Starr Quant Lab Desk Research

US · Recap

US Market Recap | 2026-07-27 (ET) Monday

Mon US Recap · 10 tables America/New_York

Machine-translated from the Chinese original. In case of any discrepancy, the Chinese version prevails.

Single-name entries

This is a reconciliation report — it carries no single-name score table. Recaps are structured around hit-rate reconciliation, theme verification and next-day outlook.

Coverage window: 2026-07-27 09:30–16:00 ET regular session + 16:00–18:00 ET after hours Reconciled against: reports/us/2026-07-27.md (that morning's pre-market list) Data retrieved: 2026-07-27 approx. 22:05 UTC (= 18:05 ET, two hours after the close) Sources and basis: closing and after-hours prices for single stocks/ETFs come from the stockanalysis.com quote API, with timestamps standardized as Jul 27, 2026 4:00 PM EDT (close) and 6:00 PM EDT (after hours); VIX comes from the CBOE official delayed_quotes endpoint; the 10Y Treasury, Russell 2000 and crude oil come from yfinance; index levels come from media accounts and have been cross-checked against index-ETF percentage moves (see the verification note in §1). ⚠️ yfinance was broadly rate-limited this time, so at the single-stock level stockanalysis.com was used as the primary source, cross-checked item by item against news sources; anything that could not be cross-verified is marked as such.

0. Recap in one line

  1. Today was not risk-off, it was risk-rotation (rotation, not retreat): the Dow +0.51%, Russell 2000 +0.62%, financials +1.01% and consumer staples +1.46% all rose, but the Nasdaq was −0.18% and semiconductors (SMH) −2.25%. Risk appetite did not contract across the board — the single most crowded track, AI hardware, was pulled out on its own.
  2. The pre-market call that was right was oil, the one that was wrong was memory, and the one completely missed was semiconductor equipment. The oil chain delivered 100% (WTI −8.29%, JETS +2.89% beating SPY by 2.87pp, XLE −2.11%); whereas the memory chain, tagged "short-term long" pre-market, went with the long-term bear narrative right from the open — SNDK −11.02%, MU −2.25%.
  3. Today's genuinely strongest theme was a negative that did not exist pre-market: The Information reported that Chinese state-backed firms have begun mass-producing key chipmaking equipment, compounded by "circular financing" worries around Nvidia and a record one-day jump in its 5-year CDS — ASML −5.80%, NVDA −4.99%, AMD −5.17%, LRCX −4.46%.
  4. The pre-market list got 14 of 18 judgments directionally right ≈ 78%, but its usefulness was far below its hit rate: most of the names that delivered printed their high of the day at the open, giving back on average about half of the pre-market gain (UAL +3.5%→+1.94%, LUV +3.1%→+0.67%).
  5. Tone for the next session: wait-and-see dominates, but semiconductors are the only direction with follow-through momentum. VIX closed at 18.67 (+0.48%), after spiking to 19.93 intraday and pulling back — opened at 17.80, panicked intraday, repaired into the close — and that intraday path by itself refutes the pre-market read of "VIX falling = risk-on". The 10Y Treasury at 4.641% (−3.8bp), with stocks and bonds strengthening together, supports the "rotation, not retreat" characterization. The 7/28 FOMC meeting begins.

1. Market overview

Indicator 7/27 close Change Prior close (7/24)
Dow Jones Industrial Average 52,210.08 +0.51% (+262.8) 51,947.25
S&P 500 7,413.18 +0.02% (+1.20) 7,411.98
Nasdaq Composite 24,932.08 −0.18% (−43.74) 24,975.82
Russell 2000 2,948.04 +0.62% (+18.04) 2,930.00
VIX (CBOE official) 18.67 +0.48% (+0.09) 18.58
10Y Treasury yield 4.641% −3.8bp 4.679%
WTI crude $81.91 −8.29% $89.31
Brent crude $87.68 −9.40% $96.78
US Dollar Index no reliable data (CBOE endpoint refused, yfinance rate-limited) — —

✅ Cross-verification of the index levels (why this set of numbers is the one we trust): two media outlets gave Dow values for the "7/27 close" that differ by 263 points (52,210.08 vs 51,947.25). On checking, the Zacks piece reporting 51,947.25 is actually describing Friday 7/24's tape (that article also gives WTI $89.31, Brent $96.78 and VIX 18.58, all three matching the 7/24 closes from yfinance/CBOE exactly, item by item; and VIX 18.58 × 0.9602 = 17.84, which is precisely the 7/27 early-morning VIX recorded in the pre-market list). Taking 7/24 as the base × each index ETF's actual 7/27 percentage move fully reproduces all three values 52,210.08 / 7,413.18 / 24,932.08. This table therefore adopts the latter set and reassigns the Zacks set to 7/24.

