Share
Research
← All trending days

Market Pulse · Daily Scan

Daily Market Pulse

Jul 31, 2026 · sentiment/flow window ≈ 48h

Pre-marketnot publishedPost-market
TLDR;
The AI-capex revolt did not end this week - it became DISCRIMINATING, and the market spent two sessions sorting the complex into names that can show a revenue offset and names that cannot. Over Jul 30-31 four hyperscalers collectively guided 2026 capital projects to $720-745B, more than any prior quarter's aggregate, and the market rallied anyway. The sorting mechanism was brutal and consistent: Microsoft (+15.5%) and Meta (-7.95%) reported the SAME night, both raised capex, and only the one with a third-party cloud revenue line was paid. Amazon then repeated the experiment on Jul 31 with AWS +37% y/y - its fastest since 2021 - and capex to $220B, and was rewarded +15.63%, while Apple fell 7.13% on a supply-constrained guide. The second thread is memory: Samsung's record print carried CFO guidance that the shortage deepens in 2027 and persists through 2028, which rebutted the CXMT-oversupply panic that had driven the July semi rout. Micron +18.36%, SanDisk +23.7%, Lam Research +18% (best day since 1999), and Korea's KOSPI up 17.9% for its largest single-day gain on record. The reflexive twist worth holding onto: the same memory shortage that made Micron the window's best long is exactly what Apple cited to cap its own guidance. Breadth: Narrow and mega-cap-driven on Jul 31, broader on Jul 30. On Jul 31 the Russell 2000 FELL 0.50% to 2,931.34 while the S&P added 0.70% and the Nasdaq 1.00% - small caps went the other way. One intraday breadth read at 11:56 ET had five megacap Dow components contributing roughly 114% of the index's entire gain while the other 25 subtracted about 22 points. On Jul 30 the S&P rose 1.7% but shortly after the open only 243 of 500 constituents were higher and just five of eleven sectors were positive. Two consecutive up-sessions on the headline, neither of them broad..
A · Daily Pulse

