Market Pulse · Daily Scan
Daily Market Pulse
Jul 20, 2026 · sentiment/flow window ≈ 48h
Pre-marketnot publishedPost-market
TLDR;
A failed rally, not a rout. Equities opened sharply higher — S&P +0.6%, Nasdaq +1%, the PHLX Semiconductor index up ~3% back above 12,000 — as memory (SNDK/MU/SK Hynix) and the AI complex (AMD/AVGO/MRVL/NVDA) mean-reverted off a bear-market low, and as Iran's foreign ministry signalled that mediated talks could be pursued. The bid faded through the afternoon: Trump posted that Iran would pay 'many times over' for three American deaths, Yemen's Houthis declared a maritime embargo on Saudi Arabia, WTI closed +0.9% at $83.23 and Brent +1.3% at $89.22, and the 10-year rose ~5bp to ~4.59%. Apple (−2 to −3%, off Friday's record) and Tesla were the two biggest S&P point decliners and single-handedly held the index red against a rising chip complex. Net: SOX +0.6% — its first up-day in six after a −10% week, the worst since April 2025 — but the cap-weighted indices still closed lower. Breadth: Negative and narrowing. Roughly two-thirds of US stocks declined (TheStreet); Russell 2000 −0.67%, underperforming for a second straight session; NYSE Composite −0.62%. The advance was concentrated in semiconductors and AI-datacenter names while defensives (Health Care −1.15%, utilities, REITs, staples) gave back recent gains — an unusual configuration: the AI trade bounced and the hiding places sold, yet the index still fell because the two largest single names went down..A · Daily Pulse
What moved — and why
▲ Rising — dailycomposite
| # | Ticker | Move | Driver | Score |
|---|---|---|---|---|
| 1 | IREN | +18.55% | Signed ~$2.8B of new multi-year AI Cloud contracts (Microsoft, Nvidia and Perplexity named as customers) and lifted its ARR target above $4B. Day's cleanest large-cap catalyst. PUMP-RISK ASSESSMENT: a vertical move in a former-bitcoin-miner/AI-datacenter name with a heavy retail following would normally raise the flag, but a specific, dated, company-announced contract catalyst is present — so the flag is NOT raised. Caveated instead: the named counterparties have not independently confirmed, and a contract-announcement gap is a non-repeatable event, not persistent momentum. retail-heavy | 81 |
| 2 | AMD | +1.6% | Closed +1.6% on an announced expansion of its strategic partnership with Microsoft spanning AMD GPUs, CPUs, networking and software on Azure — a confirmed same-day company catalyst, unlike most of the chip complex, which rose on mean-reversion alone. | 64 |
| 3 | MU | +2% | Climbed almost 2% as a key leader of the memory bounce (SNDK/MU/SK Hynix led the early tape) after five straight down-days. No fresh single-name catalyst — this is mean-reversion off an oversold chip complex, not new information. Also a Track B upside carry; appearing on both tracks is legitimate and disclosed. ↔ deep-dive | 60 |
| 4 | TER | +3.5% | Best confirmed percentage gain in the semi-test/equipment complex at +3.5%. No confirmed single-name catalyst — attributable to the sector bounce (SOX +0.6% after entering a bear market last week), so the catalyst sub-score is held low rather than credited. unverified | 58 |
| 5 | ALAB | +1.9% | +1.9% on the AI-connectivity bounce alongside the broader semi recovery attempt. No confirmed single-name catalyst. unverified | 54 |
| 6 | NVDA | — | DIRECTION CONFIRMED, MAGNITUDE NOT. Bloomberg headlined the session on Nvidia and other chipmakers rising, and the AP wrap noted Nvidia and Sandisk 'steadied' after last week's tumble — but no free-tier source gave a confirmed closing percentage, so move_pct is left null rather than estimated. Lost the world's-most-valuable-company title to Apple intraday Friday, then Apple gave it back Monday by falling harder. ↔ deep-diveprint unverified | 54 |
▼ Falling — dailycomposite
| # | Ticker | Move | Driver | Score |
|---|---|---|---|---|
