A collection of analysis on financial markets, investment strategy, and personal finance topics — combining rigorous data analysis with practical, actionable insight.
Specific-ticker calls, cap-table intelligence, M&A histories, sector positioning, and comparative screens. The market-facing side of the archive.
Coupang (CPNG) Q3 2026 estimate: company guide is 8–9% constant-currency growth and 300–400 bps of AEBITDA margin contraction on Chuseok timing plus the incident cohort. MajorHolders base $9.50B revenue, $105M AEBITDA. Bilingual (EN/KO).
Read the estimate →CPNG at $15 is 0.75x TTM sales and a ~0% FCF yield. The risk that reprices it is that the incident cohort never fully returns and mid-2027 margin recovery slips. Ranked factors for holders and new money. Bilingual (EN/KO).
Read the risk book →CPNG cash-flow history: FCF $1.76B (2023) to $1.01B (2024) to $522M (2025) to ~$100M TTM. Product Commerce still prints mid-single-digit AEBITDA. Developing Offerings and the $410M fine ate the year. Bilingual (EN/KO).
Read the cash-flow note →Standalone competitive landscape for Coupang: Naver owns intent, Coupang owns the last mile. Baemin still leads Eats. Taiwan and Rocket Now are the expensive second act. Temu is the new guest. Bilingual (EN/KO).
Read the landscape →CPNG cap-table note: SoftBank Vision Fund still 289.5M / 16.1%. Bom Kim owns all Class B and 74.3% of votes on ~8.8% of economics. Baillie Gifford cut 35% in Q2. Institutions own ~75% of shares out, ~84% of Class A. Bilingual (EN/KO).
Read the ownership note →Tesla entered SPX on 21 Dec 2020 at 1.69% after ten years public and four GAAP-profit quarters; indexers bought ~$80–94B. SpaceX is larger and already in NDX, but seasoning and earnings gates still block — calendar conversation opens 12 Jun 2027. Bilingual (EN/KO).
Read the comparison →Not one market order: APs pre-position 15–17 Sep and print the bulk in the 18 Sep closing auction. QQQ’s slice ~48M / $7.4B; as of 11 Sep it still held the old 1.26% weight. Companion to the NDX weight-math and float notes. Bilingual (EN/KO).
Read the plumbing note →Not $2T ÷ Nasdaq-100 — modified market cap = price × min(listed Class A, 3 × float). The 1.28% → 2.82% step is a 2.20× float-file refresh, checked against live QQQ at 1.26%. Companion to the NDX float & unlock note. Bilingual (EN/KO).
Read the index math →Nasdaq GIW lifts SpaceX NDX weight from 1.28% to ~2.82% — ~$26B / ~173M shares of mechanical demand into the 18 Sep rebalance, before the day-105 lock-up drip of up to 319M shares on 24–25 Sep. Bilingual (EN/KO).
Read the float note →At ~$2.05T / ~64× sales, $151 is not a rocket multiple — it is an AI-factory and Starlink compounding bet under heavy capex. Risks that reprice it: AI megawatts not converting to contracted dollars, customer concentration, Starlink ARPU, and lock-up supply. Holder vs buyer framing. Bilingual (EN/KO).
Read the risk book →At ~$4.7B / ~10× FY26 EBITDA / ~8% FCF yield, $22 assumes mid-20s volume and ~21% EBITDA hold. The risks that reprice it: volume slipping to mid-teens, buying growth back with marketing, and cash tax after the DTA release. Holder vs buyer framing. Bilingual (EN/KO).
Read the risk book →Company FCF from −$6M (2021) to $283M (2025); H1’26 already $203M. Adj. EBITDA margin −2% → 17% → ~21% FY26 guide. Operating leverage on a stable transaction margin — not a take-rate miracle. Bilingual (EN/KO).
Read the cash-flow note →Corridor-by-corridor map of remittance competition: Wise as the price ceiling, Western Union as the share pool, and where Remitly wins or loses on take-rate mix. Same-quarter peer tape for June 2026. Bilingual (EN/KO).