Note: the same verification also resolves the crude-oil data conflict left over from the pre-market list (which listed three mutually exclusive values, $85.23 / $87.48 / $92.02, and would only commit to the direction). The true sequence is: 7/24 close WTI $89.31 / Brent $96.78 → 7/27 close WTI $81.91 / Brent $87.68. The pre-market caution of "confirm the direction only, do not adopt the values" turned out, in hindsight, to be correct.

The VIX intraday path (far more informative than the closing value)

Open Intraday high Intraday low Close
17.80 19.93 17.68 18.67

The pre-market list saw 17.84 and judged risk-on from it. What actually happened: it did open at a low level → spiked to 19.93 as semiconductors were sold off (a 12.7% intraday swing) → pulled back into the close at 18.67, still above the prior close. This shows today's calm was calm at the index level, not calm inside the market.

Market breadth

  • Dow +0.51% > S&P +0.02% > Nasdaq −0.18%; IWM (Russell 2000) +0.60% beat SPY (+0.02%) — small caps and equal weight stronger than large-cap tech is the classic shape of rotation, not retreat.
  • Volume clue: SMH traded 42.38 million shares, while XLK traded only 8.34 million — the only sector with elevated volume today was semiconductors, and it was heavy-volume selling. This is the most important volume signal of the day.
  • ⚠️ NYSE/Nasdaq advance-decline ratio: no reliable data. The 1.32:1 / 1.25:1 figures found in searches belong, per the verification above, to 7/24, not today; they have been removed and are not cited.

Sentiment call: risk-rotation, neither risk-on nor risk-off. Basis: ① stocks and bonds strengthening together (indices up, yields down) ② small caps beating large caps ③ VIX closing below 19 ④ money moving clearly out of semiconductors/energy into consumer staples/discretionary/financials/communications.

2. Pre-market list reconciliation

2.1 Positive-catalyst bucket (12 names)

Ticker Pre-market conclusion Today's change % After hours % Delivered? Comment
MSFT priority deep-dive (78 pts) +1.94% $389.10 +0.13% ✅ delivered Strongest of all megacaps. Money moved from "the ones getting paid for AI" to "the ones paying for AI" — the opposite of the pre-market logic, yet the direction was right
META priority deep-dive (76 pts) −0.22% $593.87 +0.40% ❌ did not deliver The only "priority deep-dive" name that fell, mildly underperforming
AMZN watch closely (71 pts) −0.31% $231.39 +0.17% ❌ did not deliver Same as META, nobody wants to bet ahead of earnings
AAPL watch closely (69 pts) worst positioning +1.17% $336.91 +0.01% ✅ delivered, but the pre-market judgment was too conservative The $336.91 close broke above the 52-week high of $334.99, an all-time high. Pre-market leaned toward avoiding it on "only −0.30% from the high, PEG 2.68", and it simply broke out to the upside
BA watch closely (63 pts) +0.95% $211.50 +0.21% ✅ delivered Mild; earnings come pre-market on 7/28, so today was a waiting game
ARM watch only (60 pts) +2.43% $266.33 −0.03% ❌ wrong call (missed-the-move type) Rose 2.43% against the tide on a semiconductor massacre day with SMH −2.25%, the second strongest on the whole list. The pre-market valuation veto on "P/S 56.4, GAAP P/E 302x" was slapped down today by the non-manufacturing nature of the IP licensing model — the Chinese equipment breakthrough hurts manufacturing and equipment, not architecture licensing
UAL watch closely (58 pts) +1.94% $120.57 −0.18% ⚠️ direction delivered, confirmation signal not met Pre-market +3.5% → closed +1.94%, giving back 45%. The confirmation signal set pre-market was "hold above +3% in the first 30 minutes", and that signal was clearly not met
MU watch only (57 pts) −2.25% $900.20 −1.42% ✅✅ precisely avoided Gapped up +2.5% pre-market → closed −2.25%; the pre-market read that "the odds of a gap-fill are not low" landed exactly. Down another 1.42% after hours
DAL watch closely (56 pts) +1.89% $86.67 +0.25% ✅ delivered Almost in lockstep with UAL
FBRX watch only (54 pts) +39.65% $76.50 +0.13% ✅✅ estimate was precise Pre-market derived "prior close about $55.00 → pre-market about $76.45 → only 0.72% from the $77 consideration". It actually closed at $76.50, 0.65% from the consideration. A 7bp error; the arbitrage-spread judgment was near perfect
LUV watch closely (52 pts) +0.67% $45.38 −0.09% ⚠️ right direction but very weak Pre-market +3.1% → closed +0.67%, giving back 78%, the worst of the three airlines
SNDK watch only (51 pts) −11.02% $1,278.23 −2.21% ✅✅✅ the most precise avoidance on the entire list Gapped up +3.6% pre-market → closed −11.02%, an intraday swing of more than 14pp. The three pre-market reasons (distorted P/E, unverified capex basis, 52-week price series suspected of not being adjusted) plus "avoid chasing a gap-up" together sidestepped the biggest trap of the day

