Insights

Low-cost Appalachian and Canadian supply, combined with below-normal regional consumption, has recently pushed prices at a major New England gas hub below Henry Hub.

The latest Short-Term Energy Outlook forecasts U.S. crude oil production averaging 13.8 million barrels per day in 2026, above the 13.7 million record set in 2025.

Agencies announced reduced regulatory burden for community banks and broader eligibility for the 18-month exam cycle.

A full review of three-year supply commitments, collection cycles and reverse DCF, distinguishing reported facts, management outlook and valuation assumptions.

Read the linked SK hynix study through authorization, purchases and cancellation. Operating research and cross-market premiums are different questions.

A guide to the revised BE study: separate revenue, earnings and cash collection. This is a research explanation, not fresh earnings news.

From memory demand to system integration, and from high earnings to capital allocation. Assess HBM competitiveness, cyclical valuation and buyback execution together rather than substituting a low P/E or high price target for analysis.

From time-to-power to orders, earnings and cash flow. The research case for SOFC extends beyond power shortages: can the company turn its technology into sustainable operating returns?

A September 9 company recap covers the September 8 forum on AI collaboration and memory direction—not a new order or earnings forecast.

The September 9 release describes up to 2 GW of buildout by 2027: planned capacity, not operational compute or a contract price.

NVIDIA’s September 10 blog outlines Raptor XPU integration. Track compatibility, delivery timing and commercial adoption separately.

September 10 announcement starts with NVIDIA’s own supply chain. Track deployment and measured outcomes rather than treating a partnership as revenue.

On 2026-09-08 Micron trailed the semiconductor ETF (SMH) by 2.80 points; on 09-09 it beat SMH by 2.65 points; on 09-10 intraday it trails again by 1.43 points. Three sessions, three sign flips, each large enough to be read as a regime change, yet the three-session sum is only -1.57 points. This piece tests the question against a pre-registered bucketed criterion and offers a way to separate intra-sector rotation from broad de-risking: SMH is falling 4.4x as fast as the S&P, while the spread between the best and worst semiconductor name is 4.81 points, which points to rotation rather than a broad exit. All data vintages are labeled; U.S. prices are live intraday.

At the 2026-09-08 U.S. close, the semiconductor ETF (SMH) finished +1.19% while Micron closed -1.61% and SanDisk -0.12%, leaving the two memory names 2.80 and 1.31 percentage points behind the index. That session admitted two mutually exclusive readings: a single-session mispricing, or a genuine leadership rotation into compute design and manufacturing with memory routed around. Because both explanations fit 09-08 equally well, a falsifiable test was pre-registered before the next open — observe Micron and SanDisk relative to SMH on 09-09, where both outperforming means mispricing, both underperforming means regime change, and one of each means flash and DRAM are still diverging and must not be judged together. Roughly one hour into the 2026-09-09 session (as-of 10:48 ET), Micron is +3.69% and SanDisk +3.09%, for excess returns of +3.41 and +2.81 points over SMH (+0.28%), while the Nasdaq 100 sits at -0.04% — a flat broad index indicating reallocation inside semiconductors rather than a rising tide, and in the opposite direction from 09-08 (Broadcom, 09-08's second-largest gainer, is -1.52% today). On a two-session compounded basis Micron is +2.020% against SMH's +1.470%, meaning the 09-08 shortfall has been erased. The article also records an irony in options positioning: at the 09-08 close Micron was the only name of nine with a put/call ratio above 1 (1.105) versus Broadcom's 0.197 — the insured name is today's biggest gainer. On the Korean leg, SK Hynix's +3.51% is recorded as unattributable, with foreign and institutional flow at just 3.31% of turnover. ⚠️ U.S. figures are live intraday and unsettled; the verdict requires the 16:00 ET close, and full falsification conditions are stated. No positions or trading recommendations.

At the September 8, 2026 close on the Korea Exchange, foreign investors flipped from net-buying 6 of the 8 Korean names we track to net-buying only 3 of 8. Close-weighted gross foreign buying fell from KRW 2.341tn to KRW 1.125tn (~USD 0.84bn), a 52.0% contraction. Yet concentration rose rather than fell: the three memory-chain names (Samsung Electronics, SK Hynix, Hanmi Semiconductor) absorbed 100.00% of gross foreign buying, up from 97.11% in the prior session, and all five net-sold names were non-semiconductors — the split ran along the semiconductor / non-semiconductor line with zero exceptions across all eight names. The novel finding here is not direction but efficiency. Samsung Electronics absorbed 5.739mn shares of combined foreign and institutional net buying, equal to 25.4% of the day's volume, and still closed down 0.19%. SK Hynix closed up 0.56% on a combined net-buy ratio of just 2.0%. Same day, same market, same sector — an order-of-magnitude gap in how much price each unit of institutional capital bought. Under the zero-sum identity of KRX investor-type accounting, the counterparty to Samsung's institutional bid was retail-and-other net selling of 5.739mn shares: the bid was absorbed in full. This piece gives the reproducible methodology, a confirmation from the slowest-moving variable (foreign ownership ratio), and falsification criteria that span the full range rather than only the endpoints. U.S. markets were still open at publication; all U.S. figures are intraday and are excluded from the conclusion.

