Position post-mortem: the $275 stop was gapped through at the 7/14 open and the exit executed at the open per discipline — this long settles at a heavy loss. Lesson archived: into any event window (earnings or warnings), event positions get halved, never carried at full size overnight.
-27.23%07-14“Friday's sub-1nm hero became Tuesday's lead actor in the worst day of a 115-year history — the stop got steamrolled by a 20% gap, and the tuition buys one iron rule: technical levels don't exist in front of an earnings warning, so halve the position before the event window and discuss faith later”
Score5/10July 14 -25.2% — worst day in 115 yearsQ2 warning Rev $17.2B vs $17.86B est; EPS $2.93 vs $3.02Market cap erased ~$67BStop execution $275 gapped through; exited at the openRSI(14) 29.2
Technicals — Gapped below every average (20-day $272, 50-day $262, 200-day $275) to a fresh 52-week low at $211.03, then lost another 2.7% on 2.27x volume. An RSI-29 bounce can come any time, but everything below $230 — the gap's lower rim — is broken structure.
Fundamentals — The warning's substance is uglier than the drop: revenue up just 1%; customers diverted budgets from software to scarce AI hardware as memory and server prices squeezed IT wallets; several large deals slipped; and Krishna admitted Anthropic's Mythos release froze big contracts while clients rethink architectures — AI is both IBM's story and its assassin.
News — Preliminary Q2 numbers dropped premarket on 7/14 with a 'worse than our expectations' warning; the 25.2% collapse erased ~$67B of value, eclipsing the 1987 record. Street targets are being cut in batches while the options market buzzes with oversold-bounce structures. The full report lands 7/22.
Short-term · 1–2 weeks (into the 7/22 full report)
NeutralSidelines
Position post-mortem: the $275 stop was gapped through at the 7/14 open and the exit executed at the open per discipline — this long settles at a heavy loss. Lesson archived: into any event window (earnings or warnings), event positions get halved, never carried at full size overnight.
Entry No bottom-fishing, no chasing shorts; there's no long structure below $230, and only a stalling bounce into $230–240 opens the short-watchStop —Target —
Long-term · months+
Neutral
The Red Hat/consulting cash flows and the Anderon quantum option still exist, but budget crowd-out by AI hardware is structural and industry-wide; wait for the 7/22 full report and guidance before repricing.
Software demand persistently crowded out by AI hardware
Slipped deals turning into cancellations
Capex strain from quantum/advanced-node bets
Signal BacktestCumulative -27.23%price itself -27.23%0/1 closed trades won
07-14Long$290.23 → closed $211.207-16-27.23%
Community Voices
WSB · Mentions collapsed from 1,132 to 243 — the whole board talked about it the night before the crash, nobody knows it after; retail attention is the fastest stop-loss there isOriginal ↗
Institutional Views
华尔街共识 · 27 analysts average ~$291 (+38% vs spot) — but post-warning cuts are rolling in; treat the number as expiring
Forbes · Worst day in 115 years: the warning plus chip-cost mismatch did itSource ↗
CNBC · The historic crash sets up unusual options trades as volatility premium spikesSource ↗
-8.17%07-07“Down 12% in two days with the excuse rotating from 'a Hynix research note' to 'China competition' — the reasons change daily, the stop line doesn't: $880 is the border between supercycle and stampede; play dead above it, accept fate below it, and never negotiate in between”
Score7/10Two-day slide -12.0%One-day cap loss ~$110BWSB mentions 661 (doubled again, still #1)To the $880 stop 2.7%Consensus target $1,579 (+74.6%)
Technicals — Closed at $904 below the 50-day ($924) on 1.28x volume — one notch worse than Monday's low-volume dip. RSI at 44 is nowhere near oversold, the $880 stop is one 2.7% candle away, and the May shelf at $850 waits below.
Fundamentals — A race between fundamentals and positioning: HBM sold out for 2026, pre-sold through 2027, guidance raised on 7/9 — none of that changed. What changed is the holder base: a +221% YTD profit cushion is being cashed through whichever headline serves, and China competition (the CXMT expansion story) was merely today's exit door.