Positive bucket subtotal: 10 of 12 directionally correct = 83.3% (6 of the 8 recommended names rose; 3 of the 4 watch-only names had their judgment hold).

2.2 Negative-catalyst / avoid bucket (6 items)

Ticker Pre-market conclusion Today's change % Delivered? Comment
CAPR avoid (S-grade negative) −64.49% $6.995 ✅✅ delivered Also resolves the pre-market basis dispute: the list recorded "−40% or −65%, basis to be verified". The −64.49% close proves the −65% set was the correct basis and the −40% set was wrong
XOM avoid −1.38% $154.77 ✅ delivered, but the reverse signal flagged pre-market has triggered Pre-market −3.2% → closed −1.38%, the decline narrowing by 57%. The pre-market note explicitly said "if XOM/CVX narrow from −3% to −1%, it means oil is being bought back and the whole theme is a one-day affair" — that self-imposed falsification signal actually lit up today
CVX avoid −2.46% $190.00 ✅ delivered Narrowed less than XOM
FANG / DVN / APA short watch (higher elasticity) −4.34% / −4.17% / −3.90% ✅✅ elasticity call was precise Pre-market said "shale falls 1.5x as much as the majors". Measured: FANG −4.34% vs XOM −1.38% = 3.1x, vs CVX −2.46% = 1.8x. Direction and magnitude both right, and even more extreme than expected
BE avoid (AI power negative) +1.78% $188.18 ❌ wrong call The pre-market note itself said "this is last week's news, no fresh catalyst today" and yet kept it on the avoid list. A negative with no fresh catalyst = an ineffective negative, and that is a methodology error
MXL avoid, and "already fully de-rated once, do not chase the short" −9.40% $64.86 ⚠️ half right, half wrong "Avoid" was entirely correct; but "do not chase the short" was wrong — it fell another 9.40% today. A valuation de-rating is a process, not an event; it does not finish in one drop

Negative bucket subtotal: 5 of 6 directionally correct = 83.3%.

2.3 Hit rate and self-critique

Overall hit rate: 14 of 18 judgments directionally correct = 77.8%

  • Positive bucket 10/12 (83.3%) · negative bucket 5/6 (83.3%) · 4 wrong calls: META, AMZN (recommended but slightly down), ARM (avoided but rose sharply against the tide), BE (avoided but rebounded)

But the hit rate overstates this list's real value, and three points of self-critique are required:

  1. [Most serious] Direction right, magnitude entirely wrong — the names that delivered generally printed their high at the open. UAL pre-market +3.5% closed +1.94% (gave back 45%), LUV pre-market +3.1% closed +0.67% (gave back 78%), and MU/SNDK flipped outright from up to sharply down. What the pre-market list captured was "the gap", not "the trend". Anyone buying at pre-market prices is today either underwater or heavily reduced, while the list shows a 78% hit rate. This hit-rate basis itself needs correcting: it should become "delivery rate measured from the opening price", not measured by the closing direction.