At the September 7, 2026 close on the Korea Exchange, foreign investors flipped from net-selling 6 of the 8 Korean names we track to net-buying 6 of 8, and the close-weighted net buy rose from KRW 0.465tn (~USD 0.35bn) to KRW 2.298tn (~USD 1.71bn), a 4.94x increase. Counted by breadth, this looks like broad re-risking across Korea. Counted by money, the three memory-chain names (SK Hynix, Samsung Electronics, Hanmi Semiconductor) absorbed 97.11% of all gross foreign buying, with SK Hynix and Samsung Electronics alone at 96.23%, leaving 2.89% for the five non-memory names. Breadth and weight point in opposite directions. This piece separates the two with a reproducible method and gives full-range falsification criteria. U.S. markets were closed for Labor Day; the last settled U.S. session was September 4.

At 10:32 ET on 2026-09-04, the four U.S. memory names were up an average of 5.29% (SanDisk +7.45%, Seagate +5.04%, Western Digital +4.50%, Micron +4.17%) while the S&P 500 ETF was down 0.30% and the Nasdaq-100 ETF up just 0.18% — a 5.59 percentage-point spread between the group and the broad index. This is not a broad rally. The article sets out three reproducible tests for distinguishing a sector move from a market move: (1) whether the broad index and the sector carry opposite signs; (2) the ordering inside the sector (the NAND/HDD trio at +5.66% leads DRAM-heavy Micron at +4.17% by 1.49pp, pointing at the enterprise-storage leg rather than the HBM leg); (3) whether Asia moved first and whether the slow variable followed (SK Hynix closed 09-04 up 3.20% with foreign investors net buying 308,833 shares, 10.92% of session volume, and foreign ownership rising 50.44%→50.49%, the first increase after two consecutive declines). We also publish one piece of contrary evidence: in the 09-03 block tape SanDisk was the second-largest outflow on the entire tape (−$104.6M) while Micron took inflows — yet SanDisk is the strongest of the four the next session. Every ratio can be recomputed from the raw figures given in the text.

At the September 1, 2026 Korea Exchange close, foreign investors' combined net across Samsung and SK Hynix was just 29.19 billion won — appearing to show no activity. Split into legs, the sell side was 184.67B won and the buy side 155.48B won, for 340.14B won of gross two-way turnover: 11.65x the net figure. Foreign ownership ratios confirm directionally (Hynix 50.55%→50.57%, Samsung 46.72%→46.71%) with the zero-sum check passing. This article explains why net-flow metrics break down during intra-sector name rotation, why Korean flow ratios cannot be compared across sessions (KRX retroactively revises the volume denominator), and presents independent same-direction evidence from the settled 08-31 US cross-section, where the semiconductor ETF rose 0.64% while TSMC fell 0.53%.

At the US close on 2026-08-28, NVIDIA fell 4.57% and the semiconductor ETF (SMH) fell 3.47%, while the S&P 500 ETF (SPY) slipped just 0.23%. The hardest test for telling a systemic de-risking from a crowded-trade unwind is not the size of the drop but index option pricing: SPY at-the-money implied volatility was 10.74 while its semiconductor constituents ran 34–77. Index vol far below constituent vol has only one mathematical explanation — the constituents are not moving together. Includes two-directional memory decoupling evidence (near-identical two-day net moves, opposite paths), a Korean investor-type breakdown for the settled 08-31 session (two opposite structures under the same 14%), and a pre-registered falsification condition with an intraday checkpoint.

At the Korean close on 2026-08-28 SK Hynix fell 4.45%. Foreign investors net-sold just 6.80% of volume while domestic institutions net-sold 28.46% — 4.18x the foreign leg — with foreign ownership essentially unchanged. Reading the foreign leg alone inverts the conclusion. Samsung's foreign leg, by contrast, flipped from net buying to net selling. Includes settled and intraday US comparisons.