News — Day two, down 8%: coverage cites China memory-competition fears plus SK Hynix/CoreWeave chain profit-taking, dragging AMD, Intel and Marvell down 5%+ together. Buffett's line — 'tough to find values when everybody is preferring gambling' on AI — got quoted everywhere.
Short-term · 1–2 weeks
LongBullish
Losing the 50-day on volume is the trend's first crack, but fundamentals (sold out, prices rising) offer no reversal evidence — so the verdict stays delegated to $880, the line written when the position opened. Mid-trade is no time to change your answer.
Entry Hold but no adds; a break of $880 means exit, and add-talk resumes only above the 50-day at $924Stop $880Target $1,100
Long-term · months+
Accumulate
HBM pricing power and 2027 revenue visibility are intact; once the holder base clears, the supercycle's next leg gets priced by fundamentals, not profit-takers.
CXMT capacity eroding non-HBM pricing mid-term
High volatility while the profit overhang clears
Escalating price competition from Korean memory makers
Signal BacktestCumulative -8.17%price itself -12.40%profitable since 07-10
07-07Long$984.75 → open → $904.28-8.17%
Community Voices
WSB · Mentions doubled again from 320 to 661, still #1 — the same crowd went from 'gift dip' to 'get out' in two days; sentiment broke before the chart didOriginal ↗
Institutional Views
华尔街共识 · 53 analysts average $1,579 — 74.6% above spot; 1.14 rating, near strong buy
24/7 Wall St · Pins the 8% drop on China-competition fears, dragging Intel, AMD and MarvellSource ↗
Motley Fool · The pullback puts the valuation back in the contested zone between 'hikes continue' and 'cycle top'Source ↗
New today“A year-long castaway resurrected overnight by a $53 billion bid — the 9% between $55.50 and $60.50 prices two gates: board approval and antitrust clearance. Merger arb is a professionals' table; retail walking in now is just carrying yesterday's buyers on a sedan chair”
Score6/10The bid Stripe/Advent/Block consortium: $53B at $60.50/shSpot vs bid $55.52 — a ~9% spreadBoard meeting 7/20Volume 6.85x, biggest day of the yearRSI(14) 80.1
Technicals — Up 17.2% on 6.85x volume to $55.52, one candle through the 200-day ($52.95) for the first time since October 2025. RSI at 80 is extreme — but technicals barely apply to a deal stock; the anchors are the $60.50 bid and negotiation headlines.
Fundamentals — The consensus target of $46.95 sits 15% below spot — fundamentals alone never priced today; the takeover premium does. A stock down 35% over the past year just drew a three-party bid ($17B in stock, the rest cash) at a 28% premium, formally opening payments-consolidation season.
News — Stripe and Advent (with Block participating) lodged a $53B joint offer at $60.50 — nearly 28% over the prior close — with the board convening 7/20. Block rose 2.3% and ACI Worldwide 3.5% the same day as the market began stamping 'takeover candidate' on the whole payments aisle.
Short-term · Into the 7/20 board meeting
NeutralSidelines
All three endings — accept, bump, or bust — are priced by people with better information; chasing at RSI 80 earns the tail and owns the gap.
Entry Pass: the 9% spread prices deal risk, not chart room; a rejection without a topping bid opens a gap-fill toward $48Stop —Target —
Long-term · months+
Neutral
Win or lose, a $53B anchor has been set: payments have entered consolidation pricing, and any standalone PYPL thesis is subordinate to the deal outcome.