| 1 | WBD | −3.76% | Closed at $25.86, −3.76%, after a California federal judge granted a temporary restraining order halting Paramount Skydance's $110B acquisition for 14 days, following an antitrust suit by 12 state attorneys general. The order cited the states' evidence on wide-release theatrical distribution share, despite prior DOJ clearance. A preliminary-injunction hearing is set for August 3; a delay past September 30 triggers a 25-cent-per-share quarterly ticking fee to WBD holders. Volume 44.2M vs a 20.6M three-month average (+115%). SOURCE DISCREPANCY: an intraday feed printed −3.4% at $25.95 and TheStreet's live blog showed −1.7%; the −3.76% / $25.86 close is used and the intraday figures discarded. | 77 |
| 2 | TSLA | — | DIRECTION CONFIRMED, MAGNITUDE NOT. Bloomberg identified Apple and Tesla as the two biggest point decliners in the S&P 500, which alone held the index in the red against a chip-led advance — but no free-tier source gave a confirmed closing percentage, so move_pct is null. Ranks high on catalyst rather than momentum: the Jul 22 after-close print is the single largest scheduled binary in the horizon. Also the #1 Track B downside carry. print unverified | 68 |
| 3 | AAPL | −2.5% | The single largest drag on the Dow and, with Tesla, one of the two biggest S&P point decliners — profit-taking off Friday's record (~$335) ahead of its June-quarter report, closing back below $325. MAGNITUDE CONFLICT: CNBC reported 'more than 2%', FX Leaders 'nearly 3%'; −2.5% is recorded as the midpoint of a genuinely conflicting range and flagged, not asserted as a confirmed print. Context: Apple briefly overtook Nvidia as the world's most valuable company on Friday, and is up ~23% YTD versus Nvidia's ~7% — the crowd is paying for AI monetization without the capex, and Monday was that trade taking a breath, not breaking. | 64 |
| 4 | NFLX | −1.96% | Closed $67.60, −1.96% — a third leg lower after the Jul 16 guidance-driven de-rating (−7.26% on Jul 17), with merger headlines keeping the whole media complex in focus. No fresh single-name catalyst; this is continuation, not new information. | 51 |
| 5 | DIS | −1.26% | Closed $96.44, −1.26%, as the media and entertainment complex traded off the Warner Bros. Discovery injunction headlines. Sector-inferred driver — no confirmed Disney-specific catalyst on the day. | 46 |
B · One-Year Conviction
Asymmetric 12-month skew
Free sources only. Carried forward from the Jul 17 book. Market caps are Jul 16 stockanalysis.com snapshots unless noted — several sessions stale. Estimate-revision direction is inferred from analyst PT/rating actions and Zacks ranks, not full institutional revision breadth (paywalled). Directional, not precise. Not financial advice.
▲ Upside skew — 1yrconviction
| Ticker | Thesis | Key risk | Score |
|---|---|---|---|
| NVDA | AI-compute leader; consensus Strong Buy, avg PT ~$302 (KeyBanc $310→$330, Jul 14). Vera Rubin ramp into 2H26; ~50% of revenue now non-hyperscaler. Sentiment restored +2 from the Jul 17 trim: the SOX put in its first up-day in six (+0.6%) and the group led the early tape, so the Kimi-K3 commoditization scare has not compounded. Bounded, symmetric adjustment — not a thesis change. ↔ deep-dive | AI-capex digestion / hyperscaler capex pause; open-source model commoditization (Kimi K3) compressing compute demand; valuation; one stabilization session is not a trend | 82 |
| TSM | Advanced-node foundry monopoly. Jul 16 Q2 beat-and-raise: record revenue ~$40.2B (+~36% y/y), GM 67.7%, capex raised to $60–64B and Arizona to ~$100B. The capex-intensity read is exactly what pressured the group — strong results that the market punished. Barclays $470→$625. | Taiwan geopolitics; capex-cycle intensity compressing FCF; cyclicality; 'good results, bad stock' persists while capex fear dominates | 81 |
| AVGO | Custom AI accelerators (XPU) plus networking. Positive revisions (JPM $500→$580, BofA $450→$530). Named among the AI plays leading the Jul 20 early bounce; sentiment restored +2 alongside the other ai_capex_roi_risk names. | rotation out of semis; XPU revenue concentrated in a few hyperscaler customers; valuation | 75 |