NVIDIA's FY27Q2 delivered record data center revenue of $89.0B on $96.2B total revenue (+106% YoY), beating consensus non-GAAP EPS by 6.14% — yet memory and storage stocks did not follow. The explanation is not a post-hoc narrative. Three structurally unrelated settled markets had already separated storage from NVIDIA before the print: US options at the 2026-08-26 close showed NVDA IV Rank maxed at 100.0 against SanDisk 3.2 and Micron 11.9 (one-year floors); Hong Kong central clearing on 2026-08-26 showed 07709 (2x long SK Hynix) down 18.625M shares while 07747 (2x long Samsung) gained 3.444M; and the Korea Exchange 2026-08-27 close showed foreign investors net selling SK Hynix 159,385 shares while net buying Samsung Electronics 1,381,786. The piece also unpacks a definitional trap: reading the foreign leg alone (30.6% to 7.79% to 5.05% of volume) says the de-risking is finished, while adding domestic institutions (7.71% to 9.49%) reverses the direction. Includes falsifiable verification nodes and a full settlement-date table.

SK Hynix (KRW 40.00tn) and Samsung Electronics (KRW 15.00tn) filed the same legal instrument on DART — a treasury share acquisition decision (자기주식취득결정). The only difference sits in Field 5, the stated acquisition purpose: SK Hynix's is cancellation; Samsung's is employee and executive share-based compensation (임직원 주식보상). Only the former permanently reduces share count; Samsung's shares return to float. Both filings fix the won amount, not the share count, so price moves rewrite the actual size: at the 2026-08-26 Korean close, SK Hynix's implied cancellation falls from the announced 3.295% to 3.244%, while Samsung's compensation pool rises from 53.29M to about 57.36M shares. Includes the 2026-08-26 investor-type zero-sum decomposition of the Korean close.

At the 2026-08-25 Korean close, foreign investors sold a net 1,234,309 shares of SK Hynix — 35.2% more than the prior session and 30.75% of volume, the stretch high — yet the stock rose 0.42%. The prior session, with lighter foreign selling (22.86%), it fell 3.41%. Volume was near-identical and combined foreign-plus-institutional selling barely changed (−1.087M vs −1.037M shares), yet price direction reversed. Using KRX investor-category data, this piece shows the line that moved with price was domestic institutions, not foreigners — and that both categories combined remained net sellers, meaning the actual absorber is an unbroken-out residual. Includes intraday US confirmation: storage stopped falling but has retraced the least (MU 29%, SNDK 9%, vs 47%-65% for semis).

On 2026-08-24 Korea's two memory giants fell together but 2.55x apart: Samsung Electronics −8.70%, SK Hynix −3.41%. The reflexive explanation is foreign outflows. KRX investor-type data for that settled session does not support it — foreign net selling as a share of volume was nearly identical (Samsung 21.21% vs SK Hynix 22.87%, slightly heavier on Hynix). The variable that actually differs by 4.45x is domestic institutions (19.32% vs 4.34%). This piece decomposes the seller mix from settled KRX data, uses primary DART filings to explain why Samsung specifically, and shows an under-computed mechanism: Hynix's cancellation fixes the won amount, not the share count, so its own decline pushed the effective cancellation ratio from 3.166% back up to 3.277%.

On 2026-08-19 and 08-21, SK Hynix and Samsung Electronics each filed multi-trillion-won share purchase disclosures with Korea's DART. They are widely lumped together as 'Korean memory returning cash to shareholders,' but the two filings are different legal instruments with different stated purposes. SK Hynix filed a Share Cancellation Decision: 24,070,000 shares, KRW 40.00 trillion, 3.295% of shares outstanding, permanently retired. Samsung filed a Treasury Share Acquisition Decision: 53,285,968 shares, KRW 15.00 trillion, 0.911% of shares outstanding, with the stated purpose of employee and executive share-based compensation, meaning those shares return to the float. This article compares both filings line by line from primary sources and adds a mechanic few have priced: both lock the KRW amount, not the share count, so SK Hynix's 4.09% rally above its reference price already shrinks the retirable share count by roughly 946,000 shares.

On 2026-08-19 after the Korean close, SK Hynix disclosed via DART the cancellation of 24,070,000 shares (KRW 40.0T, ~$28.7B) — a headline 3.295% of share base. But roughly 17.7M new ADR shares were issued over the same period, leaving a net reduction of only ~6.37M shares, or about 0.87% — one quarter of the headline. This piece breaks down the measurement gap, why Seoul and New York read the same filing on different clocks, and the storage-vs-benchmark split in the US session on 08-20.