Read the landscape →Institutional ownership from 80.2% at YE25 to ~88.6% now as Naspers/Prosus fully exited. L12M Form 4 tape is 100% sales; ~91% of the dollar volume was the Naspers block. Fidelity, BlackRock, and the index complex absorbed the float. Bilingual (EN/KO).
Read the ownership note →Cash $676M vs debt ~$38M. First full-year GAAP profit in 2025; Adj. EBITDA and FCF stepped up through H1’26. New CEO Sebastian Gunningham. Base nine-year owner’s return mid-teens if volume outruns take-rate pressure. Companion to the multi-asset framework. Bilingual (EN/KO).
Read the report →Same horizon across Delta, Boeing, SpaceX, Tesla, Remitly, the S&P 500, TLT, and TLTW. Delta’s base (9–11%) still has the tightest cone over a valuation-aware index hurdle; Remitly’s base (12–16%) is higher but wider — take rate and corridor competition are the hinge; TLT is ~5% ballast; TLTW’s ~11% distribution is not its expected total return. Bilingual (EN/KO).
Read the framework →Independent Q3 build below Street: $10.6B revenue, operating income near zero, capex still ~$18B and almost all AI. Sold megawatts can pay back in under a year; the installed fleet still yields ~$6.3B per GW-year. Cursor stub from the Aug 14 close is in the number. Bilingual (EN/KO).
Read the estimate →July 53 and August 51 jets after June’s 64. Management framed FY FCF at $1–3B, with Q3 in the low hundreds of millions after a ~$700M DOJ payment — which means Q4 must carry most of the year. Transparent delivery-path revenue and Q4 FCF arithmetic. Bilingual (EN/KO).
Read the estimate →July 10 guide: mid-teens revenue, 11–13% operating margin, adjusted EPS $2.00–$2.50, all-in fuel ~$3.15/gal. Bear/base/bull off that box, with secondary tape risk on Gulf Coast jet and July reliability labeled as such. Bilingual (EN/KO).
Read the estimate →Q2 wasn't as bad as feared — a clean beat on revenue, EBITDA, and EPS. But the second guidance cut of the year and a withdrawn cash-flow forecast confirm the equity is now a balance-sheet story. The question: can a new CFO with a transaction résumé convert a 6–7x EV/EBITDA conglomerate into a deleveraged operating company by selling units at diagnostics-deal multiples and retiring the ~$210M annual interest bill? Bilingual (EN/KO).
Read the update →Udemy consolidation flatters headline growth (+60% Y/Y, ahead of consensus) while normalized organic revenue declined ~1%. The decisive development this quarter isn't operational: a $100M related-party investment in LearnVector — a pre-product startup founded and led by Coursera's own board chairman — at an implied ~$300M valuation. Process was followed; the substance still warrants a governance discount. Bilingual (EN/KO).
Read the update →After the February 2026 SpaceX–xAI merger (xAI valued ~$80B, combined entity ~$1.25T), SpaceX owns the Colossus data centers and has become a real AI-capex player. Hyperscalers guided to ~$725B for 2026. How does SpaceX’s AI wing compare on absolute spend, Colossus deployment speed (122 days for the first 100k-GPU cluster), capital efficiency, and PUE? Bilingual (EN/KO).
Read the briefing →Latest quarterly CapEx table (Amazon ~$53–55B in a single quarter, SpaceX AI ~$7.7B in Q1), Colossus 1 & 2 scale snapshots, construction timelines, facility cost per MW (~$2.7M vs industry ~$12.3M), and PUE estimates against Google/Meta (~1.09) and AWS/Microsoft. Data-focused companion to the Capex Race briefing. Bilingual (EN/KO).
Read the analysis →Most analysis stops at the dollars raised and the rockets flown. The deeper advantage is the talent that joined and stayed for the vision over two decades — with profiles of key engineers and a comparison to Tesla’s parallel talent retention system (mission + equity + intensity at larger scale).