  2. [Methodology] Ranking a trading theme by fundamental quality ranks it backwards. Pre-market ranked the oil-price beneficiary chain UAL (1) > DAL (2) > LUV (3) > AAL (4, "heavy debt, watch only") > cruise lines (5, "peripheral beneficiary, watch only"). Today's actual ranking: RCL +3.92% > AAL +3.28% > CCL +3.00% > UAL +1.94% > DAL +1.89% > LUV +0.67%. The bottom two in the pre-market ranking rose the most today; the one ranked first came fourth. In a "one-day sentiment move with a clear direction", what determines the size of the move is beta and elasticity, not balance-sheet quality. Fundamental ranking applies to holding, not to the day.

  3. [Coverage gap] Today's strongest theme (the semiconductor-equipment negative) did not exist at all in the pre-market list. This is not a judgment error, it is a structural hole in information coverage — pre-market only scanned "overnight and pre-market news", while this report from The Information developed during the session. The list lacks any mechanism for updating "catalysts added intraday".

3. Theme verification for today

Theme Pre-market strength Today's reality Leading gainers/losers Stage Conclusion
1. Geopolitical de-escalation → oil falls S (first theme) ✅✅ Fully delivered and the only theme that delivered across the whole chain. WTI −8.29%, Brent −9.40%; JETS +2.89% vs SPY +0.02% (outperforming by 2.87pp); XLE −2.11%, the worst sector Beneficiaries: RCL +3.92%, AAL +3.28%, CCL +3.00%, UAL +1.94%, DAL +1.89% / hurt: FANG −4.34%, DVN −4.17%, APA −3.90%, CVX −2.46% The positive has been released in one go. XOM narrowed from −3.2% to −1.38%, and the self-imposed pre-market "one-day affair" signal has triggered Called right, but the durability warning was equally borne out
2. Super earnings week + AI capex guidance A+ ⏸ Not delivered (the event has not arrived). MSFT +1.94% strong on its own, META −0.22%, AMZN −0.31%, AAPL +1.17%, ARM +2.43% — scattered directions, no combined force MSFT / ARM Waiting period, pricing power sits on 7/29–7/30 Judgment holds but carried no information today
3. Memory cycle (CXMT short-term long, long-term short) A (short-term long, long-term short) ❌ Time-scale call was wrong. Pre-market estimated "the short-term long can last 1–3 days", in reality the short-term long survived only until the 09:30 open, switching to the long-term bear narrative the same day SNDK −11.02%, WDC −4.21%, MU −2.25% The long-term bear narrative has started Narrative right, duration wrong; but the "watch only" operational conclusion saved it
4. Fed rate-hike risk A (suppressive) ⏸ not delivered, but the VIX intraday path 17.80→19.93→18.67 shows there really was repricing during the session — FOMC starts 7/28 pending
5. Biotech M&A A ✅ delivered FBRX +39.65% The event is complete, only deal-closing arbitrage remains Called right
6. AI power cooling off B+ (short) ❌ wrong call BE +1.78% An old negative with no fresh catalyst Wrong call

⚠️ The surprise theme entirely missed pre-market (today's true driving force)

# Theme Trigger Today's actual performance Why it matters
A Chinese semiconductor-equipment localization breakthrough (today's strongest theme, zero pre-market coverage) The Information reported: Chinese state-backed firms have begun mass-producing key chipmaking equipment, first setting off the selloff in ASML ASML −5.80%, LRCX −4.46%, AMAT −3.61%, KLAC −3.40%, NVDA −4.99%, AMD −5.17%, SMH −2.25% (heavy volume) This is a structural narrative, not one-day sentiment. It also explains the collapse in the memory chain — pre-market attributed the fall in SNDK/MU to "the CXMT supply shock", but the real primary cause may be that same equipment story, larger in magnitude and harder to falsify
B AI circular financing and credit risk The market is again worried about circular investment between chipmakers and their AI customers; per Bloomberg, the cost of protection against a Nvidia default on 5-year debt posted a record one-day jump NVDA −4.99% (154 million shares traded, the largest in the whole market) For the first time the AI narrative is being priced from a "credit/debt" angle, rather than a "valuation/growth" angle. This is an entirely new dimension, and the pre-market scoring model has no corresponding category
C AI money rotating from "selling shovels" to "using shovels" No single event; this is a reallocation of capital NOW +6.86% (7/22 earnings beat on both lines and raised full-year subscription revenue guidance to $15.76–15.78 billion; already +6.85% on 7/24, today is the second consecutive session of large gains), GOOGL +2.13%, MSFT +1.94% vs NVDA −4.99%, AMD −5.17% Pre-market did put forward the "payer vs payee" framework, but bet on the payee (MU/SNDK) — the wrong side. Today money clearly chose the payers (software and hyperscale cloud)
D Cruise lines > airlines Internal structure of the oil chain RCL +3.92%, CCL +3.00%, both stronger than every airline except AAL Pre-market listed cruise lines as "peripheral beneficiary, watch only", on the grounds that "fuel is a smaller share of costs than for airlines". But cruise lines get a double hit from both consumer discretionary (XLY +1.31%) and oil, while airlines only get the single oil hit