Memory and semiconductor stocks sold off hard in August 2026 while the S&P 500 kept rising. This piece separates a single-cluster momentum unwind from broad de-risking using three verifiable structural tests: whether the decline gradient is monotonic, whether index implied volatility is being pushed up, and whether hedging demand sits on the index leg or single-name leg. Includes Korea Exchange investor-type data answering who is actually selling (foreign investors flipped to net sellers but on contracting volume, with retail absorbing over 30% of turnover), a volatility-scaled metric for judging whether a drop is genuinely abnormal, and four pre-committed falsification conditions.

On 2026-08-17 US memory stocks surged alone (SanDisk +8.88%, Micron +4.13%) and the consensus read was that pricing power sat in the US while Asia failed to follow. Within 24 hours that inverted: Seoul marked Samsung down 2.19% on 08-18 with full knowledge of the US move, Hong Kong's 2x leveraged wrappers liquidated 9.6%–10.5%, and the US session then handed back 79.8% (SanDisk) to 136.5% (Micron) of the prior day's gain — Micron is now below where the rally began. As of 2026-08-18 22:30 CST / 10:30 ET (first hour, not settled) the cross-section is monotonic: the closer to memory, the worse (SanDisk −6.51%, Seagate −6.38%, Western Digital −6.20%, Micron −5.41%), while the S&P ETF is down just 0.51% and Nvidia 2.14% — the signature of a single crowded cell unwinding rather than market-level deleveraging. Korean data shows foreigners did not exit (ownership rose to 46.81%/51.13%); what fell was bid intensity (Samsung at 11.8% of the prior session). Includes four pre-committed falsification conditions and a full data-gap list.

As of 2026-08-17 22:30 CST / 10:30 ET — the first hour of the US session, not settled — SanDisk is +8.34%, Western Digital +7.22% and Micron +4.86%, while the semiconductor ETF is up only 1.37%, the Nasdaq ETF 0.18%, and the S&P ETF is down 0.12%. It looks like a repeat of Friday's flat-index/split-internals session, but the arithmetic says otherwise: the between-group gap widened from 4.96pp to 6.17pp while the within-AI-compute spread collapsed from 12.44pp to 1.46pp, lifting dispersion's signal-to-noise from 0.399 to 4.226 — roughly 10.6x. Headline dispersion actually shrank, yet for the first time it points somewhere. Also included: the 08-14 options pricing layer (index put/call 3.92 and 2.093 against all six single names below 1, index implied vol at 0.64x the single-name mean, implied below realized in eight of nine names), an exclusive CCASS breakdown showing Hong Kong's 2x leveraged channel added 10.03M shares for SK Hynix while the Samsung equivalent was pure churn, three fixed falsification conditions, and full scope limitations.

As of 2026-08-14 10:31 ET (US intraday, unsettled), all four US storage names are up, +2.92% equal-weighted, while at the same instant SOXX is −0.49%, SMH −0.55%, Nvidia −0.10% and QQQ −0.15% — semiconductor indices are falling in the same minute storage is rising. Combined with Taiwan's TAIEX closing −0.46% on 08-14, Korea's KOSPI +2.42% versus KOSDAQ +0.38% (a 9.2pp three-day gap), and 08-13 block flows rotating out of optics/AI infrastructure into storage hardware, four independent markets point to one conclusion: this is not an industry-wide chip re-rating but a flow concentrated in the storage pocket. Full reproducible data, methodology warnings and four falsification conditions included.

As of 2026-08-13 10:38 ET (US intraday, unsettled), the four US-listed storage names are up +4.95% equal-weighted — 3.67x QQQ — while Nvidia is only +1.06%. At the same time, options data as of the 2026-08-12 close shows SMH and TSMC implied volatility at the 0th percentile of their own one-year range, Micron at 1.4, with VRP of −29.6% for Micron and −48.4% for SanDisk. Realized volatility sits at a six-year 99th percentile while the price paid for future volatility is at a one-year low. This piece breaks down the nature of that gap, its direction of spread (laterally across the market, not upstream into equipment), the split in Korean foreign flows between Samsung and SK Hynix, and the one scheduled settlement point: Nvidia earnings on August 20.

On August 12, 2026, the four US storage names (Micron, SanDisk, Western Digital, Seagate) rose an equal-weighted average of 7.06% intraday — 7.4x the Nasdaq 100 ETF and 23.5x SPY. But this was not an AI rally: Broadcom lagged the whole group at +1.59%, and Nvidia's +3.12% merely matched the SOXX sector ETF rather than leading it. The leadership sat in storage and memory-capex equipment. This analysis breaks down the move using a tier-by-tier performance ladder, Korean foreign-investor net buying as a share of traded volume (21.4% in Samsung Electronics, 13.3% in SK Hynix after four straight sessions of net selling), and an ADR-versus-local cross-check on SK Hynix that agreed to within 0.07 percentage points. It argues Korea's broad rally and the narrow US storage move may reflect two different mechanisms rather than one story. Both circulating catalysts — a reported Temasek initiation in Korean equities and a 15–23% server DDR5 contract price increase — lack primary sourcing and are labeled as such, alongside a contradicting Counterpoint datapoint on YMTC NAND share. Explicit falsification conditions are stated. US figures are intraday and unsettled as of 10:31 ET; Korean figures are settled August 12 closes.