Bid rejected without a bump — gap refill
Antitrust review of Stripe's scale dragging the timeline
The stock component's value swinging with the acquirers
Community Voices
StockTwits · Trending at 8.6: the debate has shifted from 'will it close' to 'is $60.50 enough' — a target's shareholders grow their appetite on the bid itselfOriginal ↗
WSB · Mentions went 1→62: arb players and dip-missers feud in the same thread, while someone just remembered their $70 cost basisOriginal ↗
Institutional Views
华尔街共识 · 46 analysts average $46.95 — 15.4% below spot; consensus lagging the bid is a deal stock's normal state, not an opportunity
Benzinga · Best day ever: +17% on volume 4.5x the normSource ↗
FX Leaders · The $60.50 level is now both the technical and event anchorSource ↗
New today“The memory-price invoice finally arrived at the assembly shop — AI servers were already low-margin sweat money, and rising DRAM discounts even the sweat; with insiders selling $1.56 billion in three months against zero purchases, management's feet cast a more honest vote than any research note”
Score6/10Day -9.8% (down 14% intraday)Trigger Evercore flags memory costs squeezing server marginsInsider selling $1.56B in 3 months, zero buysPeers HPE/Supermicro fell too; WDC rose 5%Consensus target $504 (+22%)
Technicals — Down 9.8% through the 20-day ($422) yet still 13% above the 50-day ($358) — the first serious pullback in a year-long double. A $32.50 ATR says days like this are the new normal; volume ran 1.89x.
Fundamentals — The mirror trade of the memory supercycle: DRAM and NAND dominate an AI server's bill of materials, so Micron/SanDisk price hikes are Dell's cost inflation — Evercore models a head-on margin squeeze. Demand (AI server orders) is fine; converting it into profit is what just got harder.
News — AI hardware gave back together: Dell fell 14% intraday to close -9.8% with HPE and Supermicro in tow, while WDC rose 5% the same day — the market split the memory trade clean in half: own the price-setters, dump the bill-payers. A downgrade plus AI-overcapacity chatter piled on.
Short-term · 1–3 weeks
NeutralSidelines
The cost-squeeze story just started and memory contract prices are still climbing — the margin answer only arrives at next quarter's report, and everything between now and then is guesswork; sidelines.
Entry A low-volume hold at $395–400 (prior shelf plus round number) merits a tactical look; failing that, the 50-day at $358 is the first real catch levelStop —Target —
Long-term · months+
Neutral
The AI-server order book remains huge, and if memory costs can be passed through (escalator clauses, customer sharing), the margin panic gets disproven — but $1.56B of insider selling takes time to digest.
Memory costs stick and margins stay squeezed
AI-server order cadence slowing on overcapacity fears
New today“On the very day IBM cratered and memory stampeded, it rose 4% to the doorstep of an all-time high — in chaos, money isn't buying Apple's growth, it's buying Apple's certainty; 0.4% below a record there is no resistance, only vertigo, and vertigo has never been a reason to stop out”
Score7/10To the record high $328.73 — just 0.4% awayDay +4.0% on 1.24x volumeWSB mentions 68 (5.7x)RSI(14) 68.8Consensus target $318.76 (now below spot)
Technicals — Up 4% on volume to $327.50, one step from the $328.73 record. The 20- and 50-day averages sit far below at $302/$300, and RSI at 68.8 is warm but not redlined — a volume-confirmed breakout opens airless territory above.
Fundamentals — Safe-harbor logic on hardware-collapse day: the certainty trio of cash flow, buybacks and ecosystem lock-in got panic-bought after IBM's warning, while the OpenAI trade-secrets suit backhandedly certifies the hardware moat's worth. The consensus target now trails the price — analysts are chasing it upward.
News — A strength day without a press release: money fleeing fallen IBM and the memory chain crowded into megacap certainty (Microsoft +2.8%, Alphabet +3.2%), with Apple leading at +4% and WSB mentions up nearly 6x.
Short-term · 1–3 weeks
LongBullish
Three tailwinds: haven flows, a record-high structure and consensus playing catch-up. The $308 stop sits in the buffer above the 20/50-day cluster, a full 2.3 ATRs away.
Entry Hold; a volume break of $328.73 followed by a holding retest is the add-watch, with the $315 shelf as pullback supportStop $308Target $350
Long-term · months+
Accumulate
Ecosystem cash flows, services growth and the on-device AI option; the certainty premium still has room to re-rate in a messy tape.
