bitcoin move from roughly $80,000 to $90,000 was: “Who gives a fuck?”

Deep dive

1. Risk appetite is migrating toward the most reflexive assets

• Thread Guy’s opening read was that investors have moved progressively farther out the risk curve since early April: first larger assets, then high-beta equities, Korean stocks, microcaps, and potentially crypto. He named Monero, Bitcoin, Zcash, and Hyperliquid as the limited crypto choices, with a combined market cap he described as tiny.

• Thread Guy’s framing was that “it’s all the same people.” The traders now chasing tiny photonics or AI-related stocks are often the same people who chased coins two years earlier—“the same group of men who are looking for a 10x”—with little practical distinction between stock and crypto pump dynamics.

• Mando’s own crypto participation was modest. He owned HYPE, but after moving early it sat near $41 for roughly a month; he did not meaningfully capture Bitcoin’s move from $70,000 into the $80,000s. His macro longs, rather than crypto, generated the stronger results.

2. Korea’s transformation created the original long thesis

• Mando first became interested after visiting South Korea, while emphasizing that a few trips did not make him an authority. What impressed him was the reconstruction: Seoul had been repeatedly overrun and largely destroyed, yet nearly everything visible had been built within roughly 70 years.

• Thread Guy highlighted the country’s risk appetite and described it as having the world’s highest IQ and one of the best education systems. His deliberately simple framing was: “I want to be long the highest-IQ country in the world… long high IQ.”

• That experience distinguished Korea from other emerging-market trips. India, Brazil, and much of the BRICS left Mando seeing currency, political, judicial, corruption, and execution risks; Vietnam offered some similar industrial promise, but Korea felt unusually complete. Complex industries require an ecosystem—“it takes a village”—not merely one promising company.

3. Governance reform is attacking the Korea discount

• Korean equities historically traded at a discount, in Mando’s account, because chaebols owned one another through unusual structures, hoarded cash, supported affiliated companies, and rarely returned capital. Minority shareholders received few buybacks or dividends, so low price-to-earnings ratios were less compelling than they appeared: “The E is kind of a meme.”

• Thread Guy’s useful pushback was that many US companies also pay no dividend. Mando’s answer: US companies commonly return capital through buybacks; Korean companies often did neither, while directors legally answered to the company rather than owing shareholders the familiar fiduciary duty.

• The political change matters because new reforms, as Mando describes them, push companies toward dividends and buybacks, extend directors’ obligations toward shareholders, and make boards less captive to controlling chaebols. Japan’s earlier governance rerating under Shinzo Abe provided the relevant regional template.

• The timing is the thesis: after an election, shareholder reform arrived alongside the AI-memory boom, while Korean valuations remained near or below single-digit earnings multiples. Mando says the index has “basically tripled,” yet leading forward multiples still looked “ridiculous” near 4-5 times.

4. Samsung and SK Hynix express two different versions of the trade

• The Korea index, accessible through the EWY ETF, is now substantially a Samsung-and-SK-Hynix position; Mando puts their combined weight above 50%. He initially owned Samsung and the broad index because governance, not memory alone, drove his thesis.

• Samsung is his diversified expression: memory represents only about half its business, with consumer electronics and the broader chaebol rerating attached. He calls it Korea’s “Mag 7” analogue, combining governance reform with AI exposure.

• SK Hynix is the concentrated expression: “They only do memory.” That makes it the more direct memory-demand bet and leaves it more exposed to the supply-and-demand outcome.

5. AI memory may be structural, but the cycle has not been abolished

• Korea’s roughly 90-95% memory share, as Mando estimates it, sits inside a geographically dense Asian supply chain that is extremely difficult to reproduce. Numerous specialized companies must operate together, making near-term supply relatively inelastic.

• Hard memory orders appear to run through about 2028; bookings advertised through 2030 include softer commitments, sometimes from customers helping finance fabs in exchange for future priority. Mando therefore rejects “booked indefinitely” as proof of an effortless long.

• His warning comes from 20-30 years of semiconductor history: these trades repeatedly look permanent before capacity arrives and demand disappoints. “This time does feel slightly different,” but Samsung, SK Hynix, and Micron are all building fabs; cheap forward multiples can be misleading if supply rises and data-center demand disappoints three years out.

6. Korea’s external dependence is the thesis-breaking risk

• South Korea’s industrial sophistication coexists with acute reliance on imported food, energy, rare earths, and other inputs. Mando identifies a renewed Iran escalation or Strait of Hormuz closure as the principal risk because Korea cannot insulate itself the way a resource-rich economy might.