As of the August 10, 2026 close, at-the-money implied volatility for Micron, SanDisk, Western Digital and Seagate all sat below their own 20-day realized volatility — gaps of −29.5, −47.6, −20.3 and −3.4 vol points, the most negative readings across 95 available observations since 2026-03-30, with Seagate printing negative for the first time on record. Implied volatility collapsed (Micron 139.7 → 80.0) while realized barely moved. This piece also tests and rejects the expiry-roll explanation: ATM implied vol shows no term-structure bias across 3–9 days, but the put/call ratio depends strongly on days-to-expiry, so put/call divergence narratives spanning a roll date do not hold.

On August 10, 2026, the memory sector split in two intraday: SanDisk (+2.31%) and Western Digital (+1.61%), which both reported after the close on August 5, are rallying, while Micron (-0.81%) and Seagate (-0.38%) — neither with a pending event — are not participating, even as semiconductor ETFs trade lower (SMH -0.78%). That is the mirror image of last week's ordering, when SanDisk and Western Digital were the hardest hit. Yet measured against each stock's own 20-day realized volatility, no move today exceeds 0.6 of a daily standard deviation, so the information is in the sign pattern rather than the magnitude: the dividing line is the earnings window, not DRAM versus NAND. This piece builds a volatility baseline before attributing anything, documents the dispersion evidence of violent single stocks against a calm index (triple-digit annualized realized vol in memory versus 14% for SPY), flags the divergence between realized volatility at one-year highs and single-digit option IV Ranks, reports the Korean close with foreign, institutional and derived retail flows, and fixes five falsification conditions in advance.

On August 7, 2026, every major US storage name opened higher and was sold within the first hour: Seagate gapped +1.14% and traded to -9.74%, SanDisk gapped +3.94% and fell to -4.42%, Western Digital -5.47%, Micron -2.85% - while SMH held +1.04%, NVDA +1.83% and SPY +0.50%. That combination, weak single names inside a firm index, is the signature of a momentum unwind in one cluster rather than broad de-risking. Seoul had closed six hours earlier saying the opposite: SK Hynix's decline halved to -4.88%, Samsung turned positive, and foreign net selling on the KOSPI collapsed 74%. We break down the first-hour tape, the Seagate tell, the lookback-window trap behind the 'profit-taking' narrative, and what would flip this read.

On August 6, 2026, SK Hynix closed −10.37%, Samsung −6.30% and the KOSPI −4.58%. US markets took the handoff at the open — Micron opened −5.61% and traded to −7.41% — then reversed green within the first hour. As of 10:33 a.m. ET (intraday, session in progress), the only names still down more than 1% were Western Digital and SanDisk, the two companies that had just reported earnings. Western Digital beat its own guidance on revenue, gross margin and EPS, and guided fiscal Q1 2027 margins higher to 55–56%, yet was the worst performer at −11.35%, while Seagate — same hard-drive end market, no earnings this week — traded +1.65%. The article also corrects a widely repeated explanation: the 'parabolic run, now profit-taking' story holds only at a one-week lookback. Extend to one month and SanDisk is −16.5% and Micron −4.8% against an S&P 500 up 3.0%. Full data provenance and limitations included.

A session-by-session breakdown of foreign, institutional and retail net flows in Samsung Electronics (005930) and SK Hynix (000660), based on KRX investor-type data through the 2026-08-05 settled close. Key finding: the claim 'foreigners are fleeing Korean memory stocks' and its equally popular opposite, 'foreigners are accumulating the dip,' rest on the same dataset and differ only in the start date. Shift the window by one session and Samsung's foreign net flips from +7,611,626 shares to −747,385 shares. The cause: 07-31 alone accounts for 110% of the four-session net (206% for Hynix). The accurate answer is a third one — foreigners traded with the price direction (same-sign in 7 of 8 sessions), while domestic institutions were the persistent seller (−11.04M shares over three sessions, 14.8x the foreign sell) and retail investors absorbed the supply. All figures pass a zero-sum check across investor classes.

