$AMD — FinTwit mentions on Surfaced

Last 12mo stats

1.6k

mentions

56

voices

$AMD

Advanced Micro Devices

$AMD

Advanced Micro Devices

Past 12mo:1.6kmentions56voices

All voices1000 mentions since Mar 2026calls only

Loading price chart…

About this stock

$AMD attracted 1,590 mentions from 61 distinct FinTwit voices over the past 12 months, with the first mention recorded on August 23, 2025, and the most recent on July 18, 2026. The most-active voice category centered on long research discussion. @stocksavvyshay led engagement with 262 mentions, followed by @mrmikeinvesting with 169 mentions and @investmentguru_ with 135 mentions. Three additional voices—@yasutaketin, @aleabitoreddit—each contributed over 100 mentions during the period.

Recent discussion of $AMD focused on semiconductor demand dynamics within the AI infrastructure space. @yasutaketin connected $AMD's GPU capacity to anticipated demand from open-weight AI model deployment, noting that cheaper AI alternatives would require substantial HBM, NAND, DRAM, and GPU infrastructure. @kawzinvests addressed competitive dynamics in the AI hardware market, discussing how open-source model releases affect semiconductor demand across $AMD, $NVDA, and $MU. Dark pool transaction data from July 17, 2026, showed $AMD recording $1.05 billion in net positive off-exchange activity across 691 prints during the weekly OPEX session.

AI-generated from tracked posts · updated Aug 24, 2026

Related stocks

$TSMTaiwan Semiconductor Manufacturing

$NBISNebius Group NV

$MRVLMarvell Technology

Recent mentions

latest 100 of 1000

@yasutaketinJul 18, 2026 · 5:53 AM

Kimi K3が報道されているように優秀なら、無料のオープンモデルとして、例えば $NBIS のようなAIクラウドで西側諸国のユーザーもAIエージェントとして長時間利用すると考えられる。（主にコーディング用途）そこで HBM NAND $DRAM や $AVGO $MRVL などのネットワーキング、 $NVDA $AMD の GPUなども大量に必要になる。ちっとも悪い話しではない。むしろフロンティアAI企業よりも安価なAIエージェント時代の到来を加速する。

@kawzinvestsJul 17, 2026 · 9:41 PM

Did Kimi just start another DeepSeek moment? Kimi released the largest open weight model ever yesterday. It scored 57 on the Artificial Analysis index, 3 points behind Claude Fable 5, causing AI stocks to panic. The bears say intelligence is becoming a commodity. Open models from China collapse token pricing, frontier labs lose the revenue that funds their next training run, and the economics of ever larger training runs get harder to justify. Multiples compress starting with neoclouds, power plays, and data center REITs. The bulls point to Jevons paradox. Cheaper models raise the ROI on AI spend, which drives more token consumption and pushes margin out of the model layer into semis, power, and data centers. K3 also needs GB300 class hardware with 288GB per chip just to load, so wide deployment of this model means more orders for the highest end silicon. On cost, K3 runs $0.94 per task vs $1.04 for GPT-5.6 Sol, because it burns 50-70% more tokens per job. DeepSeek undercut everything on price. K3 lands at the same per-task cost as the closed models it competes with. We cover both cases, the four scenarios, and K3's self-optimization capability in the free article below Who do you think is right, the bulls or the bears? $NVDA $AMD $MU

@alphaticaioJul 17, 2026 · 8:56 PM

DARK POOL SCANNER | July 17, 2026 · Weekly OPEX +$15.96B net buy. Largest dark pool buy of the series. All eight top names net positive. $MU +$5.0B. Tuesday: -$15.7B. Today: +$5.0B. The position rebuilt. Q3 week three. 20,713 off-exchange prints. +$15.96B net. NET BUY. Dark Pool Buyers (all eight, zero sellers): $MU +$5.00B (4,686 prints, TUE: -$15.7B, YEST: -$1.31B, REBUILT) $QQQ +$3.88B (2,322 prints, YEST: -$580M, FLIPPED) $SPY +$2.35B (2,132 prints, LIT: -$311M, TAPES DISAGREE) $AMD +$1.05B (691 prints) $NVDA +$770M (1,326 prints, BOTH AGREE BUY, LIT: +$3.77B) $IWM +$557M (303 prints, 4 SESSIONS SELLING, FLIPPED) $META +$290M (770 prints) $AAPL +$161M (921 prints) +$15.96B. The largest dark pool buy session of the series. By nearly 5x. The previous record was +$3.29B (Jul 10). Today shattered it. 20,713 prints. $30.10B in buys vs $14.14B in sells. Sellers were active in every name. The buyers overwhelmed them at scale. All eight names in the top 8 are net positive. Yesterday: seven names both tapes agreed sell (broadest cross-venue sell of Q3). Today: all eight net buy on the dark pool. One session apart. $MU: +$5.0B. The reversal arc completed. Tuesday: -$15.70B (largest single-name sell of the series). Wednesday: -$1.31B dark pool sell. Thursday: -$30M lit (nearly flat). Today: +$5.0B dark pool buy. The institution that unwound rebuilt the position in one session. $IWM: +$557M. Four consecutive sessions of selling. -$772M combined across both tapes yesterday alone. Today: +$557M. The small-cap sell exhausted and reversed. The internals showed AMEX 1.56x after four sessions below 0.70x. The flow confirms: small caps are being bought again. $QQQ: +$3.88B. Yesterday: -$580M. The dark pool that agreed with the lit tape on selling yesterday is now buying at the largest QQQ dark pool session of the series. $SPY: +$2.35B dark pool. Lit: -$311M. The tapes disagree. The Q3 pattern persists. The dark pool buys what the lit tape sells. Weekly OPEX. Indexes red. The dark pool bought at the largest level of the series. The data follows the data. Watching the tape.

@stocksavvyshayJul 17, 2026 · 4:26 PM

@FuturumEquities $NVDA, $AMD, $AVGO and any future architectures will compete to determine who wins the next phase of AI but almost all of them will still need $TSM to manufacture the chips.

@wliangJul 17, 2026 · 4:00 PM

@Lazarus_Capital OCD kept me at 4 tickers each category. Also didn't mention my largest position, $AMD. Many unnamed tickers, hahah

@investmentguru_Jul 17, 2026 · 11:06 AM

Long-term buying opportunities across the current setup- $RKLB is becoming the second full-stack launch and space systems player behind SpaceX, with Neutron set to open up the medium-lift market. $ASTS is building direct-to-cell satellite connectivity with major carrier partnerships already in place. Still early and capital intensive, but the addressable market is enormous if the constellation build-out executes. $MU sits at the center of the memory/HBM bottleneck for AI infrastructure. Every GPU and AI server needs high-bandwidth memory, and Micron’s pricing power in this cycle has been a standout story. $AMD is the clearest alternative to Nvidia in AI compute, with MI-series accelerators gaining hyperscaler traction. Datacenter and client CPU share gains add a second growth engine beyond just the AI narrative. $CRDO is a picks-and-shovels photonics/connectivity play, supplying high-speed interconnects that hyperscalers need as AI clusters scale. $MRVL Marvell’s custom silicon and networking chips are embedded deep in hyperscaler AI infrastructure. Diversified exposure across compute, storage, and optical networking makes it a core semiconductor holding in this cycle. $IONQ is a leading pure-play in trapped-ion quantum computing, still pre-profitability but with government and enterprise partnerships building out. High-risk, high-optionality bet on quantum reaching commercial relevance this decade. $NBIS is a NeoCloud name providing GPU infrastructure-as-a-service, spun out of Yandex with clean Western structure. Positioned to capture AI compute demand overflow that hyperscalers can’t build fast enough themselves. $SPCX offers direct exposure to launch, Starlink, and the broader space infrastructure stack from the dominant player. Rare opportunity to own the category leader rather than just the challengers. $TSLA long-term thesis has shifted from EVs alone to autonomy (FSD/robotaxi) and robotics (Optimus). Volatile and expensive on traditional metrics, but the optionality on physical AI is the real long-term bet. $GOOG combines search cash flow with genuine AI leadership through Gemini, TPUs, and Waymo. $OKLO is a small modular reactor developer, targeting the power bottleneck that AI datacenters are running into. Pre-revenue and regulatory-dependent, but squarely positioned in the nuclear layer of the energy buildout. $HIMS is scaling telehealth with a direct-to-consumer model in weight loss, hair, and men’s health. Margin expansion and category breadth make it a differentiated healthcare-tech growth story. $SOFI is building a full digital banking stack with a young, sticky customer base. Loan Platform Business and membership growth give it multiple paths to sustained profitability. $HOOD has expanded well beyond commission-free trading into crypto, options, and now prediction markets and international expansion. $GLXY offers regulated, institutional exposure to crypto infrastructure, trading, and asset management. A way to play digital asset adoption without taking direct token risk. $INTC is a turnaround and onshoring/foundry bet, with government backing tied to U.S. chip manufacturing independence. Higher risk than peers, but foundry execution and node progress (18A and beyond) are the catalysts to watch. $AAOI is a smaller-cap photonics/optical component supplier riding the same CPO/interconnect bottleneck thesis as larger names. More volatile, but high torque to datacenter optical demand. $COST — Costco is the steady compounder in this list — membership model, pricing power, and loyalty that holds up in any macro environment. Not a thematic AI play, but a ballast position for a portfolio otherwise leaning aggressive. $LLY — Eli Lilly’s GLP-1 franchise (obesity/diabetes) is one of the most durable growth stories in pharma right now. Deep pipeline beyond GLP-1s adds ballast to a name already compounding at a premium multiple. Not financial advice.

@frenchie_Jul 17, 2026 · 8:16 AM

@MYptf1ot I already reduced high beta exposure After a violent move from my best runners $AXTI $IQE and some $VPG I’m now focused on monitoring exposition, its going to be a hard time for anyone who never experienced a tech drawdown, ultimately I see most of the names I bought recover $MRVL $CIEN $STM $NVDA $AMD and so on Tech always do that, we’ve seen meta dip 88%, GOOGL 50% High risk means high rewards, the « everything easy » era is done This will permit us to see what companies were tourist stocks and what were the ones with solid fundamentals https://t.co/EW0Qcf9I1Q

@josenajarroJul 16, 2026 · 9:58 PM

$TSM CEO Doesn't Fully Trust the AI BOOM!! And I Find That Bullish. Wei said customers each tell their own truth, but the aggregate isn't the full picture "I believe every customer tell me the truth. Everyone. You put all the truths together, it's not the truth." So TSMC is verifying instead!! "checking the AI data centers of progress, the building, the location, the demand, the rest. We checking all that to make sure that TSMC's chips will not be put in inventory." This is cycle trauma turned into an advantage. TSMC got burned before. Now they're protecting the entire supply chain below them from building on phantom demand. That discipline is exactly why this cycle can last longer. A greedy company would jack prices, build everything, and let the downstream players figure it out. TSMC is doing the opposite. They are verifying first and showing pricing restraint. On margins, Wei was self-aware: "I'm really jealous about memory companies, 86% gross margin. About 68%, I would be happy about that." On pricing philosophy: "We don't suddenly increase our price by, which I like to have, a 4x or 5x." When the company that plans capacity years ahead and bears the cost of being wrong says "from this day on all the way to probably 2029, 2030, the demand is very strong" and that "the gap is very big" it means more to me than most hyperscaler commentary. The day TSMC stops verifying and starts acting greedy is the day its signal loses its edge, but for now their earnings made me very bullish for the semiconductor industry $NVDA $AVGO $AMD

@itsmichaelluuJul 16, 2026 · 9:56 PM

1. $AMD ~$1200 The "legit second source" to $NVDA. Agentic AI is CPU-heavy, and EPYC owns that lane every step up the AI stack adds general-compute demand. Catalysts stacked: Advancing AI 2026 (Jul 22–23), Zen 6 "Venice" on TSMC 2nm, Q2 earnings Aug 4. Consensus ~$800 | Street-high $900 (Cantor, UBS) 2. $NBIS ~$400 The neocloud with receipts: 684% Q1 revenue growth, a $46–50B contracted backlog, landmark Meta and Microsoft deals, and a $2B NVIDIA investment. Raised 2026 power guidance past 4 GW. Consensus ~$244 | high $380 → +25% / +94% Rich (~96x P/E), insider selling, whippy on neocloud sentiment. 3. $ASTS ~$240 Direct-to-phone satellite broadband, ~60 carrier deals, BlueBird constellation scaling, and now fully funded (or close) for 90-satellite deployment after the $1B note raise. The pure moonshot of the list. Consensus target ~$81 | high $108 → +21% / +61% Consensus is only a Hold still pre-profit (2025 loss ~$342M), dilution + execution risk. 4. $MU ~$2500 The memory bottleneck trade. HBM/DRAM demand outrunning supply, 16 multi-year non-cancelable agreements worth $22B, and an Anthropic partnership plus equity stake. Just pulled back ~22% from its $1,255 ATH that's the dip. Consensus ~$1,486 (Strong Buy) | high $1,750 (KeyBanc) → +65% / +94% Burry short + SK Hynix IPO + "memory peaking" chatter = volatile (beta 3+). 5. $LITE ~ $2400 Optical picks-and-shovels for AI. Co-packaged optics, a major OCS contract win, an NVIDIA $2B investment, and a stated $8B annual revenue target. Earnings Aug 10. Median target ~$1,100 (range $600–$1,400) → +29% / +65% 6. $AVGO ~ $700 Custom AI silicon (XPUs) + networking + the VMware software annuity. Just landed a $30B custom-chip deal with Apple. 1-yr target est ~$524 | bull case ~$600 → +36% / +56% 7. $IONQ ~$150 Quantum pure-play converting science into revenue: ~80% YoY organic growth, $3.3B in cash, and a diversifying commercial base. 256-qubit roadmap is the catalyst. Consensus ~$70 | high $200 → +27% / +82% Still burning cash hard (Q1 operating loss ~$272M), beta 3.05 drawdowns are violent. 8. $NOW ~$300 Enterprise agentic AI at scale. $1B AI ACV target for 2026 and a $15B+ subscription revenue target, plus Now Assist monetization and the IBM tie-up. Down ~23% YTD = the entry. Consensus ~$141 (Strong Buy) | high $400 → +33% / +123%