| CEG | Nuclear power for AI data centers. Buy consensus, PT ~$358 (~43% implied upside). Walmart nuclear PPA (Jun 23), Calpine closed, prior Microsoft/Meta data-center deals. Structural power-scarcity long — materially less exposed to the semiconductor-commoditization fear than the chip names, which is why it stays in the book while the semi sleeve is caveated. | power-price/regulatory; data-center demand normalization; valuation; utilities gave up recent gains on Jul 20 as defensives sold | 70 |
| MU | HBM/AI-memory super-cycle; Strong Buy revisions (KeyBanc $1,600→$1,750, Jul 14). Record fiscal-Q3 revenue ~$41.46B, guided ~$50B for Q4. Sentiment restored +2: memory led the Jul 20 bounce and MU closed ~+2%, its first up-day in six. The parabolic-reversal and 2027-oversupply caveats stand — one green session does not repair a broken ~$897.80 support level. ↔ deep-dive | memory pricing cyclicality and 2027 oversupply (amplified by TSMC capex + China CXMT capacity); parabolic-reversal / momentum risk after a ~6x run; valuation | 69 |
| AMAT | Wafer-fab-equipment beneficiary of next-gen AI chips, DRAM and advanced packaging into 2H26 (Needham $530→$740). Upside modest — price sits near consensus PT (~$606), so the thesis is revision-dependent rather than valuation-driven. Sentiment restored +2 with the WFE complex bounce. | WFE cyclicality; China export controls; trades near PT — limited margin of safety | 68 |
| BWXT | Nuclear-component and SMR play. Deutsche upgraded Hold→Buy ($205→$255); record backlog ~$8.65B. Power-scarcity structural long. | federal-budget dependence; SMR timeline slippage; valuation; smallest cap in the upside book | 66 |
| MPWR | Power management for AI servers; Moderate Buy, PT ~$1,599 (KeyCorp $1,500→$2,000). Enterprise-data revenue growing rapidly. Reports ~Jul 30 — just outside the 5-day catalyst horizon but inside the next run's. | NVIDIA-platform concentration; high valuation; cyclicality | 65 |
| GEV | Grid/power equipment plus SMR (GE-Hitachi BWRX-300) levered to AI electricity demand. WEAKENING for a second consecutive run: Jefferies cut $1,350→$1,210, Zacks to Hold, PT ~$1,090 only ~5% above price. Reports Jul 22 — a binary inside the horizon that will decide whether this holds its slot. | trades ~at consensus PT; mixed/flattening revisions; Jul 22 earnings is binary; project execution | 62 |
| CSCO | AI-era networking (NVIDIA partner); BofA $135→$150, Zacks Hold→Strong Buy. The slower-growth, lower-multiple way to hold AI infrastructure — its defensive character is the point when the high-beta sleeve de-rates. Upside modest after re-rating (PT ~$123, ~13%). Reports ~Aug 12. | slower structural growth vs pure-play AI names; competitive networking pressure; limited upside to PT | 61 |
▼ Downside skew — 1yrconviction
| Ticker | Thesis | Key risk | Score |
|---|---|---|---|
| TSLA | Margin quality against a robotaxi-priced multiple. Analyst split (PT ~$425); four straight quarterly misses. Q2 deliveries are already known (~480k, +25%), so the Jul 22 AMC print is about margins, not volume. Mild confirmation on Jul 20: one of the two biggest S&P point decliners on a day the chip complex rose. | robotaxi/FSD monetization inflects and justifies the multiple; auto margins reaccelerate; Jul 22 earnings is a binary that can end this thesis in one session | 66 |
| CRWV | Debt-funded, customer-concentrated AI buildout — >7x leverage, with sell-side PTs (avg ~$141) still lagging a price collapse to ~$73. HEADWIND THIS RUN: sentiment raised +2 because Jul 20 delivered the opposite of what this thesis wants — IREN signed ~$2.8B of new AI-cloud contracts (+18.55%) and Hut 8 a 15-year, $9.8B data-center lease. Contracted AI capacity demand is the explicit invalidation leg of the bear case, and two peers just printed evidence against it. Position retained, but this is the name to watch for a genuine invalidation rather than a drawdown-driven exit. | durable contracted AI capacity demand plus successful debt refinancing invalidates the bear case — partially evidenced by peer deals on Jul 20; anchor-customer concentration resolves favorably; short interest / squeeze risk after a −52% drawdown | 64 |