Read the report →Connectivity segment Adjusted EBITDA margin reached ~63% in 2025 with operating margin near 39%. Scale, vertical integration, and network efficiency drove the expansion even as ARPU declined with international growth. Trajectory, drivers, headwinds, and outlook. Bilingual (EN/KO).
Read the analysis →Each V3 delivers ~10× downlink and ~22× uplink versus V2. One Starship launch can add ~60 Tbps. Manufacturing cost estimates (~$1.0–1.2M), lower cost-per-Gbps, Starship synergy, latency benefits, and competitive implications. Bilingual (EN/KO).
Read the analysis →Claude’s “aha moment” came from deep Excel/PowerPoint-style office integrations. The next structural shift is physical AI. A thesis on how Grok (xAI) + Tesla Optimus + Starlink connectivity may create a durable data and deployment moat in 3–5 years. Risks included. Bilingual (EN/KO).
Read the thesis →Why a large post-IPO unlock does not automatically imply a large selloff — the interaction between the first public earnings, an unusually high short position (~32% of float), price-sensitive insider supply below the $135 IPO price, and deep retail demand. Three early-August scenarios. For discussion and research only. Bilingual (EN/KO).
Read the setup →Why a few hundred million on a Norwegian AI startup can matter across a multi-billion-dollar service base — the margin math, the headcount-decoupling question, and what could go wrong. AI-generated, reference only. Bilingual (EN/KO).
Read the thesis →Endra, Augmenta, Autodesk, Bentley and rivals' in-house AI — mapped by threat level, with the one risk that really matters: commoditization. AI-generated, reference only. Bilingual (EN/KO).
See the landscape →A customer buys its own AI supplier: the ~$390M / NOK 4.0B deal, a reported ~67% stake, the founder joining as Head of AI Engineering, and the competitive read-through. AI-generated, reference only. Bilingual (EN/KO).
Read the note →SpaceX and Tesla are worth about the same (~$1.5T each), yet Musk owns very different slices — ~42% of SpaceX vs ~20% of Tesla. The merger ratio decides where his combined stake lands between those two numbers. An analysis of how the deal structure reshapes his ownership of the combined company. Bilingual (EN/KO).
Read the analysis →The acquisition label matters less than the share classes. A SpaceX-led transaction can leave Musk with roughly one-third of the economics yet potentially majority voting control — if high-vote founder shares survive and Tesla holders receive ordinary-vote shares. An interactive voting-power simulator.
Open the simulator →Move the deal-value split and watch Musk's blended ownership of the combined company. He owns ~42% of SpaceX but ~20% of Tesla, so the ratio decides where his stake lands between those two. The companion economic-ownership tool to the voting-power simulator.
Open the simulator →SpaceX arrived as the largest IPO in history — yet fewer than 5% of its shares actually trade. That scarcity drove the early spike; now a wave of lock-up expirations flips the equation. A look at the supply-demand structure, the lock-up schedule, and the Korean ETF exposure routing retail demand into a very thin float.
Read the analysis →Can structural demand from ETFs, index funds, active managers and retail investors exceed the shares actually offered for sale? An interactive model of the SPCX float: adjust the inputs and watch tradable supply meet structural demand. The companion tool to the float & ETF analysis.
Open the simulation →Seven developments across franchise deals and distress: Jersey Mike's IPO filing putting a premium on documented ~42% franchisee returns, Pizza Hut's $2.7B breakup sale, Papa Johns' proposed take-private with a major franchisee, Inspire Brands' balance-sheet IPO, and two franchisee bankruptcies (Hardee's, Popeyes) showing that sales scale cannot offset negative unit economics. Value the restaurant cash flow first, the brand second.
Read the brief →Pizza Hut's $2.7B sale as a benchmark for challenged franchise systems, Wendy's take-private interest, Red Robin's refranchising math, Texas Roadhouse's premium unit economics, a Hardee's franchisee bankruptcy, and MTY's closure program. Six developments, one theme: valuation discipline is shifting from brand recognition to unit-level cash conversion.