@itsmichaelluuJul 16, 2026 · 9:11 PM

The AI "bubble" already popped. Average drawdown: -42% from peaks so far: 🔴 NuScale $SMR -85% 🔴 Oklo $OKLO -76% 🔴 Redwire $RDW -69% 🔴 Navitas $NVTS -63% 🔴 Oracle $ORCL -62% 🔴 Intuitive Machines $LUNR -62% 🔴 AST SpaceMobile $ASTS -58% 🔴 IonQ $IONQ -57% 🔴 Planet Labs $PL -57% 🔴 Super Micro $SMCI -56% 🔴 Applied Opto $AAOI -56% 🔴 Rocket Lab $RKLB -55% 🔴 IREN $IREN -54% 🔴 Ondas $ONDS -53% 🔴 Tempus $TE -51% 🔴 ServiceNow $NOW -50% 🔴 CoreWeave $CRWV -50% 🔴 Applied Digital $APLD -43% 🔴 Ouster $OUST -42% 🔴 Arm $ARM -42% 🔴 Bloom Energy $BE -39% 🔴 Nebius $NBIS -38% 🔴 Palantir $PLTR -38% 🔴 Marvell $MRVL -37% 🔴 AeroVironment $AVAV -34% 🔴 Cameco $CCJ -33% 🔴 Coherent $COHR -32% 🔴 Western Digital $WDC -31% 🔴 Vistra $VST -30% 🔴 Constellation $CEG -30% 🔴 Microsoft $MSFT -28% 🔴 Intel $INTC -28% 🔴 SanDisk $SNDK -27% 🔴 Lumentum $LITE -26% 🔴 Micron $MU -24% 🔴 Kratos $KTOS -22% 🔴 Broadcom $AVGO -21% 🔴 Vertiv $VRT -19% 🔴 Dell $DELL -17% 🔴 Meta $META -17% 🔴 Red Cat $RCAT -15% 🔴 Alphabet $GOOGL -10% 🔴 Nvidia $NVDA -10% 🔴 AMD $AMD -9% 🔴 TeraWulf $WULF -8% 🔴 Amazon $AMZN -8% 🔴 GE Vernova $GEV -7% 🔴 Taiwan Semi $TSM -6% 🔴 ASML $ASML -5% 🔴 Arista $ANET -5% UPDATED cheatsheet for you to save:

@stocksavvyshayJul 16, 2026 · 7:02 PM

$TSM is expected to nearly double its earnings over the next three years. TSMC does not need to predict whether the next major winner is $NVDA, $AMD, $AVGO, a hyperscaler’s custom accelerator or a new CPU architecture because nearly every contender depends on its advanced manufacturing. As long as global demand for compute continues rising, TSMC remains the manufacturing toll road beneath nearly every major AI platform.

@aleabitoredditJul 16, 2026 · 2:54 PM

Lot of people were curious about $SIVE capacity volume ramp modeling through fab-light (Win Semi + others): Using 10% of Win's wafer capacity as a low-end allocation (65% yield assumption, $50-$75 ASP): Sivers would support $341-$512M worth of annual array revenue. Given upper end of managements 50-60%+ gross margin target, would be roughly: $205–307M of annual gross profit. Against Sivers current ~$1.1B MC, would be ~: 3.6–5.4× MC/gross profit if this capacity scenario plays out in 2028. And at 15% would be $307–461M in gross profit (2.4–3.6× MC/gross profit) Sivers CEO also replied that they're working with more fabs for capacity. And from an older deck, there looks to be more qualifications since 2024. So capacity targets might be larger than what's stated here as CPO takes off. I also expect to see revenue pipeline projections hiked in future quarters, as more qualification suppliers to into HVM. _ As for demand side, CW also happens to be very bottlenecked. Lumentum are buying CW off the open market due to EML obligations from their ER transcript. And $AMD are signing LTAs to secure CW capacity (from Trendforce). So when Sivers is ramping with $GFS, $JBL, Ayar, $POET, O-NET and others... Given the current constraints, it's highly likely any independent qualified capacity that comes online would be absorbed. And as a cherry on top, Morgan Stanley named $SIVE (~$1.1B) as one of the three leading CPO laser players . Alongside $55B+ players Coherent and Lumentum in their recent note for a reason... _ TLDR: Sivers only needs a low end allocation from Win to make substantial gross income relative to current valuations. I think the largest revenue upside that isn't modeled in if they TAM expansion with M&A after US NASDAQ listing. By copying the Lumentum playbook with Cloud Light to build out entire transceiver modules or with optical engines.

@investmentguru_Jul 16, 2026 · 11:19 AM

$TSM Q2: The Foundry Confirms the Thesis TSMC just delivered the clearest AI capex signal of earnings season. Q2 beat across the board: • Revenue: $40.2B (+34% YoY), above high end of guide • Gross margin: 67.7% (vs 67.1% est) • Net income: $22.4B (+74% YoY) The real story: HPC (AI accelerators, custom silicon, networking) now 66% of revenue, up from 61% last quarter. This isn’t smartphones carrying TSM anymore — it’s AI infrastructure. Three signals for the Silicon-to-Substation thesis: 1️⃣ Demand accelerating, not slowing — 2026 revenue guide raised to “slightly more than 40%” growth (from “more than 30%”). Management is seeing more demand than 3 months ago, not less. 2️⃣ Capex is the tell — 2026 capex raised to $60-64B (from $52-56B), plus another $100B pledged for Arizona ($265B total US commitment). You don’t spend like this unless the order book demands it. 3️⃣ N2 ramp underway — new node already 3% of wafer revenue in first reported quarter. Advanced nodes (3nm/5nm/7nm/N2) = 77% of wafer revenue. The leading edge is where the AI chips live. bullish for $NVDA, $AVGO, $AMD, and every custom silicon program riding TSM’s advanced packaging capacity. The recent semi selloff looks more like valuation reset than a demand problem. Q3 guide: $44.6-45.8B (+37% YoY at midpoint). Not financial advice.

@investmentguru_Jul 16, 2026 · 11:13 AM

Long term nothing changed. $MU $RKLB $AMD $ASTS $OKLO $CRDO $MRVL $NBIS $AAOI $OUST

@alphaticaioJul 15, 2026 · 8:07 PM

+$17.95B. $AAPL +$12.08B. $2.2M sells. $NVDA +$11.23B. $MSFT +$5.65B. $AMZN +$4.45B. Four names near-zero sells. $MU -$15.70B. The consistent buyer for two weeks just got unwound. Q3 week three. $74.26B in buy blocks. $56.32B in sells. Net: +$17.95B. 27,652 prints. The buyers: $AAPL +$12.08B (1,457 blocks, $2.2M SELLS, NEAR ZERO, JULY PLAY) $NVDA +$11.23B (1,742 blocks) $MSFT +$5.65B (891 blocks, $25.3M SELLS, NEAR ZERO, BOTH AGREE) $AMZN +$4.45B (791 blocks, $28.1M SELLS, NEAR ZERO, FLIPPED) $AMD +$2.96B (1,314 blocks) $SPY +$377M (3,148 blocks) The sellers: $MU -$15.70B (4,967 blocks, LARGEST SINGLE-NAME SELL OF SERIES) $QQQ -$2.00B (2,389 blocks) The biggest numbers the series has produced. $AAPL: $12.08B in buys with $2.2M in sells. That's $5,490 buying for every $1 selling. The July play at its peak. We flagged the $51M dark pool buy on Q3 day one. The accumulation built session by session to this. $MSFT: +$5.65B with $25.3M sells. Near zero. $AMZN: +$4.45B with $28.1M sells. Near zero. Four mega-caps bought at the highest conviction level the series has recorded. The opposition on $AAPL, $MSFT, and $AMZN combined: $55.6M. On $21.5B in buying. The conviction is absolute. $MU: -$15.70B. The largest single-name sell of the series. For two weeks $MU was the "only consistent buyer." The $200M pre-market alert on July 8. Both weeks. Both tapes. Today: -$15.70B. 4,967 blocks. The entire position unwound in one session and then some. The data documented the two-week accumulation. The data documents the unwind. $AMZN: +$4.45B. Last Friday: -$854M with $19.6M buys. Today: +$4.45B with $28.1M sells. A $5.3B swing. The reversal is complete and then some. The tape documents the conviction and the reversal. Both on the record. Watching the tape.

@stocksavvyshayJul 15, 2026 · 4:38 PM

Semiconductors are getting hit hard today: • $SNDK -15% • $AAOI -15% • $LITE -12% • $MU -11% • $MRVL -10% • $COHR -9% • $AMKR -9% • $INTC -8% • $CRDO -8% • $AMD -7% • $ALAB -7% • $AMAT -7% • $ARM -6% • $NVDA -3%

@wliangJul 15, 2026 · 1:56 PM

Just a shower thought... If you look at the actual numbers of AI agents, 79% of companies have adopted them, but only 11% have them running in production. That gap isn't AI failing. It's that a human still has to manually check the work, and humans don't scale like that. So all the compute demand we're already seeing is coming from companies in testing phase. The real wave hasn't even started. This whole AI supercycle is far from topping. From $NVDA / $AMD / $MU (and more), all the way to neoclouds, like $NBIS / $IREN / $CIFR / $WYFI (and more), everything will benefit.

@damnang2Jul 15, 2026 · 7:37 AM

After listening to $AEHR ’s latest earnings call, I’ve become even more convinced of my bullish view on $AMD.

@vulturetradesJul 14, 2026 · 11:27 PM

Tomorrow morning I’m loading up $200,000 in call options on a single ticker. My biggest play since $GME $OPEN $TSLA $NVDA $AMD Mark my words 99% of traders will miss this next $OPEN type mover. I WILL NOT and my my followers won’t either. I’m only sending this ticker to those who hit the “❤️” and comment “$SPY” (Must be following in order to DM)

@alphaticaioJul 14, 2026 · 8:06 PM

+$5.0B. Cool CPI day. $QQQ +$1.90B. $NVDA +$1.12B. Back to buying. $MU +$902M. $IWM -$1.11B. Small caps sold at scale. The mega-caps rallied. The small caps didn't. Q3 week three. $16.27B in buy blocks. $11.28B in sells. Net: +$5.0B. 15,980 prints. The buyers: $QQQ +$1.90B (1,528 blocks, $242M DP ALERT AT 1:06 PM) $NVDA +$1.12B (1,538 blocks, YEST: -$816M, FLIPPED BACK) $SPY +$943M (1,382 blocks) $MU +$902M (1,982 blocks, CONSISTENT BOTH WEEKS) $AAPL +$445M (398 blocks, YEST: -$819M, FLIPPED BACK) $META +$57M (512 blocks, YEST: -$1.36B, BARELY POS, PROFIT-TAKING OVER) The sellers: $IWM -$1.11B (159 blocks, SMALL CAPS SOLD AT SCALE) $AMD -$250M (873 blocks) $QQQ: +$1.90B. We alerted the $242M dark pool buy at 1:06 PM. Cool CPI sent the Nasdaq +1.26%. The lit tape confirms: $QQQ is the session's top buyer. $NVDA: +$1.12B. Yesterday: -$816M. Flipped back to buy. The CPI print resolved the positioning uncertainty. $NVDA back to accumulation. $AAPL: +$445M. Yesterday: -$819M. Flipped back. The July play resumed after one session of profit-taking. $META: +$57M. Barely positive. Yesterday: -$1.36B. The heavy profit-taking from yesterday subsided but the aggressive buying hasn't returned. The dark pool took $META from +$1.50B to -$1.36B to +$57M. The position reset. $IWM: -$1.11B. The outlier. On a day where $QQQ +$1.90B and $SPY +$943M, the Russell 2000 was sold -$1.11B. Both tapes agree sell (-$488M dark pool). The mega-caps rallied on cool CPI. The small caps didn't. The divergence between mega-cap and small-cap is the widest of the session. $MU: +$902M. The constant. Both weeks. Both tapes. MU continues to be accumulated regardless of the macro environment. Cool CPI resolved in favor of the buyers. The lit tape confirms. PPI tomorrow. Watching the tape.