What moved — and why

▲ Rising — dailycomposite
#TickerMoveDriverScore
1AMZN↗—The cleanest hyperscaler print of the quarter and the name that closed the AI-capex argument for the week. Q2 revenue $200.6B beat $196.5B (+20% y/y); AWS grew 37% y/y against a 31% consensus - its fastest expansion since 2021 - at a reported 36.7% operating margin; capex guided to $220B for 2026. Wedbush called it the cleanest beat among the hyperscalers and specifically flagged that management gave the most explicit walk-through of how it reaches ROIC on the spend. That last point is the whole story: the market is no longer pricing capex level, it is pricing capex PROOF. Amazon was the largest single positive contributor to the Dow on the session.
89
2MU—Largest verified single-session gain in the window: +18.36% on Jul 30 to close $874.66, off a $739 Jul 29 close - roughly +24% in two sessions from Wednesday's low near $707. The trigger was NOT company-specific: Samsung reported record quarterly earnings and its CFO told analysts the memory shortage deepens in 2027 and persists through 2028, which directly rebuts the CXMT-oversupply thesis that had taken MU down 41% from its $1,255 high. Korea confirmed the read on Jul 31 with SK Hynix +30% and Samsung +28%. TWO-SIDED CAVEAT, stated rather than resolved: Michael Burry disclosed an INCREASED short near $880, MU traded down more than 5% intraday on Jul 31 (close not confirmable on the free tier), and a company sitting on record revenue needing a competitor's print to move its own stock is itself a signal about who is setting the price. ↔ deep-dive
85
3MSFT↗—The print that turned the week. FQ4 revenue $90.01B beat $87.62B; Azure grew 43% in constant currency against a 40.2% StreetAccount estimate; annual Azure revenue crossed $100B for the first time. Shares closed +15.5% on Jul 30 and added roughly another 3% on Jul 31. The structural read matters more than the number: Microsoft and Meta reported the same night, both raised capex, and the market rewarded one and punished the other. That is the AI-capex revolt becoming DISCRIMINATING rather than ending. Also appears as Track B upside rank 3 - and the Track A gain is precisely what cut the Track B valuation sub-score. ↔ deep-dive
84
4LRCX↗—Best single day since 1999 on strong earnings and AI-driven guidance. Analytically this is the more important semicap datapoint of the two sessions: on Jul 29 the market refused to pay KLA for a 60% backlog build, and forty-eight hours later it paid Lam 18% for the same category of news. That reversal is the cleanest available evidence that the July semicap de-rating was positioning-driven rather than a demand re-rate. Market cap not independently pulled this run - Lam is a SOX constituent that Yahoo Finance measured as having lost more than $100B of value since Jun 25, so eligibility is not in doubt, but the figure is flagged rather than estimated.
81
5SNDK—Largest percentage gain of any ranked name in the window: +23.70% on Jul 30 (one source chain says +26%; the two figures were not reconciled, so treat magnitude as provisional). Same Samsung shortage catalyst as Micron, amplified by a much smaller and far more crowded float. PUMP_RISK stays flagged on flow characteristics - this name ran +2,881% over 52 weeks, fell roughly 36% in five sessions, then added 24-26% in one - but the flag is a position-sizing warning, not a fraud signal: the underlying business carries roughly $3.0B of operating cash flow and 56% gross margins. Reports Aug 5 AMC, which is the binary that resolves whether this was a real re-rate or a squeeze. pump risk
81
6AMD↗—Rose more than 13% on Jul 30 in the broad semiconductor snapback. No AMD-specific catalyst - this is high-beta participation in the Microsoft/Samsung read-through, which is exactly why momentum is the only high sub-score here. Reports Aug 4 AMC with weekly IV above its 52-week range, so the event is the story from here, not the bounce. Also carried as Track B downside rank 9; the two tracks are scored independently and point in opposite directions by design - do not net them.
78
7INTC↗—Up 12.53% on Jul 30 on a better-than-expected Q2 and a raised full-year guide. This is the single most uncomfortable datapoint in the artifact, and it is being reported rather than smoothed: Intel is Track B downside rank 2, and a beat-and-raise that the market actually PAYS for is direct evidence against that short. The counterweight is that the same stock had already fallen more than 30% from its late-June peak into this print and finished July down 41.36%, so a 12.5% rally does not by itself re-rate a 54.8x-forward company with no trailing earnings. The Track B entry has been moved to watch_for_invalidation rather than quietly re-scored.
77
8GOOGL—Second-largest Dow contributor on Jul 31 at +7.12%, on a two-part catalyst: Google DeepMind released a set of embodied-AI models (Gemini Robotics 2, Gemini Robotics ER 2, On-Device 2), and Amazon's AWS print retroactively validated the hyperscaler-cloud demand case that Alphabet's own Jul 22 report - which took the stock down roughly 15% over two days on a first-ever negative free-cash-flow quarter - had failed to sell. MAGNITUDE CONFLICT, flagged not averaged: Trading Economics' post-close Dow component list gives +7.12%, its own body text +6.9%, one aggregator +6.7%, and TradingKey's intraday market-movers page +3.12%. The close-basis Dow list is used; the spread is disclosed.
76
9NVDA↗—Rose on both sessions (+1.67% then +3.46%) and ranks last on this list for exactly that reason - it is the smallest mover here and it lagged the complex badly on Jul 30, gaining under 2% on a day SOXX added more than 8% and Micron added 18%. That lag is the useful signal: the snapback was concentrated in the names that had been liquidated hardest, not in the sector's anchor. Separately, CEO Jensen Huang is scheduled on Capitol Hill to discuss AI leadership, open-source AI and a stated plan for up to $500B of domestic AI infrastructure - a policy catalyst with no scheduled date inside this horizon.
66
▼ Falling — dailycomposite
#TickerMoveDriverScore
1AAPL↗—The largest single drag on the tape and the most instructive print of the window. Apple BEAT on the quarter - FQ3 revenue $109.4B (+16% y/y) and EPS $2.02 (+29%), a record June quarter with double-digit growth across iPhone, Mac and Services - and fell 7.13% anyway, because it guided FQ4 revenue growth to 9-11% against a 12% consensus, citing advanced-chip and MEMORY supply constraints rather than weak demand. Services at $30.74B and Greater China both missed. The reflexive irony is the point: the same memory shortage that made Micron the window's best performer is what capped Apple's guide. Tim Cook's final earnings call before John Ternus takes over Sep 1. PT reactions split - Barclays $253 to $245, Goldman $370 to $360, Morgan Stanley $364 to $360, but Wells Fargo RAISED to $350 from $310 and BofA reiterated $380. MCAP CONFLICT: the $4.90T figure is the Jul 30 stockanalysis basis; one Tier-3 source put the post-move cap at $4.4T, and headline estimates of value destroyed ranged from ~$350B to ~$500B. Not point-estimated. ↔ deep-dive
87
2RBLX—The largest single-name decline in the window that passes the universe filter - and the one where this pipeline explicitly refuses to publish a number. Roblox issued a weak next-quarter outlook, WITHDREW its full-year guidance and missed key user-growth estimates; withdrawn guidance is a company-disclosure event of Tier-1 severity regardless of the price print. MAGNITUDE IS NULL BY DESIGN: the widely circulated -22.5% comes from TheStreet's 08:11 ET PREMARKET movers block, and no close-basis figure was obtainable from any free source before publication. A technical-analysis note put the post-fall trading level near $35.50 against a strike-50 straddle the prior evening, which corroborates a decline of roughly that order without confirming it. Market cap $34.85B is a Jul 31 close-basis figure from S&P Global via stockanalysis, so universe eligibility is confirmed even though the move is not.
84
3COIN—Third consecutive quarterly loss. Q2 total net revenue $1.22B (-14% y/y) missed a ~$1.35B estimate, transaction revenue fell 21% to $599M, net loss $359.5M, and headcount was cut 14% to 4,321. Bitcoin fell below $63,000 the same session, so the equity and its underlying beta broke together rather than the stock absorbing an idiosyncratic miss. MAGNITUDE FLAGGED: post-close sources cluster around -10%, but the same session produced a -4.5% premarket print, a -6.53% after-hours print on Jul 30, and a -14% intraday figure. The -10% used here is the post-close consensus of two independent sources, marked approximate rather than exact. Market cap not independently pulled; eligibility is not in doubt at any plausible figure.
79
4META↗—Fell 7.95% on Jul 30 after Q2 EPS of $6.18 missed a $7.22 estimate on higher legal and severance charges - even with revenue up 28% to $60.8B - then recovered roughly 3.3% on Jul 31. The company raised the LOW END of full-year capex guidance to $130-145B and guided Q3 revenue below consensus. Paired with Microsoft the same night, this is the defining structure of the week: both raised capex, one was rewarded 15.5% and one was punished 7.95%. The differentiator was a demonstrated third-party revenue offset, which Meta does not have. MAGNITUDE CONFLICT: CappNotes gives -7.95% close-basis, CNBC 'more than 9%', one Tier-3 aggregator -9.64%. The CappNotes close-basis figure is used and the spread disclosed. Carried as Track B upside rank 9 with watch_for_invalidation - the daily evidence actively cuts against the 12-month thesis, which is stated rather than reconciled. ↔ deep-dive
78
5NVO—The strongest Tier-1 catalyst in the window and a direct confirmation of an existing Track B downside position (rank 3). Novo announced that ziltivekimab FAILED its primary endpoint in the Phase 3 ZEUS trial: hazard ratio 0.99 against placebo (95% CI 0.88-1.11) across more than 6,300 patients with atherosclerotic cardiovascular disease, chronic kidney disease and inflammation. The drug demonstrated target engagement - free IL-6 and hsCRP both fell as expected - and that biology simply did not convert into fewer cardiovascular deaths, heart attacks or strokes. Jefferies sized the MACE opportunity above $10B annually. UBS modelled only 1.7% of NPV but expects a market reaction 2-3x the pure NPV impact because the failure compounds a run of R&D setbacks and closes off cardiovascular as a second growth pillar for a company with ~94% sales reliance on one franchise. A non-cash Q3 impairment follows; 2026 adjusted operating profit guidance is unchanged. HERMES and ARTEMIS read out in H1 2027. MAGNITUDE: sources span -7.42% (Copenhagen listing) to -10% intraday on the ADR, with post-close ADR figures clustering -8.78% to -9.4%; -8.8% is used and marked approximate.
78
6CVNA—Fell nearly 12% on Jul 30 after a disappointing full-year profit outlook and weaker gross-profit metrics - a rare clean consumer-credit-adjacent datapoint in a window otherwise dominated by AI capex. Worth isolating precisely because it is NOT an AI story: with the 30-year at a 19-year high and the July consumer complex (Roblox, Reddit, GoDaddy, Apple's China line) all disappointing on the same two days, the consumer side of the tape deteriorated while the infrastructure side re-rated upward. Market cap not independently pulled; eligibility not in doubt.
64
7BA—Second-worst Dow component on Jul 31 at -1.90%, with no confirmed single-name catalyst on the session - it reported Jul 28, so this is post-earnings drift plus industrial weakness rather than news. Included because it is close-basis verified from a post-close source and the alternative would be padding this list with premarket prints. Ranked near the bottom precisely because the magnitude is small and the driver is unconfirmed. unverified
47
8UNH↗—Third-worst Dow component on Jul 31 at -1.58%, no confirmed session catalyst. Carried here mainly for continuity of the record: UNH was formally retired from Track B as thesis-invalidated in an earlier run and later flagged in the upside runners-up on a Q2 beat with the medical cost ratio at a two-year low, so it stays visible as an ex-position rather than disappearing. Small magnitude, unconfirmed driver, ranked last accordingly. unverified
46
B · One-Year Conviction

Asymmetric 12-month skew

Valuation and market-cap figures are carried from stockanalysis.com (Tier-2 aggregator republishing S&P Global Market Intelligence) as of 2026-07-30 and are therefore ONE SESSION STALE for every name that moved on Jul 31 - materially so for AAPL (-7.13%), AMZN, GOOGL (+7.12%) and NVO (~-8.8%). They are flagged as as-of figures rather than re-stated. Tier-1 sources standing behind these theses: Micron FQ3-2026 via SEC EDGAR, NVIDIA Q1 FY2027 IR, GE Vernova Q2-2026 IR, Alphabet Q2-2026 slides, Novo Nordisk company release on ZEUS, FOMC statement.

