All Posts

SK Hynix pledged a 40 trillion won buyback-and-cancel. Two days later Samsung pledged 110 trillion. The smaller number won, and here is why.

I counted 632 instances of price reaching overhead supply. It broke through 53.5% of the time, a coin flip. Why I am shutting down weekly per-stock planning.

I held a 10M-won SK Hynix position through the KOSPI's worst month on record and its biggest single-day gain. A day-by-day log from July 21 through the August 1 review.

ServiceNow beat every line and re-accelerated cRPO to 21.5%, yet the gap-up vanished in 16 minutes. Why a beat went nowhere, and the $130 target.

The won firmed to 1,481 and SK Hynix still crashed 38% from its peak. The five new engines behind it, and a fact-check of the scenarios from two weeks ago.

Korea's 2026 dual-listing ban fact-checked: the 3% rule, the board's five duties, the exemption trap, holdco re-rating, and three market scenarios.

SanDisk fell 11% on the day Bernstein raised its target to $3,000. Four reasons good news turned into a sell signal, plus the downside scenarios.

Foreigners dumped 150 trillion won on the KOSPI as the won hit 1,540. Why the most-held names crack first, the ADR myth, and the downside scenarios.

The M7, semis, the KOSPI and SK Hynix all fell together. This is not rotation, it is a cash grab. The mechanics of the deleveraging, and what survives.

HBM swallowed commodity DRAM and flipped memory pricing power for the first time in 30 years. The reversal, chipflation, and the 2028 to 2030 signals.

The Kospi doubled in H1 2026 as foreigners sold a record 100 trillion won. How retail margin debt absorbed it, the 56% chip concentration, and the outlook.

SK hynix lists a Nasdaq ADR as early as July. Via the TSMC mirror: how high the local shares can run, Korea vs US tax, the Kospi impact, and the real risks.

Low rates propped up valuations for 15 years and excess liquidity has now turned negative. How the shift from beta to alpha reshapes how you invest.

Post-earnings deep dive on Nebius, CoreWeave and IREN across EV/Sales, EV/ARR and SOTP. Targets: NBIS $245, CRWV $115, and IREN $62 in the base case.

NBIS, CRWV and IREN are up 70% to 642% in a year. Share price misleads; market cap over sales tells you cheap from expensive. Plus an allocation guide.

Intel +97% against AMD +28% after earnings, decomposed with P/S normalization, SOTP, and a cycle-peak scorecard. Targets: INTC $104, AMD $413.

Intel rose 97% and AMD 28% after earnings. A plain-English look at the two rivals, the 'AI inference needs CPUs too' story, and which one to buy.

Samsung, SK Hynix, Micron and SanDisk across revenue growth, P/S normalization and SOTP. Top pick SK Hynix, with Micron at a 16.7x peer premium.

The four memory makers rose 420% to 4,109% in a year. Share price misleads; market cap over revenue shows who is actually cheap, in plain English.

Cloudflare (NET) was thriving, then cut 1 in 5 employees and crashed 24%. What the company actually does, and whether it is really in trouble.

Cloudflare beat with revenue up 34% and still fell 24.3% in one session. We decompose the 1,100 layoffs, margin compression, and our $225 target.

Datadog beat revenue by 32%, lifted FY26 guidance from 19% to 26%, and re-accelerated non-AI customers. What it does, and whether to buy now.

A 32% revenue beat, FY26 guidance raised 7 points, and a re-accelerating non-AI cohort drove a 31% single-session surge. How we get to a $230 target.

A $1,000 stake in DigitalOcean last May is worth $5,130 today. AI revenue up 221%, FY27 guidance above 50%. What drove the 5x, and can you still buy?

DigitalOcean double-beat, AI ARR rose 221%, and FY27 guidance jumped from 30% to over 50%, sending shares up 40% in a session. Our $185 target.

SanDisk posted revenue up 251% with 78.4% gross margin and $42B locked into multi-year contracts. The NBM model, and how we reach an $1,800 target.

₩1M invested last May would be ₩33M now. The USB-stick maker rose 3,685% after its spin-off as AI giants pre-booked $42B of memory. What happened.

ServiceNow was perfect on revenue, profit and cash flow, but one line about a 'slight slowdown' triggered a 17.75% crash. Markets trade acceleration.

ServiceNow beat on revenue, EPS, cRPO and FCF and still fell 17.75%. Dissecting the multiple compression: three quarters of cRPO deceleration, and Armis.

Russian refinery strikes and the Hormuz blockade pushed diesel crack spreads to $65 a barrel. The winners by region, and why margins may stay high.

Korea's 2026 separate taxation on high dividends: the four rate tiers from 14% to 30%, who qualifies, how to apply, and the savings by income level.

Palantir scores near-perfect on fundamentals, but one critical weakness holds it back. Ten metrics that set a fair value for PLTR right now.

ServiceNow stock crashed 53% from highs, but 98% of customers never leave. We score 10 key fundamentals, compare NOW vs CRM vs WDAY, and estimate fair value.

Q4 revenue +70%, Rule of 40 = 127, 1,300+ AIP Bootcamps. Beyond geopolitical momentum — five structural reasons behind Palantir's rebound, backed by data.

The US-Iran war has closed the Strait of Hormuz. From WTI $75 and Brent $82, three scenarios including a $120 bull case, and the ETFs that benefit.

On March 4, 2026 the won passed 1,505 per dollar for the first time since 2009 and the KOSPI fell 12.06%, its worst day ever. Causes and responses.

A Korean soy sauce holdco that has surged and crashed five times in three years. The structural reason behind it, and a weekly-candle scaling strategy.

DigitalOcean climbed from $25 to $68 in under a year. The AI pivot, the financials, four key risks, and what to watch in the coming Q4 print.

Seventeen years of Doosan Enerbility data tested three ways: DCA, averaging down, and a green-candle filter, plus a calculator for your own multiple.

A style self-assessment, stop-loss rules and mental discipline for beginners, with portfolios by risk level drawn from more than ten years investing.

A trillion dollars gone from software stocks. Fact-checking the AI fear trade with seat compression data and dot-com comparisons, plus rebound picks.

Three real success stories plus 5%, 6%, and 7% yield simulations. 19% effective tax rate, $500-600K needed, and why most success stories include capital gains.

TQQQ held 6 months: 12%. Held 2 weeks: 436%. One strategy change creates a 35x return gap, proven with 19 years of QQQ, QLD, and TQQQ real data.

1x hold, 2x 10-week MA, 3x 5-week MA. Optimal strategy per leverage level from 5.8 years of weekly data, ranked by return, Sharpe ratio, and max drawdown.

Kelly criterion position sizing, three portfolio profiles from conservative to aggressive, and the seven triggers that mean liquidate immediately.

Friday close check, Monday execution. MA calculation, peak-based stop-loss, and real QLD/TQQQ trade records analyzed for a complete strategy execution guide.

Why does money vanish when the market goes nowhere? 10-year simulations prove the structural cost of daily rebalancing in bull, sideways, and bear markets.

The 40% deduction is not a refund. The 25M won combined limit trap, real savings by salary, and optimal investment amounts: every condition analyzed.

Reading the market after Korea's change of government. Housing supply expansion and commercial law reform aimed at dividend policy look like the drivers.

Wall Street analysts' words and trades diverge by 67% to 88%. A data-driven case for caution on NVIDIA and for avoiding Palantir at current prices.