@itsmichaelluuJul 13, 2026 · 9:27 PM

The entire AI trade is ON SALE. 24 stocks, avg –33% off their highs: 🔴 Oracle $ORCL –59.3% 🔴 IonQ $IONQ –49.4% 🔴 ServiceNow $NOW –49.2% 🔴 Applied Optoelectronics $AAOI –48.6% 🔴 IREN $IREN –46.5% 🔴 Rocket Lab $RKLB –46.3% 🔴 AST SpaceMobile $ASTS –45.2% 🔴 Applied Digital $APLD –38.6% 🔴 Palantir $PLTR –37.8% 🔴 Arm Holdings $ARM –33.9% 🔴 Marvell $MRVL –33.6% 🔴 Microsoft $MSFT –30.5% 🔴 Nebius $NBIS –29.0% 🔴 Western Digital $WDC –27.7% 🔴 Intel $INTC –27.5% 🔴 Micron $MU –26.1% 🔴 Lumentum $LITE –26.1% 🔴 SanDisk $SNDK –25.5% 🔴 Ouster $OUST –25.0% 🔴 Bloom Energy $BE –23.3% 🔴 Broadcom $AVGO –22.3% 🔴 Meta $META –16.1% 🔴 Nvidia $NVDA –13.2% 🔴 AMD $AMD –7.9% UPDATED cheatsheet for you to save

@alphaticaioJul 13, 2026 · 8:04 PM

DARK POOL SCANNER | July 13, 2026 +$1.48B net buy. The tapes disagree. Again. $SPY +$1.81B. The dark pool is buying what the lit tape sold. $META -$407M. Both agree sell. $SOXX -$177M. CPI tomorrow. Q3 week three. 360 off-exchange prints. +$1.48B net. NET BUY. Dark Pool Buyers: $SPY +$1.81B (50 prints, LIT: -$713M, DISAGREE, $191M ALERT) $MU +$438M (23 prints, BOTH AGREE BUY, BOTH WEEKS) $QQQ +$415M (52 prints, LIT: -$2.09B, DISAGREE) $MSFT +$239M (9 prints, BOTH AGREE BUY, OSCILLATION BACK) $AMD +$22M (9 prints) Dark Pool Sellers: $META -$407M (29 prints, BOTH AGREE SELL, $577→$660, DONE) $SOXX -$177M (10 prints, -3.7% ON SCANNER) The tapes disagree at the macro level. Lit tape: -$7.17B net sell. Dark pool: +$1.48B net buy. An $8.65B gap between the two tapes. $SPY: +$1.81B dark pool. Lit: -$713M. We alerted $191M in SPY buys at 2:46 PM. Two equal $95.5M prints. Nine seconds apart. One institution splitting a $191M order. The EOD confirms: the dark pool bought $SPY at the largest single-session level of the series. The lit tape is selling the index. The dark pool is buying it. $QQQ: +$415M dark pool. Lit: -$2.09B. We alerted the $178M QQQ sell at 12:43 PM. That was the lit tape. The dark pool disagrees at +$415M. $META: -$407M. Both tapes agree sell. The only name where both venues are aligned. We alerted the $105M sell at 1:44 PM. The profit-taking from $577 to $660 is complete on both tapes. The cycle: the dark pool bought at $577, we alerted it, the stock ran to $660, the dark pool sold. Both directions documented. $MU: +$438M dark pool. +$1.99B lit. Both tapes agree buy. The only name with cross-venue consensus on the buy side at scale. MU is the consensus long. $SOXX: -$177M dark pool. The rotation scanner showed $SOXX -3.7%. The dark pool is selling SOXX. Last week the $163M after-hours SOXX buy called the +5.1% bounce. Today: the dark pool is on the other side. Worth watching. The pattern from last week: The lit tape sold at series extremes. The dark pool bought. The dark pool was right 13/17 times. Today: the lit tape sold -$7.17B ahead of CPI. The dark pool bought +$1.48B. The same pattern. The same divergence. Tomorrow's CPI + bank earnings test whether the dark pool is right again. The data is positioned. The catalyst arrives tomorrow morning. Watching the tape.

@alphaticaioJul 13, 2026 · 8:03 PM

-$7.17B. The lit tape de-risked ahead of CPI. $META -$1.36B. Profit taken. $AAPL -$819M. $NVDA -$816M. Last week's winners sold. $MU +$1.99B. The only consistent buyer. Q3 week three. $15.45B in buy blocks. $22.62B in sells. Net: -$7.17B. 21,640 prints. The sellers: $QQQ -$2.09B (2,839 blocks, $178M DP SELL ALERT AT 12:43 PM) $META -$1.36B (886 blocks, LAST WEEK: +$1.50B, +14.81%, PROFIT TAKEN) $AAPL -$819M (795 blocks, 4 SESSIONS NEAR-ZERO SELLS, NOW SELLING) $NVDA -$816M (1,605 blocks, FRI: +$3.99B, NOW SELLING) $SPY -$713M (2,326 blocks) $AMD -$42M (1,100 blocks) The buyers: $MU +$1.99B (2,881 blocks, BOTH WEEKS, BOTH TAPES, CONSISTENT) $MSFT +$785M (605 blocks, FLIPPED, 6 SESSIONS ALTERNATING) Last week: three consecutive +$13B sessions. $46.8B net buy. The largest buying run of the series. Today: -$7.17B. The institutions took profit on last week's winners ahead of CPI + bank earnings. $META: -$1.36B. Last week: +$1.50B dark pool buy. +14.81% return. We alerted the $255M buy Thursday. Receipts posted Friday. Today: -$1.36B. The cycle completed. $577 to $660. Profit taken. $NVDA: -$816M. Friday: +$3.99B with $6.5M sells. We alerted the $326M dark pool buy. Today: selling. The name that had near-zero opposition on Friday now faces selling pressure ahead of CPI. $MU: +$1.99B. The exception. Both weeks accumulated. The $200M pre-market alert from last Wednesday continues to validate. The only mega-cap with consistent buying across both the sell week and the CPI positioning day. $MSFT: +$785M. Six sessions alternating: +$2.02B, -$503M, +$226M, -$731M, -$305M, +$785M. The oscillation continues. The lit tape sold -$7.17B ahead of CPI. But the dark pool tells a different story. +$1.48B net buy. The tapes disagree again. Last week when the tapes disagreed, the dark pool was right 13/17 times. CPI tomorrow. Watching the tape.

@midascabalJul 13, 2026 · 6:22 PM

Here are the top stocks I found by using @asymmetrix_xyz, the platform that can spot AI bottlenecks in seconds: $OUST $MP $QCOM $AMD $TER Follow me for more.

@investmentguru_Jul 13, 2026 · 10:33 AM

$MU $AMD $CRDO $MRVL $AAOI $AXTI $COHR $INTC $SNDK $NBIS AI infrastructure stocks are under pressure in the premarket as the sector goes through a broad rotation and consolidation. Nothing goes straight up forever. Pullbacks are a normal part of healthy market trends and can create opportunities for long-term investors to gradually build positions in high-conviction names. Patience often pays more than chasing strength. NOT A FINANCIAL ADVICE

@alphaticaioJul 12, 2026 · 2:19 PM

WEEKLY DARK POOL RECAP | Week 17 | Q3 Week Two Our classification algorithm called 13/17 correct. 76.5%. Top 10 by conviction: 10/10. $24.2B net buy. The dark pool bought the worst internals of the series and was right. Jul 2 close to Jul 10 close. 17 names tracked. DARK POOL BUYS: $NVDA +$7.34B → +8.28% ✅ (4 sessions, zero DP sells) $AMD +$5.58B → +7.74% ✅ (4 sessions) $QQQ +$4.69B → +1.81% ✅ (5 sessions) $AAPL +$2.90B → +2.17% ✅ (3 sessions, near-zero sells) $MU +$2.44B → +0.38% ✅ (5 sessions, $200M pre-market alert) $AVGO +$1.53B → +10.96% ✅ (2 sessions, zero sells) $META +$1.50B → +14.81% ✅ (5 sessions, $255M alert, receipts) $SOXX +$628M → +2.65% ✅ (3 sessions, $163M AH alert) $SPY +$290M → +1.37% ✅ (5 sessions) $AMZN +$271M → +1.10% ✅ (1 session) $IWM +$219M → -0.53% ❌ $TLT +$184M → -1.22% ❌ $GLD +$84M → -0.30% ❌ $GOOGL +$76M → -0.76% ❌ DARK POOL SELLS: $PG -$1.48B → -2.89% ✅ (2 sessions) $MSFT -$677M → -1.38% ✅ (4 sessions) $TSM -$376M → -0.01% ✅ (2 sessions) 13/17. 76.5%. The top 10 by conviction: 10/10. Perfect. The pattern: when the dark pool has high conviction, it is right. Every position above $270M was correct. Every miss was a small position under $220M with a return within -1.3%. The misses were $IWM (-0.53%), $TLT (-1.22%), $GLD (-0.30%), $GOOGL (-0.76%). Small positions. Small losses. The conviction calls delivered. The best call: $META. +$1.50B dark pool buy. +14.81% return. We alerted the $255M buy at 3:41 PM Thursday. $META +4% pre-market Friday. Receipts posted. The semi trade: $NVDA (+$7.34B, +8.28%), $AMD (+$5.58B, +7.74%), $AVGO (+$1.53B, +10.96%), $SOXX (+$628M, +2.65%). Four semi names. All bought. All correct. Combined: $15.0B in dark pool buying on semis. The institutional thesis was clear and it was right. The sell call: $PG (-$1.48B, -2.89%), $MSFT (-$677M, -1.38%), $TSM (-$376M, -0.01%). All three correct. $PG was the only high-conviction sell. The dark pool sold Procter & Gamble at $1.48B and the stock dropped 2.89%. The week's arc: Monday: Internals balanced. Dark pool nearly flat (+$141M). Tuesday: Volume 0.42x. Series low. Dark pool: +$472M. Wednesday: Breadth 0.42x. VWAP 27%. All factors negative. Worst of series. Dark pool: +$2.54B. Thursday: Bounce. +$18.44B lit. +$983M dark pool. Every alert validated. Friday: Indexes green, internals red. Mask returned. Dark pool: +$3.29B. New series record. The dark pool bought the sell. The dark pool bought the bottom. The dark pool bought the bounce. The dark pool bought through the mask. Four consecutive buy sessions. $7.28B combined. The bounce was called before it arrived. The data was published in real time. The receipts are on the record. Week 17. Q3 week two. 13/17. 76.5%. Top 10: 10/10. Bank earnings this week. The $299M $JPM dark pool sell from Wednesday is on the record. Watching the tape. $META $NVDA $SPY

@stocksavvyshayJul 12, 2026 · 1:45 PM

THE MOST PROFITABLE AI SEMICONDUCTOR STOCKS Operating margin breakdown: • Best-in-class (50%+) | $MU, $NVDA, $SKHY, $TSM • Elite (40–49%) | $ANET, $AVGO, $KLAC, $SNDK • Strong (30–39%) | $ASML, $LRCX, $CRDO, $WDC • Healthy (20–29%) | $AMAT, $ALAB • Solid (Below 20%) | $ARM, $MRVL, $AMD, $COHR, $LITE, $ON

@aleabitoredditJul 12, 2026 · 12:37 PM

@softmaxedx My friend at $AMD tripped on a rock and had to spend a few hours at home. So we’ll likely see a major delay at AMD’s entire GPU program.

@itsmichaelluuJul 11, 2026 · 2:11 PM

$DRAM — the first-ever memory ETF. Everyone spent this whole AI run fighting over which stock to own $NVDA, $MU, $SNDK, $AVGO, $AMD, $AAPL, $META, $MSFT, $GOOG, $TSLA. Here's what none of them run without: memory. HBM is sold out. $DRAM contract prices ripped ~90% in a single quarter. Memory is THE bottleneck of the entire AI buildout and the hardest part to scale. You don't have to pick the winner of the AI race. You just have to own what every winner is desperate to buy. $DRAM packages the whole memory bottleneck $MU, $SNDK, SK Hynix, Samsung into one ticker.

@stocksavvyshayJul 11, 2026 · 1:57 PM

KEY LEVELS TO WATCH FOR POPULAR STOCKS • Bullish: $NVDA, $AMD, $MU, $HOOD, $HIMS • Hold: $TSLA, $AVGO, $NBIS, $AEHR, $CIFR, $RKLB, $OPEN, $SOFI, $AAOI • Bearish: $PLTR, $ONDS, $OKLO, $ASTS, $IONQ, $IREN https://t.co/aGEiqUTUys

@alphaticaioJul 10, 2026 · 9:05 PM

DARK POOL SCANNER | July 10, 2026 +$3.29B. New largest dark pool buy of the series. $NVDA +$1.34B zero sells. Both tapes agree. $AMD +$1.30B. $QQQ +$1.32B. $SPY -$768M. Tapes disagree. Again. Q3 week two closes. 364 off-exchange prints. +$3.29B net. NET BUY. Dark Pool Buyers: $NVDA +$1.34B (25 prints, $0 DP SELLS, ZERO, BOTH AGREE, ALERT VALIDATED) $QQQ +$1.32B (43 prints) $AMD +$1.30B (13 prints) $MU +$213M (44 prints) $SOXX +$157M (9 prints) $IWM +$109M (19 prints) $MSFT +$88M (4 prints, LIT: -$305M, TAPES DISAGREE) Dark Pool Sellers: $SPY -$768M (40 prints, LIT: +$3.70B, TAPES DISAGREE) $TSM -$188M (3 prints, BOTH AGREE SELL) $META -$139M (44 prints, LIT: +$998M, TAPES DISAGREE) +$3.29B. New series record. The prior record was +$2.54B on Wednesday when the internals hit series extremes. Today: +$3.29B on a day where the indexes are green but the internals are red (volume 0.46x, breadth 0.72x). The dark pool is buying through the mask. $NVDA: +$1.34B dark pool. Zero sells. Not one. Lit: +$3.99B with $6.5M sells. Combined: $5.33B. Both tapes agree at maximum conviction. We posted the $326M dark pool buy at 11:02 AM. "Someone picked a side." The EOD confirms: $5.33B combined across both tapes. Zero dark pool sells. The strongest single-name cross-venue signal the series has recorded. $SPY: -$768M dark pool. Lit: +$3.70B. Tapes disagree. The $SPY tapes-disagree pattern persists. The lit tape buys $SPY at scale. The dark pool sells. This has been the Q3 theme. $META: -$139M dark pool. Lit: +$998M. Yesterday: +$491M zero dark pool sells. Today: -$139M. The $META dark pool consensus broke on day two. The lit tape still buys. The dark pool flipped. $MSFT: +$88M dark pool. Lit: -$305M. Tapes disagree in the other direction. The dark pool is buying what the lit tape is selling. Fourth consecutive dark pool buy session: +$472M, +$2.54B, +$983M, +$3.29B. Four days. $7.28B combined. The dark pool bought the sell, bought the bottom, bought the bounce, and bought the mask day. The positioning is clear. Week 2 dark pool arc: Mon: +$141M. Nearly flat. Tue: +$472M. Volume series low. Dark pool buys. Wed: +$2.54B. Breadth series low. Dark pool buys harder. Thu: +$983M. Bounce. Dark pool buys the bounce. Fri: +$3.29B. New record. Indexes green, internals red. Dark pool buys through the mask. $7.28B net dark pool buying in four sessions. The data documented every step. Watching the tape.