▲ Upside skew — 1yrconviction
TickerThesisKey riskScore
TSMThe unavoidable chokepoint of the AI buildout, and the only large-cap in the complex that converts the capex boom into cash rather than consuming it. Full-year revenue growth guided above 40% with capex lifted to $64B; pricing power sits simultaneously over Nvidia, AMD, Broadcom and Apple. The window added a second-order confirmation rather than a direct one: Apple explicitly blamed its FQ4 guide on competition for advanced-chip manufacturing capacity, which is what a foundry monopoly looks like from the customer's side of the table. Taiwan geopolitical tail risk is the entire bear case and is not diversifiable; Customer concentration in Nvidia; Rising capex intensity compresses FCF conversion; A 2027 hyperscaler capex cut hits with a 2-3 quarter lag
87
NVDA↗Q1 FY2027 (Tier 1, company IR): revenue $81.6B +85% y/y, data center $75.2B +92%, non-GAAP gross margin 75.0%, Q2 guide $91.0B +/-2% while assuming ZERO China datacenter compute revenue, plus an $80B buyback authorization. The window's contribution is that the customer base publicly recommitted: four hyperscalers now guide to $720-745B of cumulative 2026 capital projects, and both Microsoft and Amazon demonstrated the revenue offset. At roughly 19.6x forward against 85% growth, a substantial deceleration is already discounted. Memory input-cost inflation is now visible in the bill of materials - memory at 29% of total cost against a ~20% target - and Samsung has now guided that shortage through 2028, which raises the cost floor rather than lowering it; $410.6M of insider selling over three months with zero buys; Custom-ASIC share loss to Broadcom, TPU and MTIA; Beta of 2.21 makes it the wrong vehicle if the capex thesis is wrong; Lagged the Jul 30 snapback badly (+1.67% on a day SOXX added 8%), which says the marginal buyer is chasing beaten-down names, not the anchor
85
MSFT↗Thesis confirmed on the tape and therefore partly spent. FQ4: revenue $90.01B vs $87.62B expected, Azure +43% constant currency against a 40.2% estimate, annual Azure revenue past $100B for the first time. The stock added roughly $450B of market value on Jul 30 and contributed more than $600B for the week. The 12-month case no longer rests on a de-rating being unwound - that has now happened - so from here it rests on Azure holding a 40%-handle growth rate against a much larger base. This is a deliberately honest downgrade of the valuation leg rather than a victory lap. ↔ deep-diveThe cheap-entry argument is gone: this is now a growth-durability call, not a re-rating call; Azure growth is materially OpenAI-linked, concentrating exposure to one counterparty's funding; Capex up sharply y/y and raised again with this print; The pre-print options market underpriced this outcome by ~2.2x, which cuts both ways - the market's read on Microsoft is unstable in both directions
84
AVGO↗The custom-ASIC leg is the structural AI trade: every hyperscaler dollar spent reducing Nvidia dependence is by definition spent with Broadcom. Q1 FY2026 AI semiconductor revenue $8.4B +106% y/y on total revenue $19.31B +29%, with a $73B AI backlog deliverable over 18 months. AVGO appeared in Market Rebellion's Jul 31 'popular stocks' and 'active options' lines, so positioning is live even though no name-level metric was obtainable. Next earnings Sep 3 - outside this horizon, which is itself relevant: there is no scheduled event to resolve the thesis inside five sessions. Customer concentration (Google, Meta, OpenAI); The '$100B AI revenue in 2027' figure is a CEO statement on a call, NOT formal guidance; Trailing 61.6x reflects acquisition amortization so GAAP optics screen badly; No catalyst inside the horizon means no near-term falsification test
81
MUSPLIT CONVICTION, still stated rather than resolved - but the split has narrowed. Bull: FQ3 2026 (Tier 1, SEC-filed) revenue $41.46B against $9.30B a year earlier, GAAP diluted EPS $24.67, operating cash flow $25.39B, FQ4 guidance $50.0B +/-$1.0B at ~86% gross margin, 16 supply agreements with 14 carrying minimum commitments of roughly $100B through 2030. Samsung - the largest competitor, with every incentive to talk supply UP - instead told analysts unmet 2026 demand carries into 2027 and constraints worsen through 2028. That is the strongest available rebuttal to the oversupply thesis and it came from an adversarial source. Bear, undiminished: an ~85% gross margin is roughly 25 points above any historically sustainable level, customer inventories have moved from 2-4 weeks to 7-9 against an 8-week warning threshold, and Michael Burry publicly INCREASED a short near $880 into the bounce. A 5x forward multiple on peak-cycle earnings remains the textbook signature of a cycle top. ↔ deep-divePeak-cycle earnings at a trough multiple - the classic value trap in semis; Needed a competitor's earnings call to move its own stock, which says who is setting the price; Customer inventories 7-9 weeks vs an 8-week warning threshold and a 10-week exit trigger; Highest-variance name on either list - the 12-month distribution is bimodal, not merely uncertain; Market cap and Jul 31 close both carry unresolved source conflicts
78
GEVElectricity, not silicon, is the binding constraint on AI. Q2 2026 (Tier 1, company release): orders $24.2B +88% organic, backlog $176B, gas power equipment backlog 100GW to 116GW, revenue $11.1B +22%, free cash flow $5.1B in the quarter - more than all of 2025 - with FY26 FCF guidance raised from $6.5-7.5B to $11.5-12.5B. Strongest verified estimate-revision evidence on either list. The window's capex reconfirmation is a direct read-through: turbine slots are the scarcest physical asset in the buildout. Highest headline multiple on this list at ~52x forward; Wind segment remains a drag; Turbine slot reservations are not firm revenue; A one-year New York State moratorium on 50MW+ data centers shows regulatory bottlenecks are real; Fell more than 7% on its own Jul 22 print on margin concerns - the market is not paying for backlog alone
77
LLY↗Clear structural winner of the GLP-1 duopoly, and the window made the contrast quantitative rather than narrative. Lilly is compounding roughly 29% EPS while Novo has guided 2026 to a 5-13% constant-currency sales DECLINE and, on Jul 31, failed the ZEUS cardiovascular outcomes trial that was its most visible attempt to build a second franchise. Oral orforglipron extends Lilly past injectable supply constraints. Included as much for portfolio construction as for absolute upside - the only top-10 upside name with zero exposure to the hyperscaler capex question. Reports Aug 5 BMO, inside the horizon. Consensus PT implies only ~11% upside - the easy multiple expansion is behind it; Oral pill data was characterised as disappointing by some outlets; US drug pricing / most-favored-nation policy risk; ~34x forward leaves no room for a manufacturing stumble, and Aug 5 is a live binary inside the horizon; A rival's trial failure is not the same as own-pipeline success - this is a relative, not absolute, improvement
75
GOOGLBest full-stack AI position in the world - TPUs, models, distribution and cloud - and the only Mag-7 name whose revenue growth ACCELERATED on AI. Q2 2026: revenue $119.8B +24%, operating income $40.8B +30%, Google Cloud $24.8B +82% with cloud operating margin expanding to 35.6% from 20.7%, and a $514B cloud backlog. The stock had fallen roughly 15% over two days on its own print; it recovered 7.12% on Jul 31 when Amazon's AWS number made the same argument more credibly than Alphabet had managed itself, alongside DeepMind's embodied-AI model release. Why it ranks 8th and not higher: 2026 capex guided to $195-205B, Q2 free cash flow -$5.9B (first negative quarter since the 2004 IPO), buybacks suspended; Street models FCF going from +$13B in 2026 to -$13B in 2027; EARNINGS QUALITY: Q2 diluted EPS of $9.11 was materially inflated by unrealized equity-securities gains; that EPS is deliberately excluded from the valuation sub-score; Antitrust remedies remain live; A robotics model launch is a narrative catalyst, not a revenue one - the +7.12% is sentiment repair, not an estimate change
74
META↗CONTRARIAN, and the tension is now sharper rather than resolved. You are paying roughly 17.6x forward - the cheapest multiple in mega-cap tech - for a business that grew revenue 28% to $60.80B with ad impressions +14% and price-per-ad +12%. The core business is accelerating. But Q2 EPS of $6.18 missed $7.22, operating margin compressed to 31% from 43%, free cash flow fell to $784M from $8.55B, and 2026 capex is guided to $130-145B with the LOW end raised. Reporting the same night as Microsoft produced the cleanest possible controlled experiment, and the market's verdict was unambiguous: +15.5% for capex with a demonstrated third-party revenue line, -7.95% for capex without one. ↔ deep-diveTHE STRUCTURAL FLAW, now market-tested: unlike MSFT, GOOGL and AMZN, Meta has no external cloud revenue line to monetize the buildout; Reality Labs lost $4.03B in the quarter; Q3 revenue guided to $61-64B, the low end below the $63.15B consensus; Pre-print options were 2:1 call-skewed into a 7.95% decline - the crowd was positioned wrong and may still be unwinding; WATCH_FOR_INVALIDATION: if the Q3 print again shows capex rising with no offset, this position should be retired rather than re-scored
67
VRTPure play on data-center thermal management and power distribution - the physical bottleneck layer. Rack densities from Rubin-class systems make liquid cooling non-optional, converting a cyclical hardware business into a content-per-rack growth story. The $720-745B hyperscaler capex aggregate is the direct demand input. Ranked last on this list because the evidence remains two-sided: the Q2 revenue miss ($3.27B vs $3.38B) that took the stock down 17.26% on Jul 29 is exactly the execution slip a ~29x forward multiple cannot absorb repeatedly, and the options tape is skewed against it. Highest beta on this list to any hyperscaler capex cut; Competitive entry from Schneider and Eaton; Options positioning is actively bearish: 30-day IV 73 against a 52-week range of 45-85 with call:put 1:1.4 toward PUTS on 104K contracts (Jul 30 pre-market snapshot); Ranked last precisely because the recent daily evidence cuts against the 12-month thesis
64
▼ Downside skew — 1yrconviction
TickerThesisKey riskScore