4. After-hours earnings moves (7/27 after hours → next-day catalysts)

Figures are the after-hours prices recorded by stockanalysis.com at Jul 27, 2026, 6:00 PM EDT; EPS/guidance values come from CNBC's roundup (not checked item by item against company IR originals, pending verification).

Ticker Name Earnings highlights Close % After hours % Nature
CDNS Cadence Design Q2 adjusted EPS $2.11 vs LSEG consensus $2.05, a beat +3.79% +3.88% $351.75 ✅ strongest after hours today, and a countertrend signal — EDA software strengthened on both counts on the very day semiconductor equipment was killed, confirming again that "software is stronger than hardware"
WELL Welltower Raised full-year normalized FFO guidance to $6.36–6.44, above the FactSet consensus of $6.30 −1.48% +4.01% $258.30 ✅ guidance raised; senior housing real estate, an independent thread unrelated to the AI narrative
FFIV F5 FQ3 adjusted EPS $4.73, revenue $865 million, both beats +4.02% +0.50% $410.00 ✅ multicloud security; already up 4% during the session, limited incremental move after hours
RMBS Rambus Q2 adjusted EPS $0.77 (consensus $0.72), revenue $207 million (consensus $198 million) +0.43% +1.40% $97.77 ✅ a double beat in memory interface chips — on a memory massacre day with SNDK −11% and MU −2.25%, this is contrary evidence: memory interface IP and memory manufacturing are being priced apart
UHS Universal Health Cut full-year guidance to $22.28–23.65 (from $22.64–24.52) +2.29% −4.86% $151.56 ❌ the only major negative after hours today; up 2.29% during the session, then it erased all of that gain and more after hours
MU 美光 (Micron) no earnings −2.25% −1.42% $887.44 ⚠️ kept falling with no news; memory selling pressure is not over
SNDK SanDisk no earnings −11.02% −2.21% $1,250.00 ⚠️ same as above, two days of cumulative declines with no sign of stabilizing

The one line most worth remembering from this after-hours set: CDNS (EDA software) +3.88%, RMBS (interface IP) +1.40% — on the worst day for semiconductors, every semiconductor company that "does not touch wafers" rose. ARM's +2.43% today is the same logic. The negative from the Chinese equipment breakthrough hit precisely and only the parts that need to buy equipment / make wafers. This distinction is the most valuable actionable insight this recap produced.

5. Flows and sentiment

Sector ETF rotation panorama (7/27)

Leaders Gain Laggards Loss
XLP consumer staples +1.46% SMH semiconductors −2.25%
XLY consumer discretionary +1.31% XLE energy −2.11%
XLC communication services +1.28% XLU utilities −1.32%
XLF financials +1.01% XLK technology −0.90%
XLV health care +0.51% XLRE real estate −0.41%
XLI industrials +0.30%
XLB materials +0.25%
JETS airlines +2.89%

Seven up, five down — but all of today's volume sits in the five that fell.