@alphaticaioJul 10, 2026 · 8:37 PM

+$14.98B. Third consecutive session above +$13B. $NVDA +$3.99B. $6.5M sells. The $326M alert validated. $AMZN -$854M. $19.6M buys. Near zero. The tape picked winners and losers. Q3 week two closes. $26.85B in buy blocks. $11.87B in sells. Net: +$14.98B. 21,665 prints. The buyers: $NVDA +$3.99B (2,120 blocks, $6.5M SELLS, NEAR ZERO, ALERT VALIDATED) $SPY +$3.70B (2,681 blocks) $QQQ +$3.12B (2,356 blocks) $MU +$1.99B (2,851 blocks) $AMD +$1.92B (1,114 blocks) $META +$998M (2,124 blocks) $GLD +$342M (160 blocks, GOLD BOUGHT ON LIT) $AVGO +$232M (345 blocks) $AAPL +$148M (484 blocks) The sellers: $AMZN -$854M (388 blocks, $19.6M BUYS, NEAR ZERO) $MSFT -$305M (432 blocks, 2ND CONSECUTIVE SELL) $TSM -$126M (183 blocks) $GOOGL -$94M (288 blocks) $NVDA: +$3.99B with $6.5M sells. We posted the $326M dark pool buy at 11:02 AM. "Someone picked a side." The lit tape confirms: $3.99B in buying with virtually zero opposition. $NVDA was perfectly balanced yesterday at -$11M. Today the balance broke decisively. $AMZN: -$854M with $19.6M buys. Near zero. The opposite of $NVDA. Amazon sold with virtually no buying. On a day where the lit tape was +$14.98B net buy with seven names green, Amazon was sold with zero opposition. $MSFT: -$305M. Second consecutive sell session. The MSFT oscillation: +$2.02B, -$503M, +$226M, -$731M, -$305M. Two consecutive sells. The distribution pattern persists. Three-day lit tape totals: +$13.37B, +$18.44B, +$14.98B. Combined: $46.8B. The institutions moved $46.8B net into the market over three sessions. The sell from Tuesday and Wednesday is being bought back at nearly double the pace. $GLD: +$342M. Gold bought on the lit tape. After the $95M dark pool sell this morning and the precious metals crash on Wednesday, gold is being accumulated. The dark pool sold gold. The lit tape bought it. The tape picks winners and losers. $NVDA and $AMZN. Same market. Same session. Opposite outcomes. Watching the tape.

@evergreencap3Jul 10, 2026 · 5:29 PM

With the emergence of $SPCX and $META, the Western frontier is on the verge of expanding from 3 to 5 players. This is a phenomenal development that should have a meaningfully positive impact on the size, duration, and sustainability of the AI ecosystem more broadly. A more diversified frontier will lead to higher levels of innovation, healthy competitive intensity, and a larger, more durable end-market. The accelerating rate of frontier capabilities will only intensify, as will the revenue generation, along with the magnitude of investment dollars poured back into underlying AI infrastructure, which for all intents and purposes is still within the early portion of the S-Curve. While Anthropic and OpenAI may have preferred the dynamic to have remained status quo, particularly given the potential implications for near-term profitability, the size of this market will grow over time into the many tens of trillions of dollars, with ample room for all 5 players. Over the very long term, I expect multiple trillion-dollar-plus revenue lines, at healthy margins, to emerge across this cohort. I believe this crystallizes a multiyear investment framework going forward, which is underpinned by owning both the frontier model players and the AI infrastructure companies that supply them. To me these are the most obvious cohorts to continue seeing outsized benefits from GenAI. I believe the Big 5 will both generate more revenue and profits than ever, and invest in more AI infrastructure than ever, creating a highly attractive two-sided ecosystem. The market to date has already demonstrated willingness to reward the frontier model winners (Anthropic, OpenAI and Google have all seen extraordinary increases in equity valuation), along with the AI infrastructure providers. So this is not entirely new. But the expansion from 3 to 5 is so overwhelmingly positive for the AI infrastructure cohort, in addition to creating a healthier frontier AI ecosystem, that the sustainability of this trend is becoming much more solid and tangible across a multiyear horizon. The world will undoubtedly change and the pace of change will accelerate. There will be many more calls for an AI bubble, that capex is peaking, and that the trade is over. I expect zero of these will come to fruition. And I expect that both the value created at and invested into the frontier will march forward on an increasingly healthy trajectory, and that this will provide a steady and reliable core investment framework for many years to come. $Anthropic $OpenAI $GOOGL $META $SPCX $NVDA $AVGO $AMD $TSM $INTC $Samsung $MU $SKHYV $ARM $ASML $AMAT $KLAC $GLW $CBRS $LITE

@landoinvestsJul 10, 2026 · 4:23 PM

They’ll say it was “obvious” in the end. Everyone says that after the move already happens They said it with $MU, $SNDK, $NBIS, $MRVL, $AMD, etc I know people are extremely concerned with short term price action when it comes to $IREN. I am not. I’m an investor, with a very long time horizon compared to most. I believe this company can become a 100-200B MC company within the next few years Don’t let your emotions get the best of you. It does nothing to sulk, whine, or complain. Absolutely nothing. So don’t. Thesis/Theme didn’t change? Buy the fear

@aleabitoredditJul 10, 2026 · 11:50 AM

@stephengluck I blocked them since they're known for spreading disinformation. Especially after it was exposed around their short seller activity, that there were marketing accounts telling people to sell. These accounts had histories of promoting the same Asian crypto launches. - They spread defamation like saying I used AI to fabricate reports on CW laser LTAs from $AMD, despite the report coming directly from Trendforce. Then they tried to hide references to the actual source after they were proven wrong. - I didn't sell a good chunk of my $SIVE position, I own 1M+ shares (not USD) as one of the largest individual shareholders. Don't see a reason to exit my position since I'm trying to own as much as possible before 1.6T pluggables and CPO architectural inflections hit. Better not to give malicious accounts additional views.

@wliangJul 10, 2026 · 7:41 AM

KOSPI is finally green… Foreign investors were net buyers two days straight now, institutions stepped in hard too. Also nice to see Samsung start to stabilize after the earnings-beat volatility. Not saying it’s all-clear now but when the earnings selloff stops, it’s usually a good sign. (Same thing we saw with $SNDK and $AMD earnings months ago. Turbulence into an aggressive rally) Fundamentals never changed. Sentiment just had to catch up.

@investmentguru_Jul 9, 2026 · 11:58 PM

Bullish Calls Return to AI Semis: $MU, $SNDK, $AMD, $AVGO & $NVDA

@alphaticaioJul 9, 2026 · 8:05 PM

+$18.44B. Largest net buy of the series. $SPY +$3.66B. $AAPL +$2.78B near-zero sells. $META +$1.43B. $MU +$1.74B. Eight buyers. Zero sellers. The institutions bought everything. Q3 week two. $30.72B in buy blocks. $12.28B in sells. Net flow: +$18.44B. 23,976 prints. $SPY +$3.66B (2,271 blocks) $AAPL +$2.78B (505 blocks, $46M SELLS, NEAR ZERO, JULY 10/10, 4TH SESSION) $QQQ +$2.01B (2,803 blocks) $MU +$1.74B (4,590 blocks, $200M PRE-MARKET ALERT VALIDATES) $META +$1.43B (1,183 blocks, $255M ALERT AT 3:41 PM, BOTH AGREE) $AMD +$1.02B (1,261 blocks) $AVGO +$778M (626 blocks, $208M AH ALERT CONTINUES) $NVDA -$11M (1,963 blocks, FLAT, $1.11B BUY VS $1.13B SELL) Eight names on the board. Zero sellers. Not one red name on the lit tape. $NVDA at -$11M is perfectly balanced, not a sell. +$18.44B. The prior record was +$13.37B yesterday. Two consecutive sessions of record-breaking buying. The institutions didn't just buy the bounce. They bought it harder than the sell. $AAPL: +$2.78B. $46M sells. Fourth session. We flagged the $51M dark pool buy on Q3 day one. Combined four-day total: $7.85B in $AAPL lit tape buying with near-zero sells. The July play is executing at scale. $META: +$1.43B. We alerted the $255M dark pool buy at 3:41 PM. Both tapes agree. +$491M in the dark pool with zero sells. The cleanest cross-venue signal on the board. $MU: +$1.74B on the lit tape. But the dark pool sold -$1.19B. The biggest tapes-disagree of the series. The lit and dark are on opposite sides of $MU. The week's arc on the lit tape: sell to series extremes, +$13.37B, +$18.44B. The buying accelerated. The bounce exceeded the sell. Every alert validated. Watching the tape.

@midascabalJul 9, 2026 · 7:10 PM

Start at the top. The GPU layer. $NVDA B300s. $AMD MI300Xs. This is where every retail investor is staring. Fully priced. Fully covered. Nothing left on the table. Everyone sees this layer. Keep going.

@stocksavvyshayJul 9, 2026 · 5:33 PM

$AMD and 5C are partnering to build gigascale AI factories for training, inference and agentic AI workloads. Initial deployments are underway in Ohio and Memphis with the focus on integrating AMD compute with 5C’s power, cooling, networking and operations infrastructure. https://t.co/1GaFXpXpCh

@thevalueistJul 9, 2026 · 3:34 PM

$AMD AMD: Scaling as a Premier AI Infrastructure and Compute Platform. Investment Thesis. New: 7/9/26. AMD has executed the most consequential competitive repositioning in the semiconductor industry over the past decade, and the current strategic position reflects that execution quality. Lisa Su's leadership has transformed a company that was existentially threatened in both CPU and GPU markets into a credible dual-threat competitor across data center compute — EPYC server CPUs have taken meaningful share from Intel in a market that Intel treated as structurally protected, and Instinct GPUs have established AMD as the only scaled alternative to NVIDIA in AI accelerator infrastructure. Both achievements required sustained engineering excellence and commercial execution over multiple product generations. The data center revenue mix shift is the financial expression of that repositioning. High-value server CPU and GPU revenue carries structurally better margins and growth characteristics than the consumer and embedded markets that AMD historically depended on, and the trajectory of that mix shift is the primary driver of the earnings quality improvement the valuation reflects. Hyperscaler partnerships with Meta and OpenAI are the demand anchors that validate the Instinct GPU roadmap at production scale — these are not evaluation relationships but deployed infrastructure commitments that signal AMD has cleared the technical and operational qualification bar for the most demanding AI infrastructure customers. ROCm software maturity is the honest gap in the competitive position that the hardware progress doesn't fully close. NVIDIA's CUDA ecosystem represents decades of developer investment, tooling depth, and optimization infrastructure that AMD's ROCm framework is still working to match. Hardware performance parity or leadership on specific benchmarks is a necessary but insufficient condition for taking sustained share — the software ecosystem is what determines whether customers can deploy AMD hardware without incurring the migration friction that keeps them on NVIDIA even when the hardware economics favor a switch. ROCm improvement is happening, but the pace relative to the valuation's implied market share trajectory is the execution variable that deserves the most scrutiny. Supply chain constraints are the near-term operational limitation that demand momentum alone cannot resolve. Advanced packaging capacity — CoWoS and equivalent configurations for HBM integration — is a binding constraint that affects AMD and NVIDIA simultaneously, but AMD's ability to fulfill accelerating demand depends on securing adequate packaging allocation in a supply environment where TSMC's capacity is the shared bottleneck. That constraint is temporary but real, and it creates revenue timing risk that the strong design win pipeline doesn't eliminate. The execution compounder framing is accurate and the track record justifies the characterization — AMD has delivered on ambitious roadmap commitments across multiple successive product generations, which is a demonstrated organizational capability rather than an aspiration. The valuation demands that the execution consistency continues through a more complex product environment, a more competitive landscape as NVIDIA invests aggressively to defend its position, and a software ecosystem catch-up that requires sustained investment without near-term financial payoff. The competitive position is real, the growth trajectory is credible, and the leadership has earned the benefit of the doubt — but the current price requires that all of those elements perform simultaneously without meaningful interruption.