The certainty premium cuts both ways if growth disappoints
China demand and geopolitical noise
On-device AI shipping slower than hoped
Community Voices
WSB · Mentions 12→68 as the old 'buy before or after the record' debate reopens — the answer was always sizing, never timingOriginal ↗
Institutional Views
华尔街共识 · 52 analysts average $318.76 — now 2.7% below spot; the price is ahead of consensus and the upgrade cycle is under way
New today“The season's opening shot printed 86% equities-trading growth — volatility is retail's cost and the house's revenue, and everyone's panic this quarter got booked into JPMorgan's P&L; a mere 1.2% post-earnings gain isn't disbelief, it's big money taking its seat slowly”
Score7/10Q2 scorecard EPS $6.14 vs $5.85 est; revenue $58B, a huge beatEquities trading +86% to $6BSector echo MS profit +58%, BlackRock +6.6%, BNY raised guidanceTo 52-wk high -1.2%RSI(14) 67.5
Technicals — Up 1.2% the day after earnings to $346.91, one step from the $351.24 high. The rising 20-day at $334 now acts as support, RSI 67.5 is strong without redlining, and the sector (XLF +0.7%, KBWB +1.5%) escorts it at collective records.
Fundamentals — Q2 steamrolled: EPS of $6.14 on $58B revenue with record equities trading (+86%); IB fees rode the SpaceX IPO wave; and Dimon called the US economy 'notably resilient' on the AI-capex, fiscal and deregulation engines. It's a double hit of volatility dividend plus reopened capital markets.
News — All five megabanks reported premarket on 7/14 with JPMorgan leading the sweep; 7/15 added Morgan Stanley's +58% profit, BlackRock's $6B quarterly private-debt haul and BNY lifting full-year revenue guidance to +10–11% — an industry-wide harvest from the trading-plus-IB twin engines.
Short-term · 1–3 weeks
LongBullish
Earnings confirmation, sector resonance and rising consensus stack up; the $326 stop sits an ATR under the 20-day. With event risk released, this is the cheapest stretch of the trend to hold.
Entry Hold; a 20-day retest at $334 is the add-watch, and a break of the $351 record opens $365Stop $326Target $365
Long-term · months+
Accumulate
Scale plus the universal-banking franchise compounds through high-vol, high-rate regimes; the biggest beneficiary of a reopened capital-markets cycle.
Provision pressure if the credit cycle sours
Trading revenue's high base is hard to lap
Regulatory pendulum swings
Institutional Views
华尔街共识 · 27 analysts average $365.36 — 5.3% above spot; consensus is still catching up to the print
CNBC · A double beat with the 86% equities-trading surge as the headlineSource ↗
-6.03%07-07“Three days pinned 1.1% above the stop — the market has nailed it to the knife's edge before earnings. A stock with a $17.70 ATR sitting $4.50 from $390 no longer answers to fundamentals; it answers to whichever day the tape sneezes. Four sessions left on a discipline position — don't improvise”
Score6/10To the $390 stop 1.1%Earnings countdown 7/22 — four sessions awayThis week -1.2%, pinned under the 20-dayATR $17.7Consensus target $407 (+3.3%)
Technicals — Down 0.4% to $394.46, a third straight session hugging the $390 stop. The 20-day ($399) and 50-day ($410) press from above on 0.77x volume — both sides are waiting for 7/22, and neither will show its hand early.
Fundamentals — The 480k-delivery card (+25%) is priced; the only suspense on 7/22 is margin — whether volume was bought with cuts and financing subsidies. Flat EPS consensus at $0.27 makes the low bar itself the biggest variable of the day.
News — A newsless waiting room. After the IBM episode, tolerance for full-size positions into earnings has dropped visibly, and the high-ATR-plus-imminent-event combination is being systematically de-risked.
Short-term · Into 7/22 earnings
LongBullish
The original discipline stands, but event risk got a heavier weight: with the stop in its face and earnings imminent, choosing 'trim first' over 'wait for the stop' is exactly what this week's tuition paid for.
Entry Hold without adds; a $390 break means exit. The IBM lesson applies: if it's still hugging the line at Monday's close, halve the position before the printStop $390Target $430
Long-term · months+
Neutral
The robotaxi/FSD and storage curves are real but heavily pre-priced; no new chips before the 7/22 margin answer.