ORCLThe only name on either list where the risk is solvency-adjacent rather than valuation. Oracle levered its balance sheet to fund a ~$300B OpenAI contract and the credit market repriced it: CDS spreads hit a record 198.23bp and S&P downgraded to BBB- - one notch above junk, which would force selling by investment-grade-mandated funds. Total debt $167.43B against $43.06B of equity, TTM free cash flow -$23.69B, with roughly $40B of further financing planned for FY2027. In a regime where the front end still prices a possible HIKE and the 30-year sits at a 19-year high, this is where funding tightening surfaces first. THE HONEST COUNTER-ARGUMENT: the stock is already -49%, 14.6x forward is not demanding, and the 3-year EPS growth forecast is 27.37%; Any OpenAI equity or financing event that de-risks the counterparty; A successful $40B raise at reasonable spreads collapses the CDS and the thesis with it; The window's capex validation raises the probability that OCI demand is real, which is the wrong direction for this short
78
NVOWas already the cleanest verified downward-revision case on either list; the window converted it from an estimates thesis into a pipeline thesis. The company had guided 2026 to a full-year constant-currency sales DECLINE of 5-13% against 10% growth in 2025 and cut US GLP-1 prices by up to 70%, defending volume at the cost of mix. On Jul 31 ziltivekimab missed its ZEUS primary endpoint outright - biology worked, outcomes did not - against a market Jefferies sizes above $10B annually. UBS modelled only 1.7% of NPV but explicitly expected a 2-3x reaction because the failure compounds a run of R&D setbacks; Mizuho called it a disappointment that could derail sentiment on the whole non-obesity pipeline. Consensus PT already sat below spot before this. Reports Q2 Aug 5. A trailing P/E near 12 is genuinely cheap and limits downside velocity; HERMES (heart failure) and ARTEMIS (post-MI) both continue, reading out H1 2027 in distinct populations - a single failed trial is not a failed molecule; 2026 adjusted operating profit guidance was explicitly maintained; the Q3 impairment is non-cash; Down roughly 25% over twelve months plus ~9% on this news - a meaningful portion is now priced; The Aug 5 print is a live binary inside the horizon and could stabilise the story
77
AAPL↗Thesis confirmed on the tape. The prior framing was that the most valuable company in the world trades near 36.6x forward for ~11% growth with a consensus target BELOW spot - a rare mechanically bearish configuration. The FQ3 print delivered exactly the predicted shape: a genuine beat (revenue $109.4B +16%, EPS $2.02 +29%) undermined by a forward guide of 9-11% revenue growth against 12% consensus, with Services at $30.74B and Greater China both short. Management attributed the guide to advanced-chip and MEMORY supply constraints - meaning Apple is now a price-taker to the same shortage that is enriching Micron and Samsung, and has already raised Mac and iPad prices with iPhone expected to follow. Barclays, Goldman and Morgan Stanley all cut targets within hours. ↔ deep-dive$123.3B of TTM free cash flow and the largest buyback in history put a hard floor under it; Wells Fargo RAISED its target to $350 from $310 and BofA reiterated $380 on the same print - the sell-side is not uniformly on this side; Supply constraints are a timing problem, not a demand problem; management was explicit that demand is not the issue; A CEO transition to John Ternus on Sep 1 is a plausible narrative reset; Roughly $350-500B of value has already come out in one session - much of the easy move is done
76
CRWV↗$35B of debt against roughly $40B of equity value, funding depreciating GPUs, with -$10.6B of free cash flow. Altman Z-Score 0.36 and Piotroski F-Score 2 both sit in elevated-distress territory. The balance-sheet leg of this thesis is intact and is not affected by anything that happened this week. What HAS changed is the demand leg: this short was partly a bet that AI infrastructure spending would be cut, and two hyperscalers just told the market the opposite with numbers attached. The position is retained under the no-churn rule but the burden of proof has moved. Contracted revenue with investment-grade counterparties is real and materially de-risks the near term; ~99% 3-year revenue growth forecast; Nvidia's strategic support is an explicit backstop; Consensus PT of $138.03 is far above spot and borrow cost is likely punitive; WATCH_FOR_INVALIDATION, now on its second consecutive run: if the credit leg does not produce a financing event within the next quarter while capex keeps rising, this should be retired rather than carried
73
TSLA↗$1.22 trillion of market capitalization resting on $3.81B of earnings. The underlying auto business is a ~14%-growth, 3.7%-net-margin manufacturer; the valuation is optionality on robotaxi and Optimus with no verifiable revenue line. This is the largest pure discount-rate exposure of any US mega-cap, and the discount rate moved decisively against it this window: the long end printed multi-decade highs on both sessions even as equities rallied. TSLA remained a top-three active option line on both days without producing a directional print. ESTIMATE-REVISION DIRECTION IS UNVERIFIED - the sub-score is neutral 50 rather than guessed; Persistent retail bid; TSLA is a top-three active option every session in this window; Any credible autonomy milestone re-rates it violently; Consensus PT of $398.30 sits well above spot - the Street is not on this side; Rates being high is a necessary but not sufficient condition; the market has ignored this before
72
MRKMerck rallied 57% INTO the Keytruda loss of exclusivity, and the deterioration is visible in aggregate estimates: FactSet attributed the Health Care sector's Q2 downward revision specifically to Merck and Gilead. The company announced $3B of annual cost cuts ahead of the cliff - a defensive signal. Paying roughly 21x forward for 5.6% EPS growth against a patent cliff is the worst growth-adjusted multiple among mega-cap pharma. Reports Aug 4 BMO inside the horizon, with weekly implied vol printing above its 52-week range. Pre-print options skew is 2:1 toward CALLS - the market is positioned for a beat, not a miss; Subcutaneous Keytruda conversion could materially blunt the cliff; a KEYTRUDA QLEX sNDA has a PDUFA date of Aug 17; A 2.6% dividend limits downside velocity; Pipeline or M&A can re-rate it - the company is actively acquisitive; Consensus PT is still above spot
72
INTC↗A 4.5x move in twelve months on a company still losing $11B a year, driven by policy and partnership headlines rather than profits - $8.9B of CHIPS funding converted to a government equity stake, a $5B Nvidia investment, an Apple chip-making agreement. At roughly 54.8x forward it remains the most expensive large-cap semi on either list with the weakest competitive position and no trailing earnings. But the window produced the first genuinely uncomfortable datapoint for this thesis: Intel beat AND raised, and the market paid 12.53% for it. Intel also topped the free-tier options-activity lists on both sessions, so the crowding is now two-sided rather than one-sided. THESIS UNDER PRESSURE: a beat-and-raise that the market actually pays for is the specific evidence this short cannot absorb repeatedly; A sovereign/strategic backstop makes a fundamental short genuinely dangerous - the US government is an equity holder; 18A/14A yield news is binary and headline-driven; Server CPU demand is genuinely strong; Consensus PT sits well ABOVE spot - the Street is not on this side of the trade; WATCH_FOR_INVALIDATION: one more rewarded beat should retire this position rather than re-score it
72
WDAYThe cleanest pure-play on the seat-based-SaaS de-rating. HCM headcount licensing is the most literally AI-exposed pricing model in enterprise software: if AI reduces headcount, billable units shrink mechanically. Morgan Stanley initiated at Underweight with a $145 target on 2026-07-21 - below spot - while conceding Workday has the strongest moat in enterprise software. 15x forward is not expensive; Consensus is still Buy across 41 analysts with a PT above spot; Morgan Stanley itself concedes the moat, and the same firm published in March 2026 that AI-disruption fears in software are overblown - that internal inconsistency is why this cohort ranks low rather than high; No catalyst inside the horizon and no new evidence this run - a stale position by construction
66
PLTR↗Still trades at roughly 56x SALES after a 22% decline. The balance sheet is pristine and the growth is real, so this is explicitly NOT a business-quality short - it is a pure multiple-compression short in a regime where the long end just printed a 19-year high. At 56x sales a 53% grower needs roughly five years of flawless execution just to reach a normal software multiple. Reports Aug 3 AMC, the marquee single-name event of that session and the first live test of this position inside the horizon. retail-heavyThe options crowd is positioned 2.4:1 long calls into the print with weekly IV above its 52-week range - short-term risk/reward is against the short even if the 12-month thesis holds; Government and defense budget tailwind ($1.5T FY27 request); Crowded short; consensus PT well above spot; A beat-and-raise on Aug 3 in a tape that has started paying for beats again (LRCX, INTC, AMZN, MSFT) would be a genuine problem for this position
66
AMD↗AMD trades at roughly 2.5x Nvidia's forward multiple while being the structurally weaker accelerator franchise - squeezed from above by Nvidia's ecosystem and from below by Broadcom, Google TPU and Meta MTIA custom silicon. It faces the same memory-cost inflation now hitting Nvidia's bill of materials WITHOUT Nvidia's pricing power, and Samsung's shortage-through-2028 guidance makes that input problem structurally worse rather than transitory. This remains a relative-value short against NVDA far more than an absolute one. Rallied 13% inside the window - the entry is materially worse than it was; Aug 4 AMC with weekly IV above its 52-week range is a live binary inside the horizon; Genuine second-source demand from hyperscalers wanting a Nvidia alternative is real and structural; Server CPU share gains against Intel are real and margin-accretive; A 46.6% revenue growth forecast is not fantasy
65
C · Options Heat