@stocksavvyshayJul 9, 2026 · 1:50 PM

Semiconductor stocks are getting bought aggressively today: • $AEHR +14% • $LITE +13% • $KLAC +12% • $ARM +11% • $AMAT +10% • $AMKR +10% • $AAOI +9% • $ALAB +9% • $MU +7% • $MRVL +7% • $COHR +6% • $AMD +6% • $SNDK +6% • $INTC +5% • $ASML +5%

@investmentguru_Jul 9, 2026 · 1:06 PM

MORNING MOVERS Semis roaring back — AI demand + easing geopolitical fears lighting a fire under the whole space: GAPPING UP $MU +6.7% — $3B US supply-chain investment $KLAC +5.8% — MS upgrade to Buy, TD Cowen PT $260 $AMAT +5.7% — analyst hikes + China easing H200 curbs $SNDK +5.5% — Wedbush PT $2000, Outperform $LRCX +5.4% — MS reiterates Buy, PT $404 $INTC +5.0% — semis ETF inflows $ARM +4.3% — sector strength $AVGO +3.7% — $30B Apple custom chip deal thru 2031 $AMD +3.6% — GS reiterates Buy, PT $640 $GEV +3.4% — SMR deal w/ Hitachi (Europe) $MARA +4.0% — Texas powered facility acquisition (HIF) GAPPING DOWN $AZN -7.7% — Wainua Phase 3 miss (cardio endpoint) $IBM -4.5% / $MSFT -1.9% — Starbucks building in-house AI, cutting reliance $CRM -4.4% — KeyBanc downgrade, Agentforce skepticism $NOW -4.2% — GS cuts PT to $145 ahead of Q2 $META -3.8% — in-house Iris AI chip prod. starts Sept, 14GW compute by 2027 $ADBE -3.3% — BofA downgrade to Underperform, GenAI competition risk $FIG -3.1% — JPM cuts PT to $28 Semiconductor supply chain (WFE, memory, foundry) getting a broad-based bid — this lines up with the Silicon-to-Substation thesis. Enterprise SaaS ($CRM, $NOW, $ADBE) facing real “AI eats software” pressure as hyperscalers build in-house.

@amitisinvestingJul 9, 2026 · 11:59 AM

WHY ARE THE SEMIS UP THIS MORNING? Well…a hyperscaler called $META just reminded the market how much they are willing to spend on capex: - Meta plans to double its AI computing capacity to 14GW in 2027 and begin manufacturing its in-house Iris AI chip in September, according to Reuters citing an internal memo. - The Iris chip, part of Meta’s MTIA program, is designed with $AVGO Broadcom and manufactured by $TSM TSMC, with the goal of reducing reliance on $NVDA Nvidia and $AMD GPUs. - Meta has also secured long-term supply agreements with Samsung (memory), $SNDK SanDisk (flash storage), and Sumitomo Electric (fiber-optic equipment). Once again, the market remembers how much Big Tech is willing to spend and how serious they are about creating their own custom silicon. Trump saying that Iran wants a deal last night + META getting ready to spend even more on capex is likely why the semis are starting to recover.

@investmentguru_Jul 9, 2026 · 11:28 AM

Can Anthropic Single-Handedly Save the Entire AI Industry Chain? SemiAnalysis just dropped numbers that should reframe how you’re thinking about the AI trade. Anthropic is tracking toward $1B in GAAP EBITDA by Q3 2026 (~6% margin) — its first real profitability inflection. ARR has gone from $9B (end of 2025) to $60B+ today. If the ~$15B/month net-new-ARR pace holds, SemiAnalysis models $300B ARR by end-2027 → a $6T enterprise value. Anthropic confidentially submitted its draft S-1 to the SEC on June 1. That’s not rumor — it’s confirmed. The race with OpenAI to IPO first is now officially on. Why this matters more than another “AI stock is expensive” headline: Anthropic’s revenue mix is the story. 75-85% of ARR is usage-based API (no per-user revenue ceiling), vs. OpenAI’s subscription-heavy model (~65% subscription, ~40% consumer ARR). Claude Code alone is now >7% of all GitHub commits — that’s the growth engine. Blended gross margin has gone from -94% (2024) to mid-60s now, with API margins above 80%. Reported NRR: 500%. Who rides the wave if this plays out: 🔹 Compute: $NVDA $AMD $AVGO $MRVL 🔹 Hyperscalers (TaaS distribution, now 15-20% of ARR): $AMZN $MSFT 🔹 Dev tooling / workflow: $TEAM $DOCN $FROG 🔹 Data/observability SaaS: $DDOG $MDB $SNOW 🔹 Edge/CDN: $NET $AKAM $FSLY these are SemiAnalysis’s bottom-up modeled estimates on a private company that hasn’t disclosed audited financials — Anthropic itself hasn’t confirmed these specific figures publicly. Treat the trajectory as directionally credible, the $6T number as a scenario, not a forecast. Not financial advice.

@alphaticaioJul 8, 2026 · 8:55 PM

DARK POOL SCANNER | July 8, 2026 +$2.54B. Largest dark pool buy of the series. $SPY +$868M. Second consecutive massive buy on a sell day. $QQQ +$832M. $TSM +$216M zero sells. Every alert validated. Q3 week two. 374 off-exchange prints. +$2.54B net. NET BUY. Dark Pool Buyers: $SPY +$868M (38 prints, 2ND CONSECUTIVE SELL-DAY BUY) $QQQ +$832M (69 prints) $TLT +$232M (10 prints, BUYING THE BOND DIP) $TSM +$216M (3 prints, $0 DP SELLS, ZERO) $NVDA +$163M (16 prints, BOTH AGREE BUY) $MU +$160M (37 prints, PRE-MARKET + SESSION) $IWM +$110M (16 prints) Dark Pool Sellers: $AMD -$119M (7 prints, LIT: +$1.68B, TAPES DISAGREE) +$2.54B. Largest dark pool buy of the series. Yesterday: +$472M. Five times larger. The dark pool bought the worst internals of the series for the second consecutive session. Yesterday: breadth 0.60x, volume 0.42x (series low). Dark pool: +$472M. Today: breadth 0.42x (series low), VWAP 27% (distribution). Dark pool: +$2.54B. The worse the internals get, the harder the dark pool buys. That's the pattern. $SPY: +$868M. Second consecutive session. Yesterday: +$784M. Two days. $1.65B combined. Both on the worst internals the series has recorded. The $SPY dark pool buy-the-dip pattern is established. $TLT: +$232M. The dark pool is buying bonds. The rotation scanner showed $TLT -0.5% and all four factors negative. The dark pool is buying what the scanner says is selling. When the dark pool buys $TLT alongside $SPY and $QQQ, the positioning is: this sell is temporary. $TSM: +$216M. Zero sells. Taiwan Semi accumulated with zero dark pool opposition. The semi supply chain accumulation across both tapes: $MU (+$4.39B lit, +$160M dark), $NVDA (+$1.89B lit, +$163M dark), $AVGO (+$1.23B lit, zero sells), $TSM (+$350M lit, +$216M dark, zero DP sells). Four semi names. Both tapes. The institutional thesis is clear. $AMD: -$119M dark pool. Lit: +$1.68B. Tapes disagree. The dark pool sold $AMD while the lit tape bought $1.68B. This is the only name where the tapes disagree today. Today's alerts, in order: 9:06 AM: $200M $MU pre-market dark pool buy. Four prints. Same second. Result: $MU +$4.39B lit, +$160M dark. Validated. 12:45 PM: VWAP velocity +13pts accumulation surge. Result: Market recovered off distribution lows. Validated. 1:08 PM: Five consecutive dark pool buys. Zero sells. Result: +$2.54B dark pool EOD. Validated. Three alerts. Three validated. The dark pool bought the sell day. The pattern is established. The data names the names. Watching the tape.

@mrmikeinvestingJul 8, 2026 · 8:34 PM

In 2024 I called $NVDA at $47. Then In 2025 I called $AMD at $76. & Finally In 2026 I called $NBIS at $73. Now these 5 stocks have the exact setups $NVDA, $AMD, & $NBIS had. 1. CoreWeave ~ $CRWV 2. One Stop Sys ~ $OSS 3. Zeta Global ~ $ZETA 4. Rocket Lab ~ $RKLB 5. AST SpaceMobile ~ $ASTS Quite literal life changing opportunities await in front of you. Don’t miss out on these…

@alphaticaioJul 8, 2026 · 8:05 PM

+$13.37B. The institutions bought the worst internals of the series. $MU +$4.39B. Pre-market alert validated. $AVGO +$1.23B zero sells. $NVDA +$1.89B. Seven buyers. One seller. Q3 week two. $26.40B in buy blocks. $13.03B in sells. Net flow: +$13.37B. 23,230 prints. The buyers: $MU +$4.39B (3,535 blocks, $200M PRE-MARKET DP BUY VALIDATED) $SPY +$2.44B (2,839 blocks) $QQQ +$2.21B (2,619 blocks) $NVDA +$1.89B (2,082 blocks) $AMD +$1.68B (1,141 blocks) $AVGO +$1.23B (788 blocks, $0 SELLS, ZERO) $TSM +$350M (233 blocks) The seller: $MSFT -$731M (453 blocks, 4 SESSIONS: +$2.02B, -$503M, +$226M, -$731M) This morning we posted: breadth 0.42x, worst of the series. VWAP 27.0%, distribution. Volume 0.49x. All four factors negative for the first time. 10.1% spread, widest of the series. Every metric at series extremes. The tape: +$13.37B net buy. Seven buyers. One seller. $MU: +$4.39B. We alerted $200M in dark pool buys pre-market at 9:06 AM. Four prints. Same second. $50M each. One institution splitting a $200M order. The pre-market alert called the session. $MU closed as the top buyer on the board. $AVGO: +$1.23B. Zero sell blocks. Not one. Broadcom accumulated with zero opposition on the worst internals day of Q3. The semi accumulation was broad: $MU, $NVDA, $AMD, $AVGO, $TSM. Five semi and semi-adjacent names. All bought. The institutions used the sell day to load up on semis. $MSFT: -$731M. The only seller on the board. The MSFT oscillation: +$2.02B, -$503M, +$226M, -$731M. Four sessions. Alternating buy/sell. The distribution pattern hasn't resolved. Every other name on the board was buy. The institutions bought the worst internals of the series. The surface said sell everything. The blocks said buy everything except $MSFT. The data is the data. Watching the tape.

@stocksavvyshayJul 8, 2026 · 3:36 PM

FUTURUM EQUITIES PODCAST EPISODE #48 Daniel & I will be live today at 1PM ET exclusively on the Futurum Equities YouTube channel covering: • Is the memory trade over for $MU, $SNDK & $SKHY after one of the strongest runs in the market? • Our updated thoughts on $NVDA, $META, $MSFT, $GOOGL, $TSLA, $AAPL & $AMZN • Is software’s nearly 50% rally off April lows a dead cat bounce or real regime change for $PLTR, $SNOW, $MDB & $ZETA? • What could save AI infrastructure drawdown across $NVDA, $AMD, $ASML, $TSM & $AVGO? • The biggest themes we are watching as the market separates durable AI winners from crowded trades We’ll leave the last 15 minutes for live Q&A so bring your questions!

@stocksavvyshayJul 8, 2026 · 2:25 PM

$TSM is one of the cleanest ways to own the AI buildout because $NVDA, $AMD, $AAPL, $AVGO, $QCOM and hyperscaler custom silicon programs all depend on the same leading-edge manufacturing layer. The real bottleneck is who has access to scarce wafer starts and TSMC remains the physical chokepoint behind almost every version of the AI trade. With N3 capacity reportedly scaling to ~220K wafers per month, the “second source” everyone wants still doesn't really exist which means every token ultimately pays an invisible royalty to TSMC.

@stocksavvyshayJul 8, 2026 · 2:09 PM

@FuturumEquities $TSM remains one of the cleanest ways to own the AI trade because it benefits from almost every version of the debate investors are having. Whether market favors $NVDA or $AMD accelerators or $AVGO XPUs, training or inference, TSMC is the AI factory powering the entire stack.

@yasutaketinJul 8, 2026 · 5:57 AM

エージェントAI需要期待で比較的まだ堅調な $AMD $INTC のデータセンター向けCPUサーバーも、メモリーは大量に必要。実運用の中心レンジ •一般的なクラウド/仮想化/DBサーバー：512GB〜1.5TB •高密度仮想化・大規模DB・分析系：1.5TB〜3TB •メモリ重視のHPC/大規模分析/インメモリDB：3TB〜6TB •AI GPUサーバーのホストCPU側メモリ：768GB〜1.5TB程度も多いが、用途次第で3TB以上もあり得る

@benitozJul 8, 2026 · 4:53 AM

B200 is sold out at @LambdaAPI and @Vultr on-demand, forcing buyers into preemptible at $3.20–$3.95 — a 45–53% discount to list. This scarcity premium mirrors the 2% util-avail signal and confirms B200 is the GPU market's bottleneck Live: https://t.co/JnGVTfO4tB $NVDA $AMD https://t.co/GvroKViVGb

@alphaticaioJul 7, 2026 · 10:17 PM

-$2.01B. $NVDA +$1.19B near-zero sells. $TSLA -$1.08B zero buys. $META +$707M. The tape split on a sell day. The institutions picked names. Q3 week two. $14.38B in buy blocks. $16.39B in sells. Net flow: -$2.01B. 18,088 prints. The buyers: $NVDA +$1.19B (1,246 blocks, $119M SELLS, NEAR ZERO) $META +$707M (545 blocks, BOTH TAPES AGREE BUY) $SPY +$335M (1,744 blocks) $MSFT +$226M (483 blocks, FLIPPED BACK TO BUY) $AMD +$115M (794 blocks) $MU +$100M (3,433 blocks, BARELY POS, CONTESTED) The sellers: $TSLA -$1.08B (575 blocks, $0 BUYS, ZERO) $QQQ -$389M (2,315 blocks) Volume 0.42x. New series low on the internals. The sell was the most extreme of Q3. But the institutions used it. $NVDA: +$1.19B with $119M in sells. Near-zero opposition. While volume ratio hit the lowest reading of the series, the institutions were quietly accumulating NVIDIA. The crowd sells. The blocks buy. This is what institutional buying on weakness looks like. $TSLA: -$1.08B with zero buys. Not one buy block. $0. Same sell day as $NVDA. Two names. Two opposite outcomes. $NVDA bought with no opposition. $TSLA sold with no opposition. The tape is surgical. $META: +$707M on the lit tape. +$480M on the dark pool. Both tapes agree buy. $28M in dark pool sells. Near zero. The cleanest cross-venue signal of the session. $MSFT: +$226M. Flipped back to buy after yesterday's -$503M sell. The MSFT oscillation: +$2.02B, -$503M, +$226M. Three sessions. Three different directions. $MU: +$100M on 3,433 blocks. Barely positive. $1.94B buy vs $1.84B sell. The MU battle is the most contested name on the board. After three consecutive days of selling, the first positive session. The post-earnings reversal may have found a floor. Net: -$2.01B. The tape was sell. But the name-level data tells a different story. $NVDA bought. $TSLA sold. $META accumulated. The institutions used the sell day to position. The tape doesn't have an opinion. It has a receipt.