Read the brief →Following the June 23 supplemental modeling call, the investment case has shifted from revenue growth to margin expansion and capital return. Revenue guided down on a normalized basis (−4% to −2%), but synergies arriving ahead of the original 24-month plan (≥$80M by year-end), the share count already shrinking (>$70M repurchased in ~5 weeks), and a Q4’26 exit margin of ~16% pointing to a materially higher 2027 earnings base. A full model for 2026–2027 EBITDA and cash flow.
Read the model →Fixed wireless poached millions of cable broadband customers by selling spare 5G capacity as home internet. Comcast's counter inverts the play: sell mobile service carried mostly on the cable network it already owns — because roughly 90% of “mobile” data is consumed at or near a building cable already serves. The three-layer network build, the Q1 2026 scoreboard, and what the strategy can and cannot do.
Read the analysis →SpaceX's constellation ended 2025 near 450 Tbps of total capacity. With Starship-launched V3 satellites adding ~60 Tbps per flight, a decade-long trajectory model asks the harder question: does “total Tbps” actually mean anything when comparing satellite to terrestrial? What the capacity curve through 2035 means for Comcast, Charter, AT&T, Verizon and T-Mobile.
Read the analysis →Two decades of dealmaking show a consistent pattern: Comcast creates value buying from distressed sellers and destroys it bidding at auction. A full deal ledger from Adelphia to Time Warner Cable to Sky, the story of how Universal ended up inside NBCU, a scorecard on every major bid, and valuation scenarios for NBCUniversal at the mid-2027 spin.
Read the note →Institutions own roughly 84% of Comcast's equity, yet no shareholder holds a majority — and one man's non-tradable Class B stock controls a third of every vote cast. A four-year review of majority owners and insider activity, with the governance implications for the coming NBCUniversal/Sky spin-off.
Read the analysis →Coursera, Boeing, Comcast, QuidelOrtho, Delta Air Lines and Credo Technology, each run through the same disciplined 6-gate value-investing checklist as of June 2026 and as if the report had been written in June 2022. Now updated with Comcast: two spin-offs in one year — Versant in January, NBCUniversal/Sky announced late June. Across four years, not one of these six was ever a clean PASS.
Open the screen →From grounded to climbing: a record $695 billion backlog, a stabilizing factory floor, and the long runway back to positive free cash flow. Seven years of altitude plotted, three pillars of the recovery case, the 2018–2026 crisis timeline, and what could go wrong — and right — from here.
Read the report →A comprehensive analysis of the 777X program's regulatory bottlenecks, mounting write-down risk, and Boeing's narrowbody-led recovery thesis heading into 2H 2026. A reading of where the FAA queue actually sits, what the accumulated program losses imply for further charges, and how the rest of the portfolio is offsetting the widebody overhang.
Read the report →Insight Partners held roughly 29.6M COUR shares at merger close on May 11. In under three weeks they've crossed below 10% — the Section 16 reporting threshold — selling into a declining stock and leaving a structural overhang. A reading of the supply pressure, the company's $500M buyback authorization that partially offsets it, and what each side of that math is actually doing to the price.
Read the analysis →Since-inception, the covered-call ETF has lost 43% of its price. Since September 2023 — once the 2022 rate spike is behind you — total returns are positive 9.8%. The same fund, the same monthly data, three different stories depending on which window you read. A reading of when the strategy works, when it doesn't, and what an investor entering today should actually expect.
Read the report →The all-stock merger closed May 11, 2026 at a 0.800× exchange ratio, creating a ~$2.5B online-learning entity. A synthesis of 13G/13D filings, Form 3/4 disclosures, the S-4 proxy, and the May 18 Insight Partners sale — mapping who owns the combined company and what the $500M buyback authorization means for the float.
Open the report →Ninety-four 13F filers hold the equivalent of 120% of QDEL's shares outstanding — a 13.5M-share over-reporting gap that maps almost exactly onto the short interest. The May 2026 snapshot, the securities-lending economics that explain the gap, and the full holder-by-holder cap table.