Over the three sessions from the 2026-07-31 close, companies doing the same thing — making memory chips — diverged by 23.36 percentage points: SanDisk +14.79%, Micron +7.59%, Western Digital +0.12%, Seagate −0.37%, SK Hynix −8.21%, Samsung Electronics −8.57%. This piece separates two effects. First, 8.57pp is pure time-zone artifact: Seoul's 8/04 close prints at 01:30 ET while the U.S. readings are 10:30 ET, with an entire U.S. session in between; net of the clock, the true spread is 14.79pp. Second, that remaining 14.79pp is structural — the same ordering repeated in two opposite directions (the 8/03 industry-wide selloff and the 8/04 industry-wide rebound) and lines up near-monotonically with HBM exposure, meaning the market is repricing HBM content per system rather than memory demand. Also covered: how Seoul's V-shaped session with ~8pp of range closing flat falsified our own morning call within the day; options positioning running exactly opposite to the price ranking (strongest name SanDisk buying protection at P/C 1.74 while laggard Micron sits at P/C 0.294); and why the semiconductor excess return is sector rotation rather than new liquidity. All baseline closes independently re-verified via yfinance K-lines and pykrx, with full replication steps and methodological limits.

On 2026-08-03 Korea's two memory majors dropped hard: SK Hynix closed at KRW 1,567,000 (-8.79%) and Samsung Electronics at KRW 239,500 (-8.76%), both underperforming the KOSPI (-5.35%). Foreign net flow as a share of each stock's traded volume flipped from +21.3%/+14.4% on 07/31 to -20.4%/-14.2%, which reads like a textbook foreign reversal. This piece runs three fully recomputable corrections. First, the apparent symmetry is an artifact of collapsing volume: Hynix turnover fell from 10.35m to 5.43m shares (-47.5%) and Samsung from 57.88m to 27.39m (-52.7%), so in share terms foreigners sold back only 50.2% and 46.6%, leaving both names net-bought across the two sessions (Hynix approx +1.10m shares, Samsung +4.463m). Second, the real liquidation came from domestic institutions: on Samsung they sold back 139.3% of Friday's buying, a two-day net of -1.421m shares, pointing the opposite way from foreigners. Third, foreigners sold the semiconductor leg, not Korea, buying Hyundai Motor (+11.0%), Kia (+11.6%) and Samsung SDI (+5.9%) the same session. A two-step percentage-versus-shares checklist and full method limits are included.

On 2026-07-31 three Korean memory-complex names closed near the +30% daily limit: SK Hynix at KRW 1,718,000 (+29.95%), Hanmi Semiconductor at KRW 214,000 (+27.98%), and Samsung Electronics at KRW 262,500 (+26.81%). But non-memory Korean large caps rose a median +9.53% the same session. This piece runs two fully recomputable corrections. First, strip market beta: roughly one third of the headline (31.8%-35.5%) was broad-market, leaving a true memory-specific excess of +16.84% multiplicative, not +30%. Second, fix the calendar: an 'extreme U.S.-Korea divergence of -11.52pp at the 99th percentile' was produced by subtracting Korea's 7/31 close from the U.S. 7/30 close, one trading day apart. Realigned to a two-day window, the spread collapses to 4.67pp and the dislocation vanishes. Includes a reusable two-step checklist and full disclosure of the method's limits.

On 2026-07-30 Samsung Electronics' board did two things on the same day: it revised Q2 2026 results upward to KRW 171.50tn revenue (+130.00% YoY) and KRW 89.49tn operating profit (+1,813.83% YoY), and it declared only a routine quarterly dividend of KRW 374 per share with zero buyback. This piece re-verifies every figure against first-party DART filings from Korea's Financial Supervisory Service and introduces a reusable analytical method — the 'clean sample test'. SK Hynix's refusal to detail a buyback has been widely excused as an ADR-related regulatory gag; Samsung faces no such constraint, upgraded its earnings that very day, and still returned essentially nothing. That makes the regulatory explanation non-load-bearing. It also unpacks a commonly miscalculated trap: Kioxia's KRW 60tn is not KRW 60tn of buyback ammunition.

On 2026-07-29 SK Hynix reported Q2 2026 results: operating profit KRW 60.54tn, +557% YoY, a 76% operating margin — all records. The stock closed down 9.61% the same day, after reversing from +4.45% to -9.61% intraday (24.1% range). This piece explains the paradox with four reproducible calculations: the miss was entirely in revenue not margin; net profit hides a KRW 62.17tn one-off gain (P/E moves from 3.7x to 6.6x); the buyback precondition can now be back-solved from the cash statement; and why a 2x leveraged ETF still cannot break even even after the underlying fully recovers. Includes full definitions, workings, and falsification conditions.