@itsmichaelluuJul 7, 2026 · 4:31 PM

200%-500% news triggers: $MU — HBM4 ramp for $NVDA Rubin and tightening DRAM supply drive memory pricing power higher → ~late Sept 2026 (FQ4 earnings); HBM4 ramp runs H2 2026 $SNDK — Enterprise SSD demand and NAND price recovery drive first earnings beats post-Western Digital spinoff → ~early Aug 2026 (Q2 earnings) $WDC — Nearline HDD demand surges as AI datacenters need massive storage capacity for training data → ~late July/early Aug 2026 (FQ4 earnings) $ARM — Armv9 royalty mix expansion and datacenter CPU share gains accelerate as hyperscalers build custom silicon → ~early Aug 2026 (FQ1 earnings) $INTC — First external 18A foundry customer volume commitments and yield milestones validate manufacturing turnaround thesis → ~late July 2026 (Q2 earnings) $AVGO — New hyperscaler custom AI accelerator design wins expand XPU pipeline beyond existing anchor customers → ~early Sept 2026 (FQ3 earnings) $META — AI ad-targeting gains and Reels monetization lift revenue growth despite heavy AI infrastructure spending → ~late July 2026 (Q2 earnings) $MSFT — Azure growth reacceleration and Copilot enterprise adoption prove AI monetization thesis at next earnings → ~late July 2026 (FQ4 earnings) $IONQ — New quantum milestone or major government contract validates roadmap toward fault-tolerant commercial systems → ~early Aug 2026 (Q2 earnings); contracts event-driven $ASTS — BlueBird satellite launches enable commercial direct-to-cell service activation with AT&T and Verizon partnerships → launch cadence through H2 2026; Q2 earnings ~mid Aug 2026 $BE — New datacenter fuel-cell power deals scale as AI energy demand outpaces grid capacity → ~early Aug 2026 (Q2 earnings); deals event-driven $NBIS — Accelerating GPU capacity buildout and new large customer contracts drive ARR guidance raise higher → ~mid Aug 2026 (Q2 earnings); Rubin deployment H2 2026 $IREN — AI cloud contract signings and GPU datacenter conversion accelerate revenue diversification away from bitcoin mining → ~late Aug 2026 (FQ4 earnings); contracts event-driven $NVDA — Rubin platform volume ramp in H2 drives datacenter revenue beats reaffirming AI compute demand → ~late Aug 2026 (FQ2 earnings); Rubin ships H2 2026 $AMD — MI400 series accelerator ramp and new hyperscaler commitments prove AMD's AI GPU share gains → ~early Aug 2026 (Q2 earnings); Helios/MI400 ramp Q3–H2 2026 $NOW — Now Assist AI agent adoption and subscription upsell accelerate as enterprises deploy agentic workflows → ~late July 2026 (Q2 earnings) $PLTR — New defense contracts and AIP commercial expansion accelerate US revenue growth beyond elevated trajectory → ~early Aug 2026 (Q2 earnings); contracts event-driven $ORCL — Massive OCI backlog conversion and new AI datacenter deals reaccelerate cloud revenue recognition sharply → ~mid Sept 2026 (FQ1 earnings) $AAOI — 800G datacenter transceiver ramp and new hyperscaler qualification wins drive revenue inflection higher → ~early Aug 2026 (Q2 earnings) $LITE — Datacenter optical transceiver demand and EML laser capacity expansion accelerate cloud revenue growth substantially → ~mid Aug 2026 (FQ4 earnings) $MRVL — Custom AI silicon ramp and new hyperscaler XPU design wins accelerate datacenter revenue growth → ~late Aug/early Sept 2026 (FQ2 earnings) $RKLB — Neutron's first launch unlocks the medium-lift market and validates reusable vehicle economics for growth → Neutron debut targeted Q4 2026; Q2 earnings ~early Aug 2026 $APLD — New hyperscaler lease agreements and Ellendale campus buildout unlock recurring high-margin AI hosting revenue → lease news event-driven; FQ1 earnings ~Oct 2026 $OUST — New automotive and robotics lidar design wins accelerate revenue growth toward sustained profitability inflection → ~early Aug 2026 (Q2 earnings); design wins event-driven

@stocksavvyshayJul 7, 2026 · 2:50 PM

Semiconductors are getting hit hard today: • $SNDK -14% • $ALAB -12% • $AMKR -11% • $AEHR -11% • $INTC -11% • $CRDO -11% • $AMAT -10% • $MRVL -10% • $AAOI -9% • $KLAC -9% • $AMD -9% • $COHR -8% • $ARM -7% • $LITE -6% • $ASML -6% • $TSM -5%

@stocksavvyshayJul 7, 2026 · 2:25 PM

@FuturumEquities Wild to see how close $AMD is to becoming the next $1T company

@stocksavvyshayJul 7, 2026 · 2:19 PM

@FuturumEquities Heres where you can see my full $AMD commentary: https://t.co/bkGt7NQZYH

@stocksavvyshayJul 7, 2026 · 2:08 PM

$AMD is becoming a two-engine AI story with EPYC CPUs for the growing compute needed around agentic workloads and Instinct GPUs for hyperscalers that need a credible alternative to $NVDA. The CPU side still looks underappreciated as AI agents require more high-core-count server compute and 2nm Venice should help AMD keep taking share as $INTC works through roadmap and manufacturing challenges. The bridge is Helios where AMD can package EPYC CPUs, Instinct GPUs and Pensando networking into a full rack-scale AI platform backed by real demand from $META, $ORCL and OpenAI.

@aleabitoredditJul 7, 2026 · 2:03 PM

Just putting it out there: If everything crashes together from $NBIS, $MRVL, $INTC, $SNDK, $AMD, $SIVE, $MU, $LITE, and others... Which are all down -4% to -10%+ today so far. Probably doesn't have anything to do with individual fundamentals. Indiscriminate selloffs from things like cascading margin liquidations, usually provide compelling opportunities if the underlying improves.

@investmentguru_Jul 7, 2026 · 10:31 AM

Silicon-to-Substation: 5 Themes Built for the Long Run Photonics / Optical Interconnects The bottleneck is shifting from compute to data movement. Co-packaged optics fixes that. $POET $LWLG $SIVEF $AAOI $AXTI $IQEPF $COHR Memory AI models are starving for bandwidth. HBM and next-gen memory are the unlock. $MU $SNDK $MRVL $CRDO $AMD $INTC Energy / Power Data centers need power that scales faster than the grid can build it. Nuclear and baseload energy step in. $OKLO $SMR $NNE $CEG $TE NeoCloud The new compute layer sitting between hyperscalers and startups. Capacity is the moat. $CRWV $NBIS $IREN $APLD $BTDR $KEEL $HIVE Physical AI / Robotics Intelligence is leaving the screen and entering the physical world. $SERV $OUST $AMBA $BB $JOBY $RR $CCXI The through-line: every layer feeds the next. Compute needs power. Power needs speed. Speed needs light. And all of it eventually walks off the screen and into the physical world. This is the stack we’re building around for the long run. Not financial advice.

@wliangJul 7, 2026 · 7:33 AM

Samsung reported the highest quarterly operating profit any tech company has ever seen... Passed both $NVDA and $AAPL. > 89.4T won operating profit (~$58B), +1810% YoY > Revenue 171T won > DRAM prices +44%, NAND +53% in Q2 As expected if you tracked $MU. HBM is sold out through 2027, DRAM and NAND ripping on supply shortage, and every memory layer is getting pulled up by AI demand simultaneously. And the stock dropped ~7%. Classic sell-the-news event. We've seen this with $SNDK and $AMD. Short-lived volatility into a rally. IMO the memory supercycle isn't slowing down. Just needed a little breathing room.

@alphaticaioJul 6, 2026 · 8:53 PM

DARK POOL SCANNER | July 6, 2026 $SPY -$783M dark pool. Lit: +$4.53B. Tapes disagree. Again. $MSFT -$271M. Alert validated. $DVN +$214M zero sells. $HOOD +$166M zero sells. Q3 week two. 336 off-exchange prints. +$141M net. Nearly flat. Dark Pool Buyers: $AMD +$253M (9 prints, BOTH AGREE BUY) $DVN +$214M (8 prints, $0 DP SELLS, NEW NAME) $HOOD +$166M (3 prints, $0 DP SELLS, NEW NAME) $NFLX +$109M (4 prints) Dark Pool Sellers: $SPY -$783M (47 prints, LIT: +$4.53B, TAPES DISAGREE) $MSFT -$271M (9 prints, BOTH AGREE, 12:37 PM ALERT VALIDATED) $MU -$227M (9 prints, BOTH AGREE, 3RD SESSION) $QQQ -$105M (9 prints, LIT: +$1.35B, TAPES DISAGREE) $SPY: lit +$4.53B, dark -$783M. Tapes disagree. The $SPY tapes-disagree pattern persists into Q3 week two. Six of the last eight sessions the dark pool sold $SPY. The lit tape bought in most. The structural divergence is the Q3 theme. The lit tape shows the headline. The dark pool shows the positioning. $MSFT: -$271M. Both agree sell. We flagged the $297M dark pool sell at 12:37 PM. The alert was validated by the closing data. The 17-week distribution paused for two sessions at the Q3 line (zero sells both sessions) and resumed today. Both tapes agree. The pattern survived the Q3 reset. $DVN: +$214M zero sells. Devon Energy. The dark pool is buying an oil producer with zero opposition while oil sits at $68. $HOOD: +$166M zero sells. Robinhood. Fintech. New names appearing on the dark pool board. The institutional universe is expanding beyond mega-cap tech. $MU: -$227M. Both tapes agree sell. Third session. The post-earnings reversal continues at scale. Net: +$141M. Nearly flat in aggregate. The dark pool is balanced but the positioning is surgical: sell the indexes and legacy tech ($SPY, $QQQ, $MSFT, $MU), buy the new names ($AMD, $DVN, $HOOD, $NFLX). The dark pool is rotating within equities, not exiting equities. Q3 week two. The tapes disagree. The data names the names. Watching the tape.

@frenchie_Jul 6, 2026 · 8:51 PM

@PhotonCap You have a position ? I feel no one talks about $AMD

@alphaticaioJul 6, 2026 · 8:19 PM

+$11.79B. $SPY +$4.53B. $AAPL +$2.00B zero sells. Third session. July seasonality confirmed. $MSFT -$503M. Distribution returned. Q3 week two. $27.93B in buy blocks. $16.14B in sells. Net flow: +$11.79B. 23,274 prints. The buyers: $SPY +$4.53B (2,810 blocks) $AAPL +$2.00B (693 blocks, $4.1M SELLS, ZERO, 3RD SESSION) $QQQ +$1.35B (2,470 blocks) $TSLA +$1.11B (1,305 blocks) $AMD +$812M (1,450 blocks) $NVDA +$507M (1,022 blocks) The sellers: $MSFT -$503M (651 blocks, DISTRIBUTION RETURNED, BOTH AGREE) $MU -$473M (3,984 blocks, 3RD SESSION SELL) $AAPL: +$2.00B. Zero sell blocks. Third session. We flagged the $51M $AAPL dark pool buy on Q3 day one. $AAPL has been positive every July since 2016. Ten for ten. Average return +6.4%. Three consecutive sessions of accumulation with zero opposition. Combined three-day $AAPL accumulation: $5.07B on the lit tape. The July play is confirmed. $MSFT: -$503M. The distribution returned. Q3 day one: +$1.67B. Day two: +$2.02B. Two sessions of buying that broke the 17-week pattern. Today: -$503M. We flagged the $297M dark pool sell at 12:37 PM. Both tapes agree sell. The distribution paused for two sessions at the quarter line and resumed. $MU: -$473M. Third session of selling. The post-earnings reversal continues. Three consecutive sell sessions. Combined: -$10.3B over three sessions. $SPY: +$4.53B on the lit tape. But the dark pool sold $SPY -$783M. Tapes disagree. The lit tape buys at scale. The dark pool sells. When the tapes disagree, the dark pool has historically been the better directional indicator. Q3 week two. $AAPL bought with zero opposition. $MSFT distribution returned. The tape picks winners. The tape doesn't have an opinion. It has a receipt.