| MRVL | AWS custom-silicon concentration inside a still-priced-for-growth multiple. Benchmark Buy→Hold; fell ~8–9% Jul 16 and extended lower Jul 17. Sentiment raised +2: MRVL was named among the AI plays leading the Jul 20 early bounce, so the freshest confirming catalyst has partially unwound. | AWS re-award or custom-silicon diversification; AI-ASIC demand stays strong; sell-side PTs still ~$253 (bullish) — this is a contested short | 62 |
| PLTR | Pure valuation-compression short (P/S ~62, P/E ~151) against genuinely strong growth. Consensus PTs bullish (~$183), so this is contrarian and needs a de-rating catalyst it does not yet have. Lowest-conviction name in the downside book. Reports ~Jul 30. retail-heavy | sustained 50%+ revenue compounding justifies the premium multiple; crowded and retail-heavy — squeeze risk; no identified de-rating catalyst | 59 |
| DE | Ag downcycle plus tariff drag. FY26 net-income guide cut to $4.0–4.75B (down y/y); ~$1.2B tariff headwind; PT ~$648 only ~8% above price. Reports ~Aug 20. | 2027 replacement-demand snapback off a 17-year-low fleet age; trade normalization; ag-cycle trough reverses the earnings decline | 59 |
| DKNG | Structural share erosion to CFTC-regulated prediction markets (Kalshi, Polymarket). PTs cut broadly (Moffett $38→$27, Buy→Neutral; BNP a rare Sell). Reports Aug 6. | regulatory clampdown on prediction markets; handle/margin reacceleration; market cap near the universe threshold | 57 |
| UNH | DEMOTED, and this is the last run it is carried on inertia. The Jul 16 Q2 beat-and-raise materially damaged the short and the stock rose again Jul 17. Only the unresolved DOJ Medicare-Advantage probe (status UNVERIFIED beyond 'still open') supports a residual bear case. Health Care was the worst S&P sector on Jul 20 (−1.15%), but that is sector-level and does not rehabilitate a broken single-name thesis. FORMAL RECOMMENDATION: drop next run unless a specific adverse catalyst lands. | the short thesis has already been invalidated by fundamentals; carrying it costs conviction credibility; DOJ probe may resolve benignly | 50 |
C · Options Heat
Where the premium is paying up
| Ticker | Signal | Read | Implication | Lag |
|---|---|---|---|---|
| VIX | iv_spike | 18.65, −0.12 (−0.64%) — declined slightly despite the index closing red and despite fresh Iran escalation over the weekend | Mildly bullish / risk-tolerant. Hedging demand plateaued rather than extended after Friday's +12.2% spike, and one vol desk noted VIX futures stayed lower into a red Monday close — no fresh panic bid. Consistent with rotation, not de-risking. | Index close from Yahoo's post-close board (Tier 2 media/aggregator, not exchange-primary). Futures observation is a Tier-3 newsletter read. No live tape. |
| TSLA | iv_spike | event premium into the Jul 22 AMC print; free digests carried from Jul 16 implied ~±5.7% to ~±7.6% — sources disagree | Uncertain, high magnitude. The largest scheduled binary in the horizon; premium is rich and direction is genuinely two-sided given deliveries are already known. | Free sources disagree on the implied-move level and none is current to Jul 20. No live tape. Free/EOD sources only — no live options tape, no true block/sweep print. Signal is inferred positioning, not confirmed flow. |
| GOOGL | iv_spike | pre-event premium into the Jul 22 AMC print | Uncertain — event premium, with a Gemini-3.5 delay and AI-capex overhang as the two-sided drivers. | Implied-move level UNVERIFIED on free tier; carried qualitatively from the prior run. No live tape. |
| AAPL | iv_spike | pre-earnings premium building after a record high and a −2-to-3% reversal; exact report date UNVERIFIED (reported as 'next week' by a Tier-3 source, consistent with a late-July print) | Hedging. Positioning into an event the stock has already run ~23% YTD into; IV-crush risk after. | No free-tier IV percentile or OI retrieved; premium is INFERRED from event proximity, and the event date itself is unconfirmed. Free/EOD sources only — no live options tape, no true block/sweep print. Signal is inferred positioning, not confirmed flow. |