Where the premium is paying up

TickerSignalReadImplicationLag
AMZN↗iv_spikeJul-31 weekly 227 straddle priced for a 7.5% move; call:put 1.6:1. Realized close-to-close +15.63%The pre-print options market underpriced the outcome by ~2.1x. Combined with MSFT (~2.2x) and AAPL (~1.4x), the free-tier implied moves systematically understated mega-cap earnings dispersion this quarter. Treat published implied moves as a floor, not a bound.Market Rebellion pre-market report Jul 30 06:03 ET, published BEFORE the event. Broker-desk snapshot, not exchange data. No post-print vol/OI: Barchart Unusual Options Activity and Market Chameleon both render client-side and return page chrome only.
MSFT↗iv_spikeJul-31 weekly call IV 111 vs August 48 (2.3x term structure); call:put 2.2:1; implied ~7%. Realized +15.5%Rich event premium that still underpriced the move by ~2.2x. The 2.2:1 call skew was directionally right for once - but note the identical skew on META (2:1) and RBLX (2.5:1) was catastrophically wrong on the same night.Market Rebellion mid-session Jul 29, intraday and pre-event. Volume ratios do NOT identify buy vs sell side.
AAPL↗iv_spikeJul-31 weekly 340 straddle priced for a 5% move; call:put 1.2:1 (balanced). Realized -7.13%The smallest relative event premium of the mega-cap slate against the largest absolute dollar exposure - the options market was not pricing an Apple surprise, and got one. Balanced skew means nobody was positioned for direction either.Market Rebellion pre-market Jul 30, pre-event broker-desk snapshot.
RBLXpc_skewJul-31 weekly 50 straddle priced for a 14% move - widest of the Jul 30 AMC slate - with call:put 2.5:1 toward CALLSBearish outcome against bullish positioning. The most decisively wrong-footed book in the window: heaviest call skew of the slate into a print that withdrew full-year guidance. A 2.5:1 call ratio is equally consistent with call buying and with overwriting, so this is evidence about DISPERSION, not about who was long.Market Rebellion pre-market Jul 30, pre-event. No post-print flow obtainable and no close-basis price move to measure against.
COINiv_spikeJul-31 weekly 160 straddle priced for a 9% move; call:put 1.3:1. Realized approximately -10%The ONLY name in this window where the free-tier implied move correctly sized the outcome. Worth recording as a base-rate datapoint: the options market priced the crypto-beta name right and the three mega-caps wrong.Market Rebellion pre-market Jul 30, pre-event. The realized figure is itself approximate (post-close sources cluster near -10% against premarket and after-hours prints of -4.5% and -6.53%).
PLTR↗iv_spikeAug-7 weekly call IV 103 vs August 74, against a 52-week range of 41 to 75 - weekly IV ABOVE the annual band; call:put 2.4:1 to CALLS, focused on Jul-31 weekly 123 and 126 strikesBullish positioning at record implied vol into the Aug 3 AMC print. Directly relevant to Track B downside rank 8: the crowd is long calls into the event, which worsens near-term risk/reward for that short without changing the 12-month multiple argument.Market Rebellion mid-session Jul 31, 11:27 ET - intraday and non-final. Pre-event.
SPCXiv_spikeAug-7 weekly call IV 156 and August 122, against a 52-week range of 71 to 116 - BOTH tenors above the annual high; call:put 2.1:1 callsThe most extreme volatility setup in the horizon. First-ever public earnings report Aug 4 AMC, followed on Aug 6 by the first insider lock-up tranche (up to 911.5M shares). Context: the stock hit an intraday all-time low of $107.01 on Jul 28, roughly 50% below its peak, with market cap reported near $1.43T.Market Rebellion mid-session Jul 31, intraday. Market-cap figure is from a Tier-3 source and conflicts with earlier $1.5-2.0T estimates - do not size off it.
AMD↗iv_spikeAug-7 weekly call IV 100 vs August 82, against a 52-week range of 39 to 89 - weekly IV above the annual band; call:put 1.2:1 (near balanced)Record implied vol with essentially no skew into the Aug 4 AMC print - a volatility event, not a direction. Relevant to Track B downside rank 9.Market Rebellion mid-session Jul 31, intraday, pre-event.
MCDvol_oiAug-7 weekly call IV 36 vs August 28; call:put 13:1 toward CALLS with a focus on 17,000 contracts of the Jul-31 weekly 275 calls into the Aug 4 BMO printThe most extreme single-name call skew observed in this window by a wide margin, in a defensive mega-cap with modest implied vol. Included because a 13:1 ratio in a low-vol staple is anomalous enough to record - but a same-week expiry concentration is equally consistent with overwriting against stock, and no side tagging exists to distinguish them.Market Rebellion mid-session Jul 31, intraday. No bid/ask side tagging on any free source.
MARpc_skewAug-7 weekly call IV 50 vs August 37; call:put 1:3.1 toward PUTS with a focus on Aug-7 weekly puts into the Aug 3 BMO printBearish. The most put-skewed name in the entire forward earnings set and the clearest single-name downside positioning in the horizon - a consumer-discretionary read worth watching alongside the Carvana/Roblox/GoDaddy consumer weakness in Track A.Market Rebellion mid-session Jul 31, intraday, pre-event.
MRKpc_skewAug-7 weekly call IV 52 vs August 37, against a 52-week range of 22 to 36 - above the annual band; call:put 2:1 toward CALLS into Aug 4 BMOPositioned AGAINST Track B downside rank 5. Recorded as evidence cutting against an existing position rather than filtered out.Market Rebellion mid-session Jul 31, intraday, pre-event.
CATiv_spikeAug-7 weekly call IV 74 vs August 59, against a 52-week range of 22 to 55 - well above the annual band; call:put 1:1.1 (marginally puts) into Aug 4 BMOEvent premium at a 52-week extreme with no meaningful skew. CAT's Aug 4 guide is the tariff and cost-inflation tell for the whole industrial complex.Market Rebellion mid-session Jul 31, intraday, pre-event.
VRTpc_skew30-day IV 73 vs a 52-week range of 45-85; call:put 1:1.4 toward PUTS on 104,000 contractsBearish, and directly relevant to Track B upside rank 10 - the options tape is positioned against that long. This is why VRT's sentiment sub-score was raised only to 50 despite the hyperscaler capex reconfirmation.Market Rebellion pre-market Jul 30, reflecting the Jul 29 close. One session stale relative to RUN_DATE.