Margin miss
Q3 sequential pressure off the delivery high base
Brand noise from political exposure
Signal BacktestCumulative -6.03%price itself -7.25%
07-07Long$419.77 → open → $394.46-6.03%
Institutional Views
华尔街共识 · 51 analysts average $407 — 3.3% above spot at a 1.77 rating; consensus is fully priced with the dispute parked on margin
+9.07%07-04“Memory cratered 12% in two days while it climbed 8.9% on the week — when the ammo plants blow up and the arms dealer still raises prices, that's what pricing power looks like. The gap between its weekly chart and the memory chain is this year's cleanest relative-strength signal; above $205, don't outsmart yourself”
Score7/10Week +8.9% (while the memory chain cratered)Day +0.3% on a day memory fell 8%Above both averages 20-day $202 / 50-day $210To 52-wk high -10.2%Consensus target $314 (+47.6%)
Technicals — Up 0.3% to $212.50, back above the 50-day ($210) with the +8.9% week scissoring away from memory. RSI 56.9 is healthy on 0.92x volume; the $215–220 June congestion sits overhead, and clearing it targets the $236 prior high.
Fundamentals — The memory panic is a cost-side tailwind — HBM is a purchase line, not a revenue line. The real tells are TSMC's rumored price hikes and ASML holding firm through earnings week: upstream capacity is queuing for its next platform. Order visibility remains the market's best.
News — No company news; the divergence day (AMD -3.5%, Marvell -7.3% vs NVDA +0.3%) confirms money rotating within the compute chain — long the strong, cut the weak.
Short-term · 1–3 weeks
LongBullish
With relative strength confirmed and both averages reclaimed, the old stop is too far away; raising it to $205 locks the floor under this leg's gains — a trend position's stop only ever moves up.
Entry Hold with the stop raised from $196 to $205 (above the 20-day); a $208–210 retest is the re-entry windowStop $205Target $236
Long-term · months+
Accumulate
The full-stack arms dealer of the AI compute race with deepening platform lock-in; upstream price hikes only entrench its pricing power.
HBM supply bottlenecks pacing shipments
Hyperscaler in-house silicon diverting spend
Export-control escalation
Signal BacktestCumulative +9.07%price itself +7.55%profitable since 07-07
07-04Long$194.83 → open → $212.5+9.07%
Community Voices
WSB · 121 mentions holds a top-10 slot; the most common line in memory-crash threads is 'so where did the money go' — its weekly chart is the answerOriginal ↗
Institutional Views
华尔街共识 · 66 analysts average $314 — 47.6% above spot; 1.14 rating, near strong buy
New today“Beat the EPS number and still fell 9.4% — insurance earnings were never about EPS, they're about the combined ratio, and a 90 paired with two megabank downgrades reads like a verdict that underwriting's golden era is entering its final act; when the defensive stock can't defend itself, step to the sideline first”
Score5/10Q2 EPS $5.67, beat the $5.30 consensusJune net income -31% YoYCombined ratio 90.0, deterioratingDowngrades JPMorgan and Wells Fargo, same dayRSI(14) 39.8
Technicals — Down 9.4% on 2.6x volume through both the 50- and 200-day ($207/$213), shattering the yearly structure in one stroke. RSI at 39.8, next support at the $189 52-week low — when a two-year consolidation breaks, the move tends to run.
Fundamentals — The headline beat ($5.67 vs $5.30) can't cover three cracks: June net income down 31%, a worsening 90 combined ratio (claims inflation plus cat losses), and premiums/book value merely in line — a slowing-growth signal that same-day downgrades from JPMorgan and Wells Fargo amplified.
News — Q2 and the June monthly dropped together, and the stock bled 8–9.4% intraday; JPMorgan and Wells Fargo double-downgraded on 'valuation ahead of fundamentals and growth set to lag peers,' while early-summer severe weather delivered above-plan catastrophe losses.
Short-term · 1–3 weeks
NeutralSidelines
Day one of an underwriting-cycle turn is never the entry; let the downgrade wave and target cuts finish before talking right side.