| IREN | vol_oi | an 18.55% single-session contract-news gap would normally be accompanied by heavy short-dated call volume relative to prior OI — NOT OBSERVED, only inferred | Bullish but low confidence, and explicitly the weakest-evidence entry here. Flagged because an unverified inference on a retail-heavy momentum name is exactly the kind of signal that should be labelled rather than quietly scored. | NO contract-level data verified on any free source. This is an inference from price action alone and should be treated as UNVERIFIED. Free/EOD sources only — no live options tape, no true block/sweep print. Signal is inferred positioning, not confirmed flow. |
| WBD | pc_skew | deal spread widened to ~19.5% against Paramount Skydance's $31 bid (close $25.86, ~16-20% below the offer depending on source); volume 44.2M vs a 20.6M three-month average (+115%) | Hedging / event-driven. The volume figure is confirmed; the options skew is inferred from the arb spread. Aug 3 preliminary-injunction hearing is the next binary. | Share volume is confirmed (Tier 2); options positioning is INFERRED from the deal spread, not observed. No free-tier IV/OI verified. Free/EOD sources only — no live options tape, no true block/sweep print. Signal is inferred positioning, not confirmed flow. |
| SMH | iv_spike | sector vol elevated after a −10% week (worst since April 2025) and a bear-market entry; SOX +0.6% Jul 20 on the first up-day in six | Uncertain. Realized vol is high and implied has repriced; a single stabilization day does not confirm a bottom. | Sector-level inference from index moves; no ETF options chain data retrieved on free tier. Free/EOD sources only — no live options tape, no true block/sweep print. Signal is inferred positioning, not confirmed flow. |
D · Catalyst Calendar
Upcoming events
| Date | Event | Tickers | IV move |
|---|---|---|---|
| 2026-07-21 | Q2 earnings (GM, 3M, Charles Schwab, Halliburton; Kinder Morgan launches midstream season). SCHW carries an added wrinkle — BMO downgraded it Outperform→Market Perform on Jul 20 after a ~20% rally. | GMMMMSCHWHALKMI | normal |
| 2026-07-22 | The week's binary cluster: Tesla and Alphabet after the close, plus GE Vernova, Texas Instruments and ServiceNow. Three of these are live positions — TSLA is the #1 Track B short, GEV the flagged demotion candidate, and GOOGL carries the AI-capex read-through. | TSLAGOOGLGEVTXNNOW | high |
| 2026-07-23 | Intel, RTX, T-Mobile and Honeywell earnings; ECB policy decision (hold expected); flash PMIs for the US, eurozone and UK. Intel is the read-through on its announced new layoff round (>5,000 US roles cut so far). | INTCRTXTMUSHON | high |
| 2026-07-24 | Exxon Mobil, American Express and Verizon. XOM lands into a tape where energy was the leading sector and WTI sits at $83.23 with Brent at $89.22. | XOMAXPVZ | normal |
| 2026-07-27 | Horizon boundary (5th trading day). Just beyond it: Starbucks ~Jul 29, Palantir and Monolithic Power ~Jul 30, and Apple's June-quarter report (date UNVERIFIED, reported as late July). | SPX | unknown |
| 2026-08-03 | OUTSIDE THE HORIZON but the dominant scheduled binary for the Track A #1 faller: preliminary-injunction hearing on the Paramount Skydance / Warner Bros. Discovery deal. A delay past Sep 30 triggers a 25-cent quarterly ticking fee to WBD holders. | WBDPSKY | unknown |
| ongoing | UNSCHEDULED, and the largest actual risk in the window: the US-Iran conflict (10th consecutive night of strikes as of Jul 21) and the Houthi maritime embargo on Saudi Arabia. Oil is the transmission channel into yields and into the whole index — Jul 20 was a live demonstration of it. | XLE | unknown |
E · Trade Ideas
What the data implies
WBDneutral · track A
Rationale Not a directional view — a spread-versus-probability view. The stock closed $25.86 against a $31 bid, a ~19.5% spread, because a TRO paused the deal for 14 days on state antitrust grounds despite prior DOJ clearance. The spread now prices litigation risk, and the Aug 3 preliminary-injunction hearing resolves a large part of it.