AMC_CNK_IMAXiv_spikeAMC 30-day IV 150 vs a 52-week range of 53-143 (above the annual high) with call:put 4.4:1 focused on August 3.50 calls; CNK 3.2:1 calls; IMAX 1:2.1 PUTSA coordinated speculative bid in film exhibition with share prices near the top of their range - the clearest PUMP-RISK-shaped flow signature in the window. AMC does not pass the $10B universe filter and is not ranked; recorded here because the pattern (record IV plus 4.4:1 calls plus no Tier-1 catalyst) is exactly what the pump-risk rule is written to catch. IMAX's opposite skew makes the group read as retail-driven rather than sector-fundamental.Market Rebellion mid-session Jul 31, intraday. ETF/sub-threshold names excluded from ranked lists by universe filter.
GDDYvol_oiHeaded the 'increasing unusual PUT option volume' list on the Jul 31 mid-session tapeBearish, and the only independent corroboration available for the GoDaddy decline whose magnitude could not be reconciled across sources. Recorded specifically to show that the name was excluded from the ranked list for a DATA reason, not for lack of signal.Market Rebellion mid-session Jul 31, intraday. Directional list membership only - no IV level, skew ratio or contract count published.
CROWDING_WATCHvol_oiNames appearing on Market Rebellion's 'popular stocks with increasing option volume' AND 'active options' lists on BOTH Jul 30 and Jul 31: INTC, MU, SPCX, MSTR, SNDK, NOK. Jul 31 adds COIN, AVGO, HOODPersistent two-session crowding is concentrated in memory (MU, SNDK), Intel, and high-variance retail names (SPCX, MSTR, NOK, HOOD) - NOT in the mega-caps that actually made the index moves. That divergence is the cleanest available read that the retail options bid and the institutional equity bid were in different places this week.Market Rebellion pre-market Jul 30 and mid-session Jul 31. These are unranked membership lists with no volumes attached - useful for crowding, useless for direction.
D · Catalyst Calendar

Upcoming events

DateEventTickersIV move
2026-08-03ISM Manufacturing PMI July (cons. 53.3) 10:00 ET; Construction Spending June; US Treasury quarterly refunding marketable borrowing estimates plus TBAC statements. FIRST market reaction to the Aug 2 (Sunday) OPEC+ ministerial setting September output - note the reported DIRECTION of the prior OPEC+ decision remains UNVERIFIED across sources.macrounknown
2026-08-03Q2 earnings, AMC. Marquee single-name report of the day. PLTR is Track B downside rank 8 with Aug-7 weekly call IV at 103 against a 52-week range of 41-75 and a 2.4:1 call skew - the crowd is long into it.PLTR↗high
2026-08-03Large-cap earnings (mixed BMO/AMC). MAR carries the most put-skewed book in the forward set at 1:3.1 puts.MARVRTXWMBOKEFANGONMELITSNMUFGJAZZCLXnormal
2026-08-04International trade in goods and services June (balance cons. -$77.6B) 8:30 ET; JOLTS job openings June (cons. 7.594M) 10:00 ET - date verified against the BLS JOLTS schedule after an aggregator conflict; Factory Orders June.macrounknown
2026-08-04SpaceX FIRST-EVER earnings report as a public company, AMC. Aug-7 weekly call IV 156 and August 122 against a 52-week range of 71-116 - both tenors above the annual high, 2.1:1 calls. Triggers a tiered insider lock-up whose first tranche opens Aug 6 (up to 911.5M shares). Stock printed an intraday all-time low of $107.01 on Jul 28.SPCXhigh
2026-08-04Q2 earnings, AMC. Track B downside rank 9 and Track A rising rank 6 - the two tracks disagree by design. Aug-7 weekly IV 100 vs a 52-week top of 89, call:put 1.2:1.AMD↗high
2026-08-04Large-cap earnings, BMO. MRK is Track B downside rank 5 with a 2:1 CALL skew against that position. MCD shows a 13:1 call ratio on 17K Jul-31 weekly 275 calls. CAT's guide is the industrial cost-inflation tell.CATMRKMCDPFEBPDUKSPOTCMIMPCAPOGWWnormal
2026-08-04Large-cap earnings, AMC. AMGN carries a 3.8:1 call skew with Aug-7 weekly IV 55 vs August 39. ANET and ALAB are the AI-networking read-through from the AWS print.ANETAMGNGILDBKNGEMREOGSUDVNALABPINSLSCCTOSTWYNNnormal
2026-08-05ADP National Employment Report July 8:15 ET - the leading tell for the Aug 7 payrolls print; ISM Services PMI July (cons. 54.0) 10:00 ET; US Treasury quarterly refunding statement plus TBAC report. With the 30-year at a 19-year high, the refunding statement is a genuine rates event this quarter rather than a formality.macrounknown
2026-08-05Q2 earnings - LLY BMO (Track B upside rank 7), NVO (Track B downside rank 3, reporting four sessions after the ZEUS Phase 3 failure). The two sides of the GLP-1 pair report on the SAME day, which makes this the single highest-information session of the horizon for Track B.LLY↗NVOnormal
2026-08-05Earnings, AMC. SNDK is Track A falling-list alumnus and current rising rank 5 with PUMP_RISK flagged; this print resolves whether the +23.7% Jul 30 move was a re-rate or a squeeze. Sandisk also has an Investor Day on Aug 13.SNDKWDChigh
2026-08-05PDUFA goal date for mRNA-1010 seasonal influenza vaccine. FDA VRBPAC voted 9-0 in favour on 2026-06-18 (separate 9-0 votes for adults 50-64 and 65+).MRNAhigh
2026-08-05Large-cap earnings (mixed BMO/AMC). UBER is on the deep-dive watchlist and did NOT surface organically in either track this run.DIS↗SHOPCVSUBER↗PSXCORKHCGFSAPPMCKDASHOXYEBAYXYZAXONEXPEHUBSCRCLIONQnormal
2026-08-06SpaceX first insider lock-up tranche opens - up to 911.5M shares - 48 hours after the maiden earnings report. A supply cliff and a potential squeeze setup resolving within two sessions of each other. Also: COP earnings BMO and Q2 productivity and costs.SPCXhigh
2026-08-07BLS Employment Situation - July payrolls, 8:30 ET. Verified against the BLS release schedule. This is the dominant macro event of the horizon: with three FOMC dissents favouring a HIKE, the 10-year at a 19-month high and the 30-year at a 19-year high, front-end index volatility should be pinned here rather than to any earlier date. The Aug 4 JOLTS and Aug 5 ADP prints are the read-throughs.macrounknown
E · Trade Ideas