Key readings

  • JETS +2.89% clearly beat SPY (+0.02%) by 2.87 percentage points — the pre-market "confirmation signal: does JETS beat SPY" was clearly met, so the oil theme was not a lone-stock move but a real sector-level flow.
  • XLU −1.32% is an overlooked signal: utilities down plus real estate −0.41%, while the 10Y yield was falling. Defensive duration assets falling while rates fall means this is not safe-haven money buying defensives, it is the AI power narrative (data-center electricity demand) receding. This is directionally consistent with pre-market theme 6; only the single stock BE failed to follow.
  • VIX 18.67 (+0.48%), intraday 17.68–19.93: a 12.7% swing. The tiny closing gain masks the real panic during the session.
  • 10Y 4.641% (−3.8bp): oil −8% transmitting into inflation expectations, yields down. Stocks and bonds strengthening together is the strongest evidence for "rotation" rather than "retreat".
  • Volume structure: SMH 42.38 million shares (heavy-volume decline), NVDA 154 million shares, INTC 133 million shares — today's money changed hands almost entirely inside semiconductors, and the gains in other sectors were passive absorption on light volume.
  • FedWatch status: the latest primary data found is on a 7/25 basis, roughly a 38.7% probability of a hike (corresponding to 61.3% for no change; the current fed funds rate is 3.50–3.75%). ⚠️ That snapshot still predates the 7/26 oil crash, and the repricing after the oil plunge could not be obtained — this is the same data gap for the second day running.

Qualitative conclusion

risk-rotation, not risk-off. Basis: ① stocks and bonds strengthening together ② small caps (IWM +0.60%) beating large caps ③ financials/discretionary/communications all up ④ VIX closing below 19 ⑤ seven sectors higher.

But it must equally be acknowledged: the only sector with elevated volume was the one that fell. The combination of light-volume gains plus heavy-volume declines means the bulls are passive and the bears are active. This is not a healthy +0.02%.

6. Next-session outlook (2026-07-28, Tuesday)

① Theme durability

Theme Durability Reason
Semiconductor equipment / AI hardware valuation kill High ⬆ The only direction backed by a structural narrative (equipment localization + circular financing + widening CDS), and today was a heavy-volume decline. The three reasons are mutually independent, not one-day sentiment. KLAC's 7/28 after-hours earnings land right in the line of fire.
Software/IP > hardware/manufacturing Medium-high ⬆ Four independent pieces of evidence pointing the same way today: ARM +2.43%, NOW +6.86%, CDNS +3.88% after hours, RMBS +1.40% after hours. But it is highly sensitive to the capex guidance from MSFT/META on 7/29
Beneficiary chain of falling oil (airlines/cruise lines) Medium-low ⬇ The positive was released in one go (WTI −8.29% in a single day). XOM's decline narrowing by 57% has already triggered the pre-market self-imposed "one-day affair" signal. And the validity period of the "pause in strikes" and the negotiation calendar still have no public disclosure — that falsifiable anchor raised pre-market is still missing today
Shorting energy Medium-low ⬇ Same as above; the reverse risk (oil being bought back) is starting to outweigh the reward
Super earnings week Entering the delivery phase BA fires the first shot pre-market on 7/28

② Tomorrow's earnings and macro calendar (ET)

Macro (highest weight)

  • FOMC meeting begins (7/28–29), decision 7/29 14:00, Chair Kevin Warsh press conference 14:30. Today is the first day of the pre-meeting blackout period.
  • Conference Board Consumer Confidence Index, S&P/Case-Shiller Home Price Index, FHFA House Price Index (May), wholesale inventories (June advance)

Pre-market earnings: BA (波音 Boeing, already on the pre-market list), KO, UPS, PYPL, SPGI, RCL, HLT After-hours earnings: V (Visa), KLAC, F (福特 Ford), MDLZ

③ Key names to watch (ticker + verification point)

Ticker Today's close Why it matters Verification point
KLAC $203.36 (−3.40%) ⭐ the name that most should have been added to the list today and was not. Earnings after the close on 7/28, and the directly falsifiable object of the "Chinese equipment breakthrough" narrative China revenue share and equipment order guidance. If China revenue shows no decline → today's equipment selloff was an overreaction; if management proactively flags domestic substitution → the narrative holds and ASML/LRCX/AMAT have a second wave ahead
BA $211.50 (+0.95%) Pre-market 7/28, the first heavyweight report of the week, setting the tone for the whole week Whether free cash flow is still negative (the pre-market list already fixed this as the only number worth looking at), whether deliveries can sustain +14% year over year, the 787 engine bottleneck
NVDA $196.51 (−4.99%) The worst-performing megacap today, 154 million shares traded Whether it can hold $196.51; more critically, whether the 5-year CDS spread keeps widening — if the credit market keeps pricing risk, technical support in the share price is meaningless
RCL $305.04 (+3.92%) The strongest name in the oil chain today, and it reports pre-market on 7/28 — a double catalyst Whether the earnings can support today's gain; this is the only chance to verify whether "cruise lines > airlines" is structural rather than a one-day phenomenon
ARM $266.33 (+2.43%) Rose against the tide on a semiconductor massacre day, the strongest sample of the "IP ≠ manufacturing" logic If semiconductors keep falling tomorrow and ARM again resists → the logic holds and can be upgraded to a theme; if it falls with them → today was just a short squeeze. Note it reports after the close on 7/29
MU / SNDK $900.20 / $1,278.23 Still falling after hours (−1.42% / −2.21%), selling pressure not exhausted Whether they retest lower. SNDK's earnings are far off on 8/5, with no fundamental support in the interim