Entry No catches on breakdown day; a low-volume hold at the $200 round number plus gap merits a bounce-watch, while losing the $189 low upgrades this to cycle repricingStop —Target —
Long-term · months+
Neutral
Best-in-class underwriting discipline and pricing data are intact; if the ratio deterioration proves to be cat seasonality, this pullback ends up a within-cycle dip — but that takes two quarters to verify.
Up 0.4% to $754.81, pressing record highs again — the cool-CPI-plus-bank-earnings combo that even IBM's crash couldn't stop(Long-term: Core allocation unchanged; manage earnings-season risk with sizing, not timing)
Down 0.3% to $717.74 as IBM and memory drags met platform offsets — the split inside the index says more than the level(Long-term: Tech core stays put; divergence is a stock-picker's market)
Down 1.1% to $181.58, dragged by the hardware side — the gap to QQQ is exactly the IBM/memory weighting difference(Long-term: With rotation this violent inside the sector, broad-market exposure is the calmer ride)
Up 0.7% to a record $56.56 — earnings week's lead sector; RSI 71 is hot but the trend holds(Long-term: The bank-earnings cycle plus deregulation is a medium-term tailwind)
Up 1.5% to $98.44 on 2.8x volume, the bank ETF at full throttle through earnings week, closing on $100(Long-term: Sector beta on a clean five-bank sweep; the pullback beats the chase)
Flat at $372.35, gold steadying on the cool CPI as last week's liquidity distortion mends(Long-term: The allocation case stands; post-panic stabilization opens the add-watch window)
Up 0.1% to $98.14 as the cool CPI pulled yields lower — core bonds' defensive value is back on the hot list(Long-term: The ballast leg of the stock-bond hedge; a hawkish Fed is the only headwind)
Down 0.4% to $35.88, Brazil pausing at the highs — the commodity-plus-carry case is intact(Long-term: A diversifier with twin commodity and currency exposure)
Up 3.1% to $27.00 as China internet ADRs chase the HK AI re-rating, with the Alibaba weighting as the engine(Long-term: The dollar-denominated lane into HK's AI re-rating; volatility is the ticket price)
Up 0.6% to $8.80 and back on the hot list — circumstantial evidence of returning risk appetite, though 62% off the high is early to call trend(Long-term: Futures-based bitcoin ETFs bleed roll costs; spot vehicles suit long holds better)
Up 0.2% to $25.51 on 235x volume — institutional creation/redemption flows in a young fund, not a trading signal(Long-term: A new flexible-allocation product with too little history to judge)
Down 6.3% to $165.55, the 3x invoice of memory's second leg, -37.8% on the month — in a divergent semis tape, leverage is self-harm(not for long-term holding)
Down 3.5% to $529.14 in the memory sympathy hit, 2.1% above the $518 stop — trailing NVIDIA by a full length is getting conspicuous; a break means exit
Down 8.5% to $390.33: the $7.45 EPS smash didn't matter — the 3.5% operating margin (4.9% last year) was the verdict, and managed care fell with it; sidelines
Down 8.1% to $1,615 in memory's second leg, -23% on the month; WSB mentions up 2.6x and all of them exit threads — the $1,580 shelf is the last line; sidelines
Down 16.8% on 4.2x volume to $44.52 in the vocational-ed double kill (APEI -15%) — synchronized collapses usually mean policy chatter; wait for the filing
Up 49% on 81x volume to $11.76 on an Amazon supply deal plus warrant incentives — real orders and real dilution; remember both amid the party, sidelines
Down 3.65% to $66.31, mystery solved: a $1B convertible raise sparked dilution panic and 5.8x WSB mentions of fury — pre-launch raises are routine, and panic signposts the opportunity zone, but don't jump the gun
Down 0.6% to $135.27, below the IPO price for the first time — the passive money is now all underwater; staying sidelined into the 8/6 earnings-plus-lockup double event
Up 28.8% to $29.63, +76% on the week — the board's new darling where those who can't explain it outnumber those who can; that's the entire signal, avoid
Stock Brief 2026-07-16: Daily U.S. & HK Market Analysis Archive · Quant Brief