On 2026-07-28 global memory-chip stocks crashed together, with KOSPI down 10.84% in a single session, its 5th worst in 30 years. Yet the Dow closed higher, VIX did not spike, and the Korean won was flat. This article lays out a five-check framework for separating a genuine systemic crisis from sector-level deleveraging, with the full loss-gradient dataset and the explicit conditions that would invalidate the conclusion.

The last week of July 2026 brings a triple catalyst convergence: Goldman raises FOMC rate hike probability to 35%, SK Hynix Q2 earnings arrive July 29, and Big Tech reports en masse. Options markets are flashing extreme signals — LRCX IV Rank 99.5, AMAT 87.2, AMD 80.6. We analyze the fund flow divergence (TSM +$66.8M vs MU -$152M), institutional de-leveraging in CCASS, and three FOMC scenarios for the memory sector.

Intel Q2 2026 revenue hit $16.1B (+25% YoY), with Data Center/AI at $6.3B beating consensus by 14% and Foundry's 18A entering risk production. We analyze three signals: semiconductor recovery broadening from memory to logic, Intel Foundry's catch-up trajectory, and the hidden margin threat from oil prices breaking $100.

A use-case-driven roundup of quant tools for individual investors, organized into 4 categories: broker terminals for A-share live trading, backtesting platforms, overseas-market tools, and open-source frameworks. With pros, cons, target users, and the key compliance line: only official broker terminals are legal for automated A-share trading.

Monte Carlo simulation reveals: when underlying annualized volatility exceeds 70%, a 2x leveraged ETF loses money even if the stock rises 40% per year. We quantify the exact vol threshold where leverage turns from amplifier to destroyer.

SK海力士7月22日发布2026年Q2财报。尽管DRAM现货价格连涨8周，存储股却从6月高点回落30%。我们拆解三大关键信号：HBM利润率走势、回购计划扩展、Q3指引——叠加韩国杠杆ETF资产从24万亿韩元缩水至17万亿的结构性去杠杆背景。

SOXX两日跌超5%，台积电/ASML财报兑现抛压叠加7月OpEx。但期权市场罕见逆向信号密集爆发——AMD PCR达4.0σ极端值、KLAC PCR触5.25历史100%胜率。解读聪明钱与恐慌的博弈。

用29只AI产业链标的的期权Put Skew数据，拆解存储芯片暴跌背后的真实恐慌分布。数据显示设备/代工恐慌反降，存储被单独惩罚。

台积电将于7月16日公布2026年Q2财报。存储股从6月高点回调30%，SK海力士ADR刚登陆纳斯达克之际，我们解读5个趋同信号，判断存储半导体周期是否正在形成拐点。

Korea's margin financing peaked at 38-39 trillion KRW in June 2026 — an all-time high concentrated in Samsung and SK Hynix. With forced liquidations peaking July 15-16 and retail purchasing power down 20%, we break down the deleveraging timeline, NPS backstop mechanics, and what it means for global semiconductor positioning.

2026年7月13日，SK海力士在韩国交易所暴跌15.4%，三星电子跌10.7%——这是2022年以来最惨烈的存储单日抛售。外资抛售海力士70.5万股，机构卖出78.2万股，买方完全缺席。但三星两周前刚公布创纪录的89.4万亿韩元营业利润。我们分析了区分恐慌性抛售与基本面恶化的六大信号，以及台积电7月16日法说会对板块走向的决定性意义。

2026年7月10日，SK海力士在纳斯达克挂牌交易，募资265亿美元，超越阿里巴巴2014年纪录，成为美国史上最大外国公司IPO。但首日恰逢三星资本开支恐慌引发的存储板块暴跌。我们分析7倍超额认购、70亿美元基石投资者、三星创纪录的89.4万亿韩元Q2营业利润，以及这个上市时点对AI存储超级周期的机构信心意味着什么。

2026年7月9日存储芯片连跌5天后暴力反弹6-15%。但期权数据讲了一个分裂的故事——美光看跌看涨比0.57偏多，西部数据3.33极度看空。我们用资金流数据、伯恩斯坦DRAM定价预测和信号分歧，拆解这是死猫反弹还是真正反转。

2026年7月存储芯片连跌5日，三星创纪录Q2利润却不及预期触发"利好出尽"抛售。但大单数据讲了另一个故事——美光单日获+1.22亿美元机构净买入，而散户净卖出2.02亿美元。我们用数据拆解这场分歧背后的信号。

Samsung Electronics posted record Q2 preliminary revenue of ₩171 trillion and operating profit of ₩89.4 trillion, beating consensus. Yet SK Hynix fell -6%, memory ETFs dropped -8%, and the sector extended its selloff. Meanwhile, UBS ($1,625), Citi ($1,564), Bank of America ($1,550), and Mizuho ($1,375) all raised Micron price targets — implying +55% upside from $985. We analyze the buy-the-rumor-sell-the-news trap, Korean leverage mechanics, and what options data says about what comes next.