④ What to avoid

  1. Chasing today's oil-price beneficiary chain (UAL/DAL/LUV/RCL/CCL) — the one-off positive has been released, the "one-day affair" signal set pre-market has already lit up today (XOM's decline narrowing by 57%), and the risk that "a pause ≠ a ceasefire agreement" has not eased in the slightest.
  2. Bottom-fishing semiconductor equipment (ASML/LRCX/AMAT/KLAC) — a structural narrative built on three independent negatives stacked together, and today was again a heavy-volume decline; it does not finish falling in one day. MXL fell another 9.40% today under the judgment that it was "already fully de-rated", which is the freshest lesson available.
  3. Taking a heavy position in any direction the day before the FOMC — the latest available basis for the hike probability is still 7/25's 38.7%, and it predates the oil crash; the true pricing after oil's −8% is something we do not know. Betting on a direction without knowing where the market is priced is gambling, not trading.
  4. CAPR — the FDA advisory committee meets 7/29, it is already −64.49% today, a binary event, unsuitable for either long or short.
  5. Chasing the four trillion-dollar giants ahead of earnings — especially AAPL, which just hit an all-time high of $336.91 today, with earnings after the close on 7/30 and IV crush risk further elevated relative to the pre-market point in time.
  6. UHS — full-year guidance cut, −4.86% after hours, likely to catch down the next day.

⑤ Input hints for the next pre-market list (retraining feedback)

The weekly recap loop resumes from today. The following six points should feed directly into the 7/28 pre-market scoring corrections:

  1. [Highest priority · coverage hole] Add "Chinese semiconductor-equipment localization + AI circular financing/credit spreads" to the theme pool and grade it S. This was the single biggest omission on today's entire list, and it is more fundamental than the CXMT thread — pre-market attributed the fall in the memory chain to CXMT capacity, but the real primary cause is very likely the equipment story. Get the attribution wrong and the time-scale judgment that hangs off it is wrong too.
  2. [New actionable dimension] Semiconductors must be priced by splitting them on "does it touch wafers". Today's evidence chain: touching wafers — ASML −5.80% / LRCX −4.46% / AMAT −3.61% / KLAC −3.40% / SNDK −11.02% / MU −2.25% / WDC −4.21%; not touching wafers — ARM +2.43% / CDNS after hours +3.88% / RMBS after hours +1.40%. Seven to three, zero counterexamples.
  3. [Ranking-method correction] Within a trading theme, rank by beta/elasticity, not by fundamental quality. Today's hard proof: pre-market ranked AAL fourth (watch only, on the grounds of "heavy debt") and cruise lines fifth (watch only), while in reality RCL +3.92% / AAL +3.28% / CCL +3.00% all beat the pre-market number one, UAL (+1.94%). Fundamental ranking answers "who to hold", not "who rises most today".
  4. [Time-scale calibration] The default for "short-term long" should be set to 0 days, not 1–3 days. CXMT's "short-term long" actually survived only until the open. Criterion: a cross-time-zone sentiment gap + no fresh domestic US catalyst = a same-day gap-fill, and no durability points.
  5. [List hygiene] An old negative with no fresh catalyst should be removed from the avoid list. The root cause of the wrong BE call (+1.78%) is that pre-market itself wrote "this is last week's news, no fresh catalyst today" and yet kept the entry. A negative with no catalyst = an ineffective negative.
  6. [Outstanding data gap · second day running] A primary CME FedWatch snapshot after the 7/26 oil crash still has not been obtained. The latest available is 7/25's 38.7%. It must be filled in before the 7/29 decision, otherwise the scoring of the entire macro theme rests on stale data.