What the data implies

NVOLLY↗short · track B
Rationale The highest-information pair in the artifact, and both legs report on the SAME day. Novo's ZEUS Phase 3 failure (hazard ratio 0.99, 95% CI 0.88-1.11, >6,300 patients) is a Tier-1 company disclosure that closes off cardiovascular as a second growth pillar for a business with roughly 94% sales reliance on one franchise, on top of company-guided 2026 constant-currency sales DECLINE of 5-13% and US GLP-1 price cuts of up to 70%. UBS modelled only 1.7% of NPV but expected a 2-3x reaction because the failure compounds a run of R&D setbacks. Lilly is compounding ~29% EPS into the same end market. Expressed as a PAIR, this isolates the share-shift thesis and neutralises GLP-1 category risk - which matters because the category itself is not the trade.
Invalidation SHORT leg invalidated by a volume inflection from the price cuts showing up in Aug 5 quarterly scripts, strong oral-formulation data, or positive HERMES/ARTEMIS signalling. LONG leg invalidated by an Aug 5 Lilly manufacturing or guidance stumble - at ~34x forward there is no room for one. Note the pair is NOT hedged against a shared regulatory shock (US most-favored-nation drug pricing hits both).
Risk A trailing P/E near 12 on NVO is genuinely cheap and limits downside velocity, and roughly 9% came out on Jul 31 alone. A rival's trial failure improves Lilly's RELATIVE position, not its absolute execution.
MUneutral · track A/B
Rationale Deliberately NEUTRAL for a third consecutive run, because the evidence got MORE bimodal this week rather than less, and saying so is more useful than picking. The bull leg strengthened materially: Samsung - the largest competitor, with every incentive to talk supply up - told analysts the memory shortage deepens in 2027 and persists through 2028, which is the strongest available rebuttal to the CXMT oversupply thesis and it came from an adversarial source. Korea confirmed it with SK Hynix +30% and Samsung +28%. Apple then independently corroborated the shortage from the demand side by blaming its own FQ4 guide on memory supply. The bear leg is undiminished: Michael Burry publicly INCREASED a short near $880 into the bounce, an ~85% gross margin is roughly 25 points above any sustainable level, customer inventories sit at 7-9 weeks against an 8-week warning threshold, and the stock needed a competitor's earnings call to move.
Invalidation LONG invalidated if customer inventories cross 10 weeks (the stated exit trigger) or if FQ4 gross margin guidance is cut. SHORT invalidated if HBM contract pricing holds into the Q3 renewal round. Watch SNDK and WDC on Aug 5 - the closest available proxy print inside the horizon.
Risk Highest-variance name in this artifact. Market cap and the Jul 31 close both carry unresolved source conflicts - size off verified fundamentals, not off a quoted price. Track A rising rank 2 and Track B upside rank 5 are scored independently; do not net them.
AAPL↗short · track A/B
Rationale The clearest thesis-confirmation event in the window. Apple beat on the quarter and fell 7.13% because it guided FQ4 revenue growth to 9-11% against 12% consensus, with Services and Greater China both short. The specific mechanism matters more than the miss: management attributed the guide to advanced-chip and MEMORY supply constraints, which makes Apple a structural price-taker to the shortage that is simultaneously enriching Micron and Samsung, and has already forced Mac and iPad price increases with iPhone expected to follow. Barclays, Goldman and Morgan Stanley all cut targets within hours. The pre-existing bearish configuration - roughly 36.6x forward for ~11% growth with a consensus target BELOW spot - now has a company-supplied catalyst behind it.
Invalidation Supply constraints easing faster than guided (this is a timing problem, not a demand problem, and management said so explicitly); any credible Siri/AI launch closing the narrative gap; a compelling iPhone upgrade cycle. Also note the CEO transition to John Ternus on Sep 1 is a plausible narrative reset with no fundamental content.
Risk $123.3B of TTM free cash flow and the largest buyback in history put a hard floor under this. Wells Fargo RAISED its target to $350 and BofA reiterated $380 on the same print - the sell-side is genuinely split, and shorting the most-owned stock in the world into a $500B one-day drawdown is a crowded entry.
SPCXneutral · track A
Rationale The most asymmetric scheduled event in the horizon and the one with the least usable prior. First-ever public earnings Aug 4 AMC with Aug-7 weekly call IV at 156 and August at 122 against a 52-week range of 71-116 - BOTH tenors above the annual high, which is unusual enough to note on its own - and a 2.1:1 call skew. Forty-eight hours later the first insider lock-up tranche opens, releasing up to 911.5M shares. The stock printed an intraday all-time low of $107.01 on Jul 28, roughly 50% below its peak. Two opposing forces (a squeeze setup and a supply cliff) resolve within two sessions of each other, against a company with no earnings history whatsoever.
Invalidation n/a (volatility view). A directional read only becomes tradeable AFTER Aug 4 clarifies the lock-up tranche mechanics.
Risk No earnings history means no base rate for how it reacts. Market cap figures conflict badly across sources ($1.43T Tier-3 vs earlier $1.5-2.0T estimates) - do not size off any of them.
PLTR↗short · track B
Rationale Position carried, but flagged as facing its first live test in three sessions with positioning against it. Still roughly 56x SALES after a 22% decline - a pure multiple-compression short in a regime where the 30-year just printed a 19-year high. This is explicitly NOT a business-quality short: the balance sheet is pristine and 53% growth is real. The problem is that at 56x sales the company needs roughly five years of flawless execution just to reach a normal software multiple.
Invalidation An Aug 3 beat-and-raise in a tape that has demonstrably STARTED PAYING for beats again (LRCX +18%, INTC +12.5%, MSFT +15.5%, AMZN +15.6% all inside this window) would be genuine evidence against the position, not noise. That regime shift is the single most important thing that changed this week for every short in this book.
Risk Options are 2.4:1 call-skewed into the print at record implied vol. Crowded short, consensus PT far above spot, and a $1.5T FY27 defense budget request behind the business.
INTC↗CRWV↗neutral · track B
Rationale Framed as a POSITION-REVIEW note, not a trade. Two Track B downside positions took direct evidentiary damage this window and both are now flagged watch_for_invalidation. Intel beat AND raised full-year guidance and the market paid it 12.53% in a single session - estimates moving up is the specific evidence a valuation short cannot absorb repeatedly. CoreWeave's thesis has a balance-sheet leg (intact: $35B debt, -$10.6B FCF, Altman Z 0.36) and a demand leg (damaged: four hyperscalers just guided to $720-745B of 2026 capex with Amazon giving an explicit ROIC walk-through). Under the no-churn rule neither is removed on two days of evidence - but the burden of proof has moved and this is recorded rather than quietly re-scored.
Invalidation INTC: one more rewarded beat retires the position. CRWV: if the credit leg produces no financing event within the next quarter while capex keeps rising, retire rather than carry.
Risk The failure mode this note exists to prevent is a book that only ever adds confirming evidence. Both of these cut the other way and both are staying visible until they resolve.