@stocksavvyshayJul 6, 2026 · 6:40 PM

$AMD announced it is contributing to Trump Accounts joining the growing list of companies funding the new investment accounts for American children. AMD CEO Lisa Su attended President Trump’s Trump Accounts event today in the White House Rose Garden. https://t.co/Y2fOxgSQME

@alphaticaioJul 6, 2026 · 4:05 PM

UNUSUAL DARK POOL PRINT DETECTED: │ 11:54:42 AMD BUY $119,175,000 NASDAQ Carte │ $AMD $SPY $QQQ https://t.co/FDgDbiHFN4

@stocksavvyshayJul 6, 2026 · 3:17 PM

$AMD UP OVER 10% TODAY https://t.co/OnEG82lmyo

@unusual_whalesJul 6, 2026 · 2:57 PM

Goldman Sachs has raised its AMD, $AMD, price target to $640 from $450 and reiterated its Buy rating.

@investmentguru_Jul 6, 2026 · 2:49 PM

@investmentguru_Jul 6, 2026 · 2:15 PM

AI infrastructure stocks are bouncing and holding so far $MU $AMD $INTC $CRDO $MRVL $AAOI $AVGO

@frenchie_Jul 6, 2026 · 2:10 PM

Jensen Huang a menti La prochaine boîte à 1 Trillion de dollar c’est $AMD https://t.co/2rAKKwUzm2

@itsmichaelluuJul 6, 2026 · 12:35 PM

You do not have to buy all of these stocks! Pick your top 3. These are all the leaders in AI, space, energy and robotics: $MU (~$977) — HBM memory is AI's core bottleneck; every GPU cluster needs more Micron yearly $DRAM (~$61) — One-ticket basket of every memory maker riding the AI storage and HBM supercycle $NVDA (~$194) — The default AI compute platform; Rubin and beyond keep it at the center $AMD (~$410) — MI-series accelerators give hyperscalers a real second source as compute demand explodes $ASTS (~$85) — Direct satellite-to-phone coverage; a working constellation could connect billions of unserved devices worldwide $OUST (~$51) — Digital lidar is the eyes of robots, trucks, and autonomous machines everywhere $VPG (~$123) — Precision sensors and load cells feed the humanoid robot and automation buildout directly $AMKR (~$70) — Advanced packaging is the new bottleneck; every AI chip needs someone to assemble it $AVGO (~$360) — Custom AI silicon plus networking makes Broadcom essential to every hyperscaler data center $SPCX (~$162) — Starlink, launch monopoly, and Starship give SpaceX unmatched dominance across the space economy $IONQ (~$50) — Quantum leader with real cloud revenue; a 2030 breakthrough could rerate it enormously $TSLA (~$393) — Robotaxi and Optimus turn Tesla from carmaker into an AI-and-robotics platform story $TE (~$8.56) — Domestic solar and battery maker positioned for AI-driven power and energy-independence demand $POET (~$8.84) — Optical interposer packaging could ride the co-packaged optics wave hyperscalers now demand $NBIS (~$216) — Fast-growing neocloud renting GPUs to AI labs desperate for compute capacity now $INTC (~$121) — U.S. foundry comeback; Apple and hyperscaler deals could validate the turnaround by 2030 $SNDK (~$1,745) — NAND and storage demand surges as AI data volumes and inference workloads explode $LITE (~$737) — Optical components move AI data between chips; photonics is the next big bottleneck $BE (~$272) — Fuel cells give data centers on-site power the grid can't deliver fast enough $AAPL (~$305) — Massive installed base plus on-device AI could turn every iPhone into an agent $MSFT (~$390) — Azure plus Copilot monetizes AI across enterprise; OpenAI stake compounds the upside If you have a smaller account less than $2000 choose these 3 for now:

@aleabitoredditJul 6, 2026 · 8:25 AM

I personally think $SIVE can be the next $LITE. In the past few months alone, we've seen: 1. Partnerships with O-Net pushing ELS into mass production 2. $JBL 1.6T LRO mass production signals with "relatively dramatic moats" for pluggables using Sivers. 3. $GFS SCALE reference level laser for hyperscalers with pluggable, NPO, CPO. -> -> where $AMD and others went to GFS for CPO. 4. Ayar, which joined $NVDA NVLink for CPO -> -> which removed Lumentum/Macom from their website and likely made Sivers their primary laser supplier. -> -> AlChip likely Trainium win from Amazon price placement (Ayar's customer) -> -> GUC rack level design in with Ayar. -> -> Raised $500m for mass production by AMD, Alchip, Mediatek, and NVIDIA 5. ~ $AEVA starting HVM H2 2026. 6. $POET starting HVM H2 2026 with hyperscaler suppliers like Lumilens ("top 3 hyperscaler initial customer") 7. TFLN + $SIVE CW Lasers with Lightium 8. Likely direct relationships with $MRVL Celestial and CPO players like Lightelligence/Lightmatter. 9. Multiple new undisclosed relationships for pluggables following Jabil in their quarterly transcripts With new Trendforce reports that $AMD and other hyperscalers are trying to source LTAs for CW laser sources, serving as a direct catalyst for independent CW sources. So when hyperscaler suppliers from Jabil to O-Net are incentivized to mass produce as many as they can: That's very material for revenue for Sivers relative to current valuations, and it looks like just a waiting game. Even in the past week: - $SIVE raised an oversubscribed institutional round for volume ramp... This is very nuanced since Sivers is fab-lite so it's not going to in-house foundry capex to scale. Likely toward Win Semi and others (they mentioned other partners too), for laser scaling + foundry allocations. So this is likely signaling material for revenue ramp is coming. - Sivers also mentioned NASDAQ listing completion targeted in the next few quarters (probably H2 2026 or Q1 2027 is my est. timeframe). This would fund M&A efforts, since it's impossible with the fundraising environments in local Swedish markets. As for becoming the next $LITE: M&A makes their lasers more valuable, so downstream IP acqusition -> into contract manufacturing like $FN, and others to make the full 1.6T pluggable or optical engines. Is how they get there, since laser array ASP scaling that people are modeling off of, wouldn't command a $60B+ valuations. There's going to be a lot of bridge architectures like NPO/pluggables, etc and noise around certain architectural delays in the meantime. But markets misunderstand laser companies like $LITE, $SIVE, $AAOI and others are used across different architectures compared to if you just look at certain passive optical components. So markets see "CPO delay headlines" algos sell off laser companies that benefit from other architectures. Being included in the pluggable 1.6T ramp to CPO scale out (Which Sivers is included in), helps bridge revenue waiting gaps until scale up inflection point H2 2027. I'm personally holding long term, since I haven't seen a ~$1.4B company mapping to this many hyperscalers before. TLDR: - Waiting on volume ramps from different architectures to play out across their hyperscaler supplier mapping -> 1.6T LRO/CPO scale out late H2 2026 start into high volume ramp 2027 -> H2 2027 CPO scale up volume ramp - Waiting on NASDAQ listing likely H2 2026/Q1 2027 for M&A efforts to fully take off, unless Sivers get more creative with equity financing in the meantime.

@stocksavvyshayJul 5, 2026 · 4:06 PM

POPULAR STOCKS BY PEG RATIO PEG < 1 usually means mispriced growth PEG > 2 starts to push into the danger zone Here’s how they stack up: • $TSLA 4.5x • $SPCX 3.8x • $GOOGL 2.9x • $AAPL 2.6x • $INTC 2.5x • $HOOD 2.0x • $PLTR 2.0x • $META 1.5x • $ASML 1.5x • $AMZN 1.4x • $MSFT 1.2x • $TSM 1.1x • $SOFI 0.9x • $HIMS 0.9x • $NOW 0.7x • $NVDA 0.6x • $AMD 0.6x • $AVGO 0.5x • $ORCL 0.4x • $MU 0.1x • $SNDK 0.1x

@stocksavvyshayJul 5, 2026 · 3:07 PM

$AMD could approach $100B in annual revenue by 2028 as $META, $ORCL and OpenAI commitments turn MI450 and Helios into a real second source for AI compute. The GPU deals are what re-rated AMD but the CPU bottleneck gives the story a second leg because every AI cluster still needs general compute for orchestration and inference workloads. That's why the CPU leg keeps compounding since agentic AI is making hyperscalers more CPU hungry with EPYC already at 46% server CPU revenue share and AMD's next-gen Venice already ramping on $TSM 2nm.

@itsmichaelluuJul 5, 2026 · 12:57 PM

24 stocks ON SALE right now and will break all time highs by Jan 1. 🔴 Micron $MU -22.3% 🔴 SanDisk $SNDK -25.9% 🔴 Western Digital $WDC -32.6% 🔴 Arm Holdings $ARM -30.4% 🔴 Intel $INTC -15.5% 🔴 Broadcom $AVGO -27.2% 🔴 Meta $META -26.8% 🔴 Microsoft $MSFT -29.7% 🔴 IonQ $IONQ -42.0% 🔴 AST SpaceMobile $ASTS -35.7% 🔴 Bloom Energy $BE -17.6% 🔴 Nebius $NBIS -28.1% 🔴 IREN $IREN -49.5% 🔴 Nvidia $NVDA -17.6% 🔴 AMD $AMD -11.4% 🔴 ServiceNow $NOW -49.7% 🔴 Palantir $PLTR -37.7% 🔴 Oracle $ORCL -59.4% 🔴 Applied Optoelectronics $AAOI -48.2% 🔴 Lumentum $LITE -32.9% 🔴 Marvell $MRVL -25.6% 🔴 Rocket Lab $RKLB -34.6% 🔴 Applied Digital $APLD -34.8% 🔴 Ouster $OUST -20.5% Here's the CHEATSHEET for you to save:

@stocksavvyshayJul 4, 2026 · 2:54 PM

@fiscal_ai This is less a knock on $AMD and more a reflection of how cheap $NVDA looks to me here.

@stocksavvyshayJul 4, 2026 · 2:45 PM

Wild to see $AMD now trading at nearly 3x $NVDA forward earnings multiple with Nvidia at ~20x despite 90% AI GPU share, 75% gross margins and a path to $1T in cumulative Blackwell and Rubin revenue through 2027. The market is paying AMD for rate of change after landing $META, $ORCL and OpenAI gigawatt commitments but the real decoupling came from the CPU bottleneck trade. GPU deals re-rated AMD but the CPU trade is what helped it decouple from Nvidia as EPYC gains pricing power into tighter supply and hyperscalers “gobbling CPUs” for agentic workloads.

@stocksavvyshayJul 4, 2026 · 2:15 PM

UPDATED FIB LEVELS FOR POPULAR STOCKS Space • $RKLB $92 • $ASTS $79 • $PL $26 • $RDW $8 Quantum • $IONQ $44 • $RGTI $16 • $QBTS $20 • $INFQ $10 Drones • $ONDS $7 • $AVAV $174 • $KTOS $52 • $MRCY $104 Nuclear • $OKLO $50 • $UUUU $12 • $GEV $906 • $LEU $132 Photonics • $AAOI $103 • $LITE $636 • $COHR $311 • $AEHR $59 CPU • $AMD $515 • $INTC $120 • $ARM $304 • $AMKR $64 Physical AI • $TSLA $382 • $AMZN $230 • $GOOGL $357 • $ISRG $397 AI Utilities • $IREN $31 • $NBIS $213 • $CIFR $17 • $CRWV $67 AI Power • $EOSE $4 • $BE $265 • $NVTS $13 • $VRT $282 AI Connectivity • $AVGO $334 • $ALAB $341 • $CRDO $210 • $MRVL $225 AI Hardware • $NVDA $192 • $TSM $419 • $ASML $1,732 • $MU $869 AI Applications • $PLTR $121 • $NOW $104 • $CRWD $171 • $SNOW $181

@investmentguru_Jul 3, 2026 · 6:14 PM

The Category Creator Portfolio Some businesses aren’t just “leaders” — they’ve built moats so wide, nobody’s even trying to cross them. $PLTR — Palantir Not a “software company.” It’s the only platform doing AI-driven ontology at classified government + enterprise scale simultaneously. Competitors sell tools. Palantir sells outcomes baked into defense and intelligence infrastructure that took a decade to build trust for. $SPOT — Spotify Global audio distribution + the algorithm that actually knows what you want to hear next. Apple Music and YouTube Music are label-first. Spotify is user-first. That’s why churn stays low and podcasting/audiobooks keep expanding the moat. $HIMS — Hims & Hers Direct-to-consumer telehealth with its own pharmacy, its own compounding, its own brand trust in categories people are embarrassed to walk into a clinic for. Nobody else owns the full stack front-to-back like this. $TSLA — Tesla Not just EVs. Vertically integrated batteries, FSD data flywheel, Optimus robotics, and a charging network competitors now pay to use. The closest comps aren’t even in the same business anymore. $AMD — Advanced Micro Devices The only credible alternative to Nvidia at scale in AI compute, while also owning x86 CPU share from Intel on the other side. Two-front dominance nobody else can claim. $DUOL — Duolingo Language learning + gamification + AI tutoring in one free-to-paid funnel with network effects (leaderboards, streaks, social) that no edtech competitor has replicated at this user count. $IBRX — ImmunityBio Proprietary IL-15 superagonist platform (Anktiva) — a completely different immunotherapy mechanism than PD-1/PD-L1 checkpoint inhibitors. First-in-class, not best-in-class. $ASTS — AST SpaceMobile Direct-to-cell satellite broadband from space, compatible with unmodified smartphones. Starlink’s version needs a partnered carrier workaround. AST built the actual satellites for it — patented spectrum + tech nobody else has cleared. $RKLB — Rocket Lab Only company besides SpaceX with a proven, reusable, frequently-flying orbital rocket AND a full space systems manufacturing arm (satellites, components, Neutron coming). Everyone else is still pre-revenue or single-product. $POET — POET Technologies Optical interposer platform integrating photonics at the chip level for AI datacenter interconnects. Nobody else has this exact manufacturing approach at this stage — first-mover on a bottleneck the whole AI buildout needs solved. $OKLO — Oklo Fast fission microreactor design with NRC engagement and a direct line to hyperscaler power deals. Most nuclear names are still legacy large-reactor. Oklo’s the closest thing to “AI power, deployed.” $CRWV — CoreWeave Purpose-built GPU cloud from day one, not a retrofit of general cloud infra. Hyperscaler-scale AI compute contracts without hyperscaler legacy overhead — a category it effectively created. $MDA — MDA Space Only Western company with proven satellite bus manufacturing + robotics heritage (Canadarm lineage) at commercial scale — critical infrastructure nobody’s replicating fast.