Invalidation A preliminary injunction granted on Aug 3, or termination of the merger agreement, invalidates the arb entirely and re-rates the stock to standalone value. Conversely, a denial collapses the spread.
Risk Binary legal outcome; no informational edge available on the free tier. The ticking fee past Sep 30 is a modest offset, not a hedge.
TSLAshort · track B
Rationale Carried Track B short into a Jul 22 AMC print where deliveries (~480k, +25%) are already known — so the print is entirely a margin question against a robotaxi-priced multiple, after four straight quarterly misses. Mild confirmation Jul 20 as one of the two biggest S&P point decliners on a day chips rose.
Invalidation Auto gross margin ex-credits inflecting higher on the Jul 22 print, or credible robotaxi/FSD monetization evidence. Either ends the thesis — not weakens it.
Risk Single largest event risk in the book. A beat can end this position in one session.
IRENneutral · track A
Rationale The day's #1 riser on a real, dated, company-announced catalyst (~$2.8B of AI-cloud contracts, ARR target lifted above $4B). Bias is deliberately neutral rather than long: an 18.55% contract-announcement gap is a repricing event, not persistent momentum, and the named counterparties have not independently confirmed.
Invalidation A close back below the pre-announcement level would indicate the contract news was faded rather than absorbed.
Risk Crowded and retail-heavy; market cap sources disagree materially ($12.0B vs $21.4B); the options read here is inference from price action alone and is explicitly UNVERIFIED.
CRWVshort · track B
Rationale Carried short on a debt-funded, customer-concentrated AI buildout (>7x leverage, PTs ~$141 versus a ~$73 price). Listed here specifically to record that the evidence moved AGAINST it on Jul 20: two peers signed large multi-year AI-capacity contracts, which is the explicit invalidation leg of this thesis.
Invalidation Durable contracted AI-capacity demand plus a successful debt refinancing. The IREN and Hut 8 deals are partial evidence toward exactly that; a third such datapoint, or a CRWV refinancing on reasonable terms, should trigger removal rather than another carry.
Risk The most likely name in the book to be wrong. Flagged watch_for_invalidation so the next run has to make an explicit decision rather than carrying it by default.
SMHneutral · track A
Rationale Sector-level scenario rather than a name. The SOX fell ~10% last week (worst since April 2025), entered a bear market, then put in a +0.6% first up-day on Jul 20 having been up ~3% intraday. That intraday give-back is the informative part: dip demand exists but did not hold.
Invalidation For the stabilization read: SOX failing to hold above 12,000, or Micron losing the ~$897.80 level again, would say the bounce was a fade rather than a base.
Risk One session. The Kimi-K3 open-model commoditization overhang that caused the de-rating is unresolved, and Alphabet, Texas Instruments and Intel all report inside the horizon.
Methodology & limitations
- NO LIVE OPTIONS TAPE. Every options entry is EOD or inferred; no block/sweep print, no verified contract-level vol/OI, no name-level put/call ratio was retrievable. The IREN options entry in particular is inference from price action alone.
- SENTIMENT VELOCITY IS NOT MEASURED. No Reddit or FinTwit mention-rate data was retrievable for Jul 20; the sub-score proxies media tone. Treat it as the weakest of the four Track A components this run.
- Both Track A lists are short (6 and 5 of 10) because free-tier coverage of a −0.19% session is thin, not because the market was quiet in a way that made names undeserving. Exclusions are enumerated by name and reason in list_completeness_note.
- Three market caps in ranked entries are null (AMD, WBD, NFLX, DIS) with documented eligibility bases rather than invented figures; several Track B caps are Jul 16 snapshots and are stale.
- 10Y close is approximate (4.558% / 4.585% / 4.59% / ~4.598% across sources); the spread is disclosed rather than averaged silently.
- Two ranked names carry null move_pct (NVDA, TSLA) — direction Tier-1 confirmed, magnitude never published on the free tier. AAPL carries a flagged midpoint of a conflicting range.
- Scores are heuristic and not backtested.
- Track B estimate-revision rigor is limited by paywalls; revision direction is inferred from PT/rating actions.
Not financial advice. Educational/analytical only — generated by Daily Market Pulse pipeline — build.py + independent validate.py gate (zero-error release requirement). Do your own diligence.