Methodology & limitations

  • THE DEFINING PROBLEM OF THIS RUN: premarket figures masquerading as closes. TheStreet's July 31 live blog publishes its mover percentages inside an 08:11-08:36 ET 'Premarket Movers' block, and search snippets strip that context entirely - the same figures surfaced in search results labelled 'late morning' and 'on the downside'. Direct fetch of the article established the timestamps. RBLX -22.5%, GDDY -16.4%, MPWR +11.4%, MTZ -16%, NWL +17.5% and AMZN +11% are ALL premarket and none was used as a close. This is a generalisation of the known TheStreet intraday failure mode and should be screened for by fetching the article rather than trusting snippets.
  • NO LIVE OPTIONS TAPE. Barchart Unusual Options Activity and Market Chameleon render client-side and return page chrome only. Every options figure in this artifact is a Market Rebellion broker-desk snapshot, published either pre-market (06:03 ET, reflecting the prior close) or mid-session (11:27 ET, intraday and non-final). No post-print vol/OI table for either session was retrievable from any free source.
  • NO BID/ASK SIDE TAGGING on any free source. Every call:put figure here is a VOLUME ratio. This window supplied three separate demonstrations of why that matters: RBLX (2.5:1 calls, fell hard), META (2:1 calls, fell 7.95%), and MCD (13:1 calls in a low-vol staple, equally consistent with overwriting).
  • NO NAME-LEVEL OPTIONS DATA for six of nine Track A rising names and five of eight falling names in the window itself. Where a name appeared only in an unranked activity list, the sub-score reflects crowding evidence and says so; where nothing was obtainable, the sub-score is neutral 50 and flagged (CVNA, UNH).
  • NO CBOE PUT/CALL RATIO obtained for either session. The official CBOE daily statistics page has returned navigation-only content across multiple runs and was not retried. No index-level hedging read is claimed.
  • MARKET CAPS ARE ONE SESSION STALE. The stockanalysis basis is 2026-07-30 for every Track B name and most Track A names, and Jul 31 moved AAPL -7.13%, AMZN +15.63%, GOOGL +7.12% and NVO ~-8.8%. Figures are flagged mcap_as_of_prior_close rather than re-estimated. RBLX ($34.85B) is the only Jul-31-dated cap in the artifact.
  • UNRESOLVED SOURCE CONFLICTS, published as ranges or flagged rather than point-estimated: the 30-year yield (5.25% CNBC vs 5.28% StreetStats); WTI (a $84.23 premarket print vs $86.80 on the Sep-26 contract near the close); AAPL's move (-7.13% close-basis vs -9.5% and ~-10% intraday) and its post-move market cap ($4.90T Jul-30 basis vs a $4.4T Tier-3 figure); GOOGL's move (+7.12% / +6.9% / +6.7% / +3.12% across four sources); META's Jul 30 move (-7.95% CappNotes vs 'more than 9%' CNBC vs -9.64% Tier 3); SNDK's Jul 30 move (+23.70% vs +26%); NVO's move (-7.42% Copenhagen vs -8.78% to -9.5% ADR); COIN's move (-4.5% premarket, -6.53% after-hours, ~-10% post-close, -14% intraday); the KOSPI level (6,695.45 AP vs 6,595.45 Tier 3); and July's Dow return (+0.7% post-close vs -1.01% from an intraday source, the latter discarded).
  • GDDY AND RBLX ARE THE TWO HONESTY TESTS OF THIS RUN. GDDY was EXCLUDED from the ranked list because three sources give three incompatible figures and all three are pre-open, even though its direction is independently corroborated by unusual put volume. RBLX was RANKED but with move_pct set to NULL, because its market cap is verified at the Jul 31 close while its price move is not. Different treatments for different failure shapes.
  • VIX CLOSE IS SEMI-VERIFIED. The 15.99 (-6.44%) figure comes from a Yahoo Finance sidebar widget carrying a 14:59 ET timestamp on the ^TNX page, and reconciles arithmetically against a 17.09 prior close. VIX settles at 16:15 ET, so this may be a near-close rather than a settled print. Yahoo sidebar widgets have a documented caching failure mode in this pipeline and this figure was accepted only because the arithmetic reconciles.
  • RUSSELL 2000 verified by arithmetic reconciliation: 2,946.10 (Jul 30 confirmed) - 14.76 = 2,931.34, matching the Yahoo sidebar exactly. TheStreet's '+1.37%' for the Russell was a 09:11 ET premarket figure and is discarded; an intraday source at 12:01 ET had it -0.91%.
  • SECTOR-LEVEL AND BREADTH DATA is thinner than prior runs. No advance/decline line, no sector performance table and no volume figure was obtained for either session. The breadth characterisation rests on two intraday observations (five Dow megacaps contributing ~114% of the index gain at 11:56 ET on Jul 31; 243 S&P gainers shortly after the open on Jul 30) and on the Russell 2000's confirmed negative close against positive large-cap indices. Stated as such rather than dressed up.
  • TRACK B SUBSCORE BASELINE. The 2026-07-30 artifact is not present in this context; only its Track B membership list and its CRWV/META watch_for_invalidation escalations were recoverable. Subscore deltas are therefore applied to the 2026-07-29 artifact, which is available verbatim as a project file, with both the Jul 30 and Jul 31 evidence incorporated. Every changed sub-score carries a subscore_delta string naming the evidence; every unchanged one says so explicitly.
  • BLOCKED OR UNAVAILABLE SOURCES: Bloomberg and CNBC return 403 to automated fetch (CNBC content here comes via search index); Globe and Mail and Seeking Alpha are robots-disallowed; FRED VIXCLS was stale to 2026-07-29; the CappNotes Closing Look for 7/31 was not indexed at build time. Zacks/Yahoo 'Stock Market News for [date]' pages were not consulted at all - this pipeline has now confirmed across three consecutive runs that they recap the PRIOR session.
  • SCORES ARE HEURISTIC AND NOT BACKTESTED. Sub-scores are analyst judgment applied to cited evidence, not a fitted model. Track A options sub-scores rest on pre-event straddle pricing and unranked activity lists, which measure expectation and crowding - not realised institutional flow.

Not financial advice. Educational/analytical only — generated by Daily Market Pulse template v1 - multi-desk research run (market regime and rates, Track A movers across two sessions, options and crowding, forward catalysts, Track B conviction review), synthesized 2026-08-01. Do your own diligence.

Built with curiosity and care

by