Also: two pre-market disputes were closed out today — ① the CAPR −40% vs −65% dispute: the −65% basis was correct (measured −64.49%); ② the crude-oil value conflict is resolved: the true sequence is 7/24 close WTI $89.31 / Brent $96.78 → 7/27 close WTI $81.91 / Brent $87.68.

Appendix: data gaps proactively flagged by this recap

Following the no-fabrication principle, the following are known but could not be confirmed:

  1. US Dollar Index (DXY) 7/27 close — the CBOE endpoint refused access and yfinance was rate-limited, no reliable data.
  2. NYSE / Nasdaq advance-decline ratio — the 1.32:1 / 1.25:1 figures found belong, on cross-verification, to 7/24, and have been removed. There is no reliable source for today's breadth data, so the breadth judgment in §1 is instead derived from relative strength across indices/ETFs.
  3. A primary CME FedWatch snapshot after the oil crash — the latest available is 7/25's 38.7%, which still predates the 7/26 event. Missing for the second day running.
  4. EPS / guidance figures for after-hours earnings — taken from CNBC's roundup, not checked item by item against company IR originals or SEC 8-Ks.
  5. NVDA 5-year CDS "record one-day jump" — relayed from media citing a Bloomberg report, the original credit-market data was not obtained.
  6. "Apple's market cap has overtaken Nvidia's for the top spot" — this statement appears in an article describing the 7/24 tape; today AAPL +1.17% while NVDA −4.99%, so the gap should have widened further, but the precise market caps of the two were not independently verified by this report.
  7. The attribution for ServiceNow — NOW's earnings were released on 7/22, and today's +6.86% is a follow-through gain on the second trading day after earnings (already +6.85% on 7/24), not driven by earnings on the day. The causal link between that gain and the "software stronger than hardware" rotation is an inference, not something confirmed.
  8. The Information's original report on Chinese chip equipment entering mass production — this report learned of it via second-hand media accounts, the original was not obtained, and neither the type of equipment nor the scale of production is confirmed.
  9. yfinance was broadly rate-limited this time, so single-stock and ETF data switched to stockanalysis.com. Consistency between the two sources on adjustment and after-hours basis was not verified item by item.

Sources

⚠️ Risk disclaimer: this recap is only a post-close information review and observation, and does not constitute investment advice. Data may differ in timeliness or basis; please rely on company disclosures/SEC filings, and do not use this directly as a basis for trading.

Sources11

Every external link cited in the body, numbered in order of appearance. · 8 domains

  1. 1The Motley Fool — Stock Market Today, July 27: Dow Rises on Oil Retreat, and Sandisk Plunges 11% on Memory Weaknessfool.com
  2. 2Yahoo Finance / Zacks — Stock Market News for July 27, 2026 (verified to actually describe the 7/24 tape; this report reassigns it accordingly)finance.yahoo.com
  3. 3The Globe and Mail — Stock Market News for July 27, 2026 (same as above)theglobeandmail.com
  4. 4Yahoo Finance — Nvidia drops nearly 5%, leading chip stocks lower amid renewed worries of circular financingfinance.yahoo.com
  5. 5Blockonomi — Semiconductor Stocks Plunge as China's Chipmaking Breakthrough Shakes Marketsblockonomi.com
  6. 6CNBC — Stocks making the biggest moves after hours: Cadence Design Systems, Rambus, Welltower and morecnbc.com
  7. 7ServiceNow IR — ServiceNow Reports Second Quarter 2026 Financial Results (7/22)investor.servicenow.com
  8. 8SEC EDGAR — ServiceNow Q2 FY2026 Earnings Release (EX-99.1)sec.gov
  9. 9CNBC — Stock market next week: Outlook for July 27-31, 2026 (next-day calendar)cnbc.com
  10. 10Schaeffer's — The Week Ahead: Fed Decision Accompanies a Big Tech Earnings Blitzschaeffersresearch.com
  11. 11Yahoo Finance — The Probability of a July Fed Rate Hike Has Tripled Over the Last Week (FedWatch, 7/25 basis)finance.yahoo.com