NVIDIA's Kyber NVL144 rack delay to 2028 triggered a second wave of selling in memory stocks on July 6, 2026. But options data tells a different story: MU's put-call ratio hit 2.12 (2.1σ extreme) — a historically reliable contrarian buy signal with 60% win rate and average +7.6% return. We separate 2028 dream discount from 2026 earnings reality.

Memory chip stocks fell another 5–14% on July 2, 2026 — but this time we can prove what caused it. Using real-time options flow, tick-level price data, and post-settlement trading, we test the Max Pain hypothesis: that mechanical forces from options expiration, negative gamma hedging, and leveraged ETF rebalancing drove 60–70% of the decline. The evidence is striking — MU pinned within $0.28 of its $950 put wall, then rebounded immediately after 4 PM settlement.

Memory chip stocks suffered a brutal selloff on July 1 2026 — SNXX -21%, MU -8.6%, SNDK -10.8% — while the S&P 500 barely moved. We analyze three competing explanations: DRAM antitrust litigation, price-peak fears, and the most overlooked catalyst — SK Hynix's ADR listing on July 10 triggering a massive capital rotation from Korean equities to US-listed shares.

On July 1, 2026, memory semiconductor composite signals flipped from +5 to -7 in a single session, with put/call ratios spiking across SNDK (2.06), WDC (1.79), STX (1.56), and MU (1.42). Yet institutional block trades showed MU net inflows of $208M and SNDK $146M. This article breaks down why bearish options activity and bullish equity flows can coexist — and what it signals for the storage cycle.

On June 29, 2026, South Korea unveiled a ₩2,000 trillion ($1.3 trillion) 10-year semiconductor mega-investment plan led by Samsung and SK Group. This analysis breaks down what it means for global memory chip supply-demand dynamics and related stocks.

Micron's fiscal Q3 2026 delivered revenue of $41.46B (+346% YoY), net income of $28.24B (+14x), and 84.9% gross margin — all company records. Investment banks have raised targets to $1,400–$1,650. We break down what these numbers mean for the broader memory stock rally and where the risks lie.

Micron's fiscal Q3 2026 revenue hit $41.5B (+346% YoY) with net income up 14x and gross margins at a record 84.9%. Investment banks raised targets to $1,400–$1,650. We break down the key numbers, sector rotation signals, and where the AI memory supercycle stands.

Micron's Q3 FY2026 earnings demolished expectations: $41.46B revenue (beat by 17.8%), $25.11 EPS (beat by 24.3%), and record 84.6% gross margins. Q4 guidance of ~$50B towers over the $39.6B consensus. We break down the key metrics, segment performance, HBM progress, and Wall Street reactions.

Line-by-line analysis of Micron CEO Sanjay Mehrotra's 45-minute earnings call. $100B SCA lock-in contracts, 86% gross margin beating TSMC and NVIDIA, HBM4 shipping >$1B, data center SSD doubling, and capex is construction not capacity expansion. Forward PE only 7.9x.

On June 23, 2026, Korean memory stocks crashed: SK Hynix -12.5%, Samsung -12.3%, leveraged 2x ETFs down over 20%, KOSPI hitting its third circuit breaker. This piece explains why it is structural deleveraging rather than fundamental deterioration, why HBM demand stays strong, and why Micron earnings are the watershed.

On June 23, 2026, Korean memory leaders fell 12% in a single session, triggering KOSPI's third circuit breaker of the year. We break down why this is structural deleveraging, not a fundamentals collapse, and why smart money in Micron quietly bought the dip.

Korea's memory chip exports surpassed $23 billion in the first 20 days of June 2026, tracking toward a record $38-42 billion for the full month. HBM shipments surged 51% MoM while DRAM ASPs doubled year-over-year. This analysis breaks down the drivers and supply-demand dynamics.

Micron Technology reports FY26 Q3 earnings on June 25, 2026. Options markets show extreme bearish positioning (PCR 2.21, 3.5σ), yet history suggests this is a contrarian bullish signal. We break down the key metrics, fund flows, and supply-demand dynamics heading into the report.

In June 2026, SK Hynix overtook Samsung Electronics for the first time to become South Korea's most valuable company, driven by structural demand for HBM (High Bandwidth Memory). This analysis breaks down the three drivers behind Hynix's rise, how it diverged from Samsung, and what it means for the memory sector's shift from cyclical to growth stock.