• His best specimen was helium: during an earlier scare, markets realized that South Korea needed helium and that the supply came from Qatar, exposing another hidden bottleneck. The AI supply chain may be locally clustered, but essential resources arrive from elsewhere.

• Thread Guy’s pushback—worth keeping—was that this made Korea the most frightening asset to buy at the war-driven low, precisely when it became the best trade. His resulting rule is to “fade every escalation and stay long” while Trump can repeatedly cool rhetoric and preserve markets; Mando broadly agreed that the war situation now felt likely to be okay.

7. US passive flows amplify the AI-capex narrative

• Thread Guy contrasted crypto with the US equity machine: Americans continuously allocate retirement savings into the S&P 500 and hold for decades, while incremental cash can create many times its amount in market capitalization. He likened major equities to “low float, high FDV” assets operating at enormous scale.

• Mando’s comparison showed how exceptional that structure is. The STOXX Europe 600 reached a high around 2000 and did not break out to a new all-time high until roughly 2021; UK, Japanese, and Korean equities experienced similarly long stagnations. US cultural conviction and international inflows make remaining underweight unusually painful.

• The emerging bull narrative is that AI gains accrue to corporations through workforce reductions, higher margins, and a small set of dominant winners, while companies spend aggressively because each believes it must build first. If everything else calms down, Mando expects investors to chase that narrative harder.

8. Breakouts, journals, and delayed exits define Mando’s process

• Mando’s prior career was high-yield and distressed credit, where he learned to analyze balance sheets. Thread Guy contrasted that bottom-oriented world—buying at 30 cents for recovery toward par—with crypto’s open-ended breakout payoff.

• Thread Guy said that experience had corrupted him into thinking about bottoms, while the breakout mindset emphasized capturing upside with a defined downside. The trade journal is Mando’s forcing mechanism: write the expected upside, downside, and cut level before entering. It rejected ETH/BTC because even being right offered perhaps 20-30%, while his preferred setup can plausibly produce a much larger multiple with a thematic tailwind and a defined stop.

• Entry is often just before an all-time-high break, when volatility is highest and the position needs more room; the danger is a double top. Exit requires losing both a four-hour trend and then a daily trend, so he accepts surrendering the top. In gold, he entered around $4,200-4,300, watched roughly $5,500, and exited near $5,050.

• Grid bots let him harvest the violent two-way movement surrounding breakouts—effectively acting as a futures liquidity provider while retaining the directional long. He has used Hyperliquid and Pionex, but preferred Binance liquidity; this volatility-overlay strategy has been his most successful approach for roughly the last year.

9. Buybacks separate compounding assets from dilution traps

• Mando sees buybacks versus issuance as the equity equivalent of crypto’s buyback-versus-unlock trade. Korea’s reform matters for exactly that reason, while Hyperliquid’s buybacks make its market-cap comparison with Robinhood especially provocative to Thread Guy.

• SaaS sits on the wrong side of the framework. Years of stock-based compensation created an issuance overhang that looked harmless while prices rose; when the trend broke, companies could keep printing shares to pay employees. Mando’s warning was categorical: “Do not buy the dip in SaaS.”

• Cheap-looking bottoms deceive because they feel low-risk. Snapchat could double supply after CZ published a thesis about it; Adobe compensation can keep diluting holders; a failed crypto bottom leaves the trader owning unwanted tokens and writing essays to defend them. The apparent floor is psychological, not structural.

10. The live watchlist contains three macro longs and two tactical exceptions

• Mando’s stated macro book is South Korea, uranium through URA, and Google; he sold gold after its trend failed. He ranks supply responsiveness with uranium most constrained, memory in the middle, and rare earths least scarce despite their name, making rare earths the more dangerous long-term position.

• URA interests him because uranium may be breaking from a roughly 20-year range. Google combines an all-time-high setup with a business thesis: at about 18 times forward earnings by his estimate, it could dominate the AI trade as companies benefit while jobs are lost and become “the most valuable company in the world.”

• Zcash is the crypto candidate, not yet the confirmed trade. He does not own it, and its roughly $700 all-time high remains the resistance separating a recovery from an all-time-high breakout: “It’s still bouncy. It’s still a bounce.”

• Mando endorsed Thread Guy’s shorter attention strategy as a separate skill: catching Hims or American Eagle after a viral catalyst can still capture the next 20-30% without pretending it is a long-term position. The edge is recognizing that repeated timeline exposure may feel late while still being “early enough”—a tactical trade, not an all-time-high thesis.