Open the tracker →A video walk-through of the AI Supercycle research — five hyperscalers, $775 billion in 2026 capex, a memory oligopoly rationing supply, and the sustainability question. The companion analysis to the written piece, in moving form.
Watch on YouTube ↗Five hyperscalers will spend roughly $775 billion in 2026 — nearly doubling 2025's record. A memory-chip oligopoly is rationing supply. End-user ROI lags badly behind capex. Reading whether this is a demand-led boom or a supply-driven mirage — and the three constraints money cannot solve.
Read the research note →Five fiscal years of guidance against what actually happened. From the COVID sugar high through the CEO transition to the May 5 reset — every guide issued, every mid-year revision, every miss. The pattern that emerges is the story.
Open the tracker →AI coding assistants have collapsed the technical floor for building software. The arithmetic of a senior operator's time has not. A viability assessment of the build-it-yourself accounting stack — for personal use, a vertical-SaaS venture, or neither.
Read the assessment →The proxy is filed; the Schedule 13Gs followed. T. Rowe halved its position, Fidelity bulked up to 11.5%, Newtyn accumulated through the drawdown. A reading of QDEL's cap table across the past three quarters, with cost basis estimates and live mark-to-market.
Open the dashboard →The fundamentals declined ~33% from peak. The stock declined ~95%. The gap is who owns it now — and what it would take to bring the long-only money back. An interactive look at QDEL's cap table, the reflexive trap, and the May 5 catalyst.
Open the dashboard →Greenfield, roll-up, hybrid — the capital, timeline, and probability of success across every viable way to assemble a QDEL-scope IVD franchise. The honest answer to "how do I build this" turns out to be "you almost certainly don't."
Open the dashboard →The instinctive comparison — QDEL vs. Roche — hides the real strategic picture. Reading QuidelOrtho's four franchises individually reveals one to lean into, one to defend, one to harvest, and one to selectively attack. Letter grades for each.
Open the dashboard →Analysis of the Institutional Investor Restriction Act and its impact on SFR, BTR, apartments, and REITs. Why the bill paradoxically raises rents — and which apartment REITs are best positioned to benefit.
Read the report →Frameworks and concepts that don’t depend on a market call — the reference material a reader may want before, during, or after reading the ticker-specific work.
If AI and robotics make labor close to unlimited, will capital ownership dominate — and will scarcity disappear or move into status, land, energy, trust and ownership? An analytical framework on the economics of automation, with the policy question shifting from how much to tax high earners to who owns the AI capital that produces national income.
Read the essay →How moderate money growth can support economic expansion, why it does not automatically create real wealth, and why total prosperity can rise even as wealth becomes more concentrated. A companion piece to the ownership thesis, working through the monetary mechanics underneath the concentration data.
Read the essay →The core framework behind MajorHolders. How ownership concentration shapes who benefits from asset appreciation, how prosperity ripples in weakening waves from the center of ownership, and the circular question of whether asset value drives ownership or ownership drives asset value.
Read the framework →Three famous benchmarks, three completely different rulebooks. What an index actually is, how each of the big three is built and rebalanced, and how to choose between cap-weighted, equal-weighted, and revenue-weighted approaches.
Read the note →A step-by-step guide to inflation, from the basic mechanics to why the headline CPI number understates what's really happening at the bottom of the income distribution. Six chapters on measurement, distributional impact, and the four ways official inflation systematically undercounts the cost-of-living squeeze on the poor.
Read the explainer →Roughly 2.7× annual economic output, owed across government, business, and households. A reading of the largest borrower on earth, the engines of business capital, and what the household debt picture actually looks like below the headline number.
Open the briefing →The S&P 500 is up 205% since 2020. Consumer sentiment hit 47.6 — the lowest reading on record. A synthesis of Fed stability data, CapEx shifts, and sector breadth, reading the widest disconnect between asset prices and lived reality in modern US history.
Read the briefing →Self-directed IRAs let investors hold real estate, precious metals, private equity, cryptocurrency, and more in a tax-advantaged account. A practical guide to how they work, custodian selection, and the rules that trip up most first-time users.
Read the guide →