@tradexwhispererJul 3, 2026 · 5:37 AM

$MU $DRAM Meta's Cost-Saving Strategy: Re-Recycle Legacy DDR4 + Leverage CXL (Memory Expansion). “CXL memory expanders give new life to DDR4 memory.” To combat soaring DDR5 prices and hardware costs, Meta is reusing legacy DDR4 memory from decommissioned servers in brand-new servers that only support DDR5 Server Configuration: A single MemServer uses one 158-core $AMD Turin CPU with: -768 GB DDR5-6400 (local, high-bandwidth: 614 GB/s) -256 GB CXL-attached DDR4-2400 (slower tier: 76 GB/s) Total: 1 TB memory. The system treats CXL memory as a separate NUMA node; Linux automatically migrates "cold" (less-used) pages to the slower DDR4 tier.

@stocksavvyshayJul 2, 2026 · 5:03 PM

Semiconductors are extending their pullback for a second straight day: • $AAOI -17% • $SNDK -14% • $KLAC -13% • $AMKR -12% • $MRVL -10% • $COHR -10% • $LITE -9% • $CRDO -8% • $ARM -7% • $MU -6% • $AMD -6% • $INTC -5% • $ASML -5%

@aleabitoredditJul 2, 2026 · 4:56 PM

Photonics is backed by actual revenue numbers and it's an architectural shift championed by $NVDA. Quantum barely has any revenue. $LITE is completely sold out for the next 2 years (per $POET AGM) likely starting into 2029. Lumentum is so strained that they buy CW lasers off competitors (earnings transcript) $COHR is bottlenecked, so they buy EML off Lumentum. Then, $AAOI is coming in with Made-in-America independent CW capacity, are projecting $1.4B/quarterly revenue ending H1 2027 of a stupid $9.3B MC today. So all the CW capacity from independent players who have it now like $AAOI or $SIVE are likely to become scarce resources. Many other hyperscalers have already started LTA discussions (per Trendforce). And players like $AMD are currently talking with players such as $AAOI (Rosenblatt channel checks). Thematically, next 2 years is 9x TAM to US$154B per GS reports, especially with 16x/45x dollar content increase in scale out/scale up. Then there's the overall thematic AI drop from $META, which is widely misunderstood because people conflate what "excess capacity" means. And as UBS mentioned, Meta planning a cloud offering is NOT NEW NEWS. Bloomberg just has a tendency to publish information that causes doom drops across the semi sector like Nvidia export controls a few months back. But I'm familiar with what I'm holding so I'm confident in these numbers playing out. Especially when all the major players are sold out, the fundamentals catch up eventually.

@itsmichaelluuJul 2, 2026 · 12:58 PM

South Korea's stock market crashed -8.5% overnight. SK Hynix and Samsung led the bleed, hit by an AI semiconductor selloff and Korea's hottest inflation print since Dec 2023. $UBS says 60% of companies are now curbing AI spending, shifting to cheaper models including open-source Chinese options like DeepSeek and Qwen. This is a buying opportunity. Global AI compute demand isn't going anywhere. This is a shakeout, not a top. Best AI infrastructure plays, and how far each sits from its 52-week high: $MU -17.7% $SNDK -14.0% $NBIS -21.0% $META -23.2% $MSFT -33.1% $GOOGL -6.0% $NVDA -16.5% $AMD -7.5% $INTC -10.8% $AVGO -25.4% $MRVL -18.0% $BE -13.8% $VRT -17.2% Still extremely bullish. Generational wealth gets made in stretches like this, for the people who stay focused. Big themes I'm not losing sight of: quantum, space, energy, AI supply chain, robotics.

@investmentguru_Jul 2, 2026 · 10:39 AM

10 Semiconductor & AI Infrastructure Stocks to Watch $MU $AMD $INTC $NBIS $MRVL $CRDO $AVGO $QCOM $ARM $AAOI This group has seen increased selling pressure and profit-taking, but don’t lose sight of the bigger picture. At some point, these names could present excellent buying opportunities as AI infrastructure spending, data center expansion, and next-generation semiconductor demand continue to be long-term growth drivers. Keep them on your watch list and let the charts confirm when momentum begins to return.

@midascabalJul 2, 2026 · 8:57 AM

Here are the top stocks I found by using @asymmetrix_xyz, the platform that can spot AI bottlenecks in seconds: $OUST $MP $QCOM $AMD $TER If this helped you out, follow me for more.

@itsmichaelluuJul 2, 2026 · 5:03 AM

1️⃣ Vera Rubin ramp (H2 2026) — Rubin already sampling, early access with $MSFT $GOOGL $META $AMZN. Mgmt has flagged ~$500B combined Blackwell+Rubin revenue visibility through end of 2026. 2️⃣ Hyperscaler capex confirmation GOOGL/AMZN/META/MSFT tracking toward ~$650-700B combined 2026 capex. Every earnings season either confirms/raises this (bullish) or shows ROI scrutiny (bearish). This matters more than NVDA's own print. 3️⃣ Sovereign AI this segment reportedly tripled in FY26 to ~$30B as governments build national AI infra outside the hyperscaler channel. Diversifies away from the "big 4" capex cycle. Risk side: China DC revenue is now fully excluded from guidance, and $AMD MI400/HBM4 platform (with a reported ~$60B Meta deal) is the most credible competitive threat Nvidia has faced.

@investmentguru_Jul 1, 2026 · 10:58 PM

Today’s selloff in AI infrastructure stocks looks more like a rotation in sentiment than the end of the AI story. Reports that $META plans to lease excess cloud-computing capacity raised concerns that data center supply could temporarily outpace demand, triggering broad profit-taking across AI infrastructure names. High-flyers like $NBIS, $CRWV, $MU, $AMAT, $AMD, $TSM, $MRVL, $LRCX, $GLW and others saw sharp declines after extended runs. This is a reminder that markets often overreact to headlines. AI infrastructure remains a long-term secular growth theme, but after massive gains, corrections and sector rotation are healthy. If AI adoption continues to expand over the coming years, quality companies in semiconductors, networking, memory and cloud infrastructure could present attractive buying opportunities during periods of weakness. Stay focused on the long-term trend, not just today’s price action.

@alphaticaioJul 1, 2026 · 9:16 PM

$MU -$5.27B. Four semis combined -$8.03B. $MSFT +$1.67B flipped to buy. $SPY +$3.86B. $META +$2.63B. The hardware/software split hit the block tape. Q3 day one. $26.36B in buy blocks. $21.55B in sells. Net flow: +$4.81B. 30,964 prints. The buyers: $SPY +$3.86B (2,917 blocks, FLIPPED, YEST: -$1.45B) $META +$2.63B (2,338 blocks, $161M SELLS, NEAR ZERO) $NVDA +$2.14B (1,585 blocks, THE SEMI EXCEPTION) $MSFT +$1.67B (986 blocks, 17-WEEK DISTRIBUTION BROKEN) $AMZN +$1.08B (740 blocks, $96M SELLS, NEAR ZERO) $AAPL +$697M (678 blocks, $96M SELLS, NEAR ZERO) $TSLA +$557M (1,147 blocks) $IWM +$231M (329 blocks) The sellers: $MU -$5.27B (5,843 blocks, POST-EARNINGS BUY STREAK ENDED) $INTC -$1.01B (1,131 blocks) $AMD -$992M (1,339 blocks) $TSM -$765M (375 blocks, $3.7M BUYS, NEAR ZERO) $QQQ -$514M (2,982 blocks) The rotation scanner showed $IGV +4.30% and $SOXX -5.31%. The block tape drew the same line. Four semi names sold -$8.03B combined: $MU -$5.27B, $INTC -$1.01B, $AMD -$992M, $TSM -$765M. The hardware sell is across the entire complex. $NVDA: +$2.14B. The only semi name on the buy side. The institutions are distinguishing between $NVDA and the rest. $NVDA is AI infrastructure. $MU is memory. $INTC is legacy. $TSM is foundry. $AMD is compute. The sell is specific. $NVDA is the exception. $MSFT: +$1.67B. We flagged the $250M dark pool buy at 12:25 PM. $MSFT was the most persistently distributed name in seventeen weeks. Every week. Every session. Through FOMC, OPEX, the great rotation, and Q2 rebalancing. Today it flipped. $1.90B in buys against $231M in sells. The distribution pattern that defined the series broke on Q3 day one. $MU: -$5.27B. Four consecutive post-earnings buy sessions. $258M in buys against $5.52B in sells. Near-zero buying. The post-earnings accumulation ended. The tape reversed. $SPY: +$3.86B. After selling -$1.45B yesterday. The S&P index flipped to buy on Q3 day one. $META: +$2.63B with $161M in sells. Near-zero opposition. The largest $META buy of the series. $AMZN: +$1.08B with $96M sells. $AAPL: +$697M with $96M sells. Three mega-caps with near-zero sell opposition. Q3 day one. The hardware sold. Everything else bought. The block tape drew the same line the rotation scanner drew. The data agrees. The tape doesn't have an opinion. It has a receipt.

@frenchie_Jul 1, 2026 · 6:51 PM

$NBIS semble sous-évalué Pour revenir factuellement sur ce qui s'est passé et pourquoi le lézard Zuckerberg fait pump $META mécaniquement au détriment de $NBIS 1. Les 90 derniers jours racontent l'histoire inverse qu'un Meta assis sur du surplus. >Le 16 mars $META dit : jusqu'à 27 milliards de dollars sur cinq ans engagés chez $NBIS pour de l'infrastructure IA. >Le 9 avril le même nous redit : accord étendu avec $CRWV pour environ 21 milliards de dollars de capacité cloud IA jusqu'en 2032, ce qui fait porter l'engagement total de Meta chez $CRWV à 35 milliards de dollars, en faisant l'un de ses plus gros clients. >Encore Il y a deux semaines : nouveaux accords avec Crusoe pour environ 1,6 gigawatt de capacité additionnelle. >Une énième fois, il y a 3 jours, la Financial Times révèle que $GOOG a dû plafonner l'accès de $META à Gemini, faute de pouvoir fournir toute la capacité demandée. Résultat ? Sur les 4 dernières news compute de Meta : achète, achète, achète, se fait rationner par son propre fournisseur. Zéro trace de surplus, absolument 0 2. Ce matin, Bloomberg rapporte que $META évaluerait deux options : >héberger des modèles pour développeurs façon AWS Bedrock Ou >vendre de la capacité de calcul brute façon neocloud. Meta n'a rien confirmé, et Reuters dit ne pas avoir pu vérifier les détails de façon indépendante Le marché achète quand même cette histoire douteuse et punit indifféremment Neocloud, $META +10%, $CRWV -13%, $NBIS -15%, et le hardware IA ( $MU , $NVDA , $AVGO , $AMD , $MRVL, $INTC dans le rouge) Le scénario n'a rien d'original en vrai , c'est celui de $SPCX, xAI se retrouvait avec des centaines de milliers de GPU sous-utilisés par Grok, loués depuis à Anthropic, Google etc... et META a vu que cette bascule neocloud a largement dopé la valorisation IPO de SpaceX à plus de 2 000 milliards de dollars. Sauf que Meta n'a, factuellement, aucun surplus visible dans le narratif : c'est un reframing IR, calculé pour soulager le ROE après un capex 2026 guidé à 125-145 milliards, faire remonter le titre, et lever du capital dans de meilleures conditions pour financer, on s'en doute, encore plus de compute Rien n'a changé sur $META ni $NBIS, pour moi c'est un arbitrage évident

@spotgammaJul 1, 2026 · 6:32 PM

$AMD 25 delta call IV. I know Americans like to consume but this is getting ridiculous. https://t.co/sdRdCHTen3

@amitisinvestingJul 1, 2026 · 5:34 PM

@Sambo_mac41 $AMD at $600 also gives $META $100B of equity they can use in many ways!

@stocksavvyshayJul 1, 2026 · 5:14 PM

Semiconductors pulling back today after reports that $META is developing “Meta Compute” to sell access to excess AI compute: • $KLAC -11% • $SNDK -11% • $AMAT -10% • $MU -10% • $ALAB -8% • $INTC -8% • $MRVL -8% • $COHR -7% • $TSM -7% • $AMD -6% • $ARM -5%

@alphaticaioJun 30, 2026 · 10:09 PM

+$14.16B. 11 buyers. 3 sellers. Five names with zero sells. $QQQ +$4.36B. $MU +$2.92B. $SPY the lone index sell. Last day of Q2. 18,740 blocks (combined sessions). June 30, 2026. The buyers (11 names, five with zero sell blocks): $QQQ +$4.36B (1,843 blocks) $MU +$2.92B (4,081 blocks, 4TH POST-EARNINGS BUY) $NVDA +$1.41B (1,043 blocks) $TSLA +$1.06B (869 blocks, $0 SELLS, 2ND SESSION) $AMD +$930M (1,115 blocks) $INTC +$643M (753 blocks, $0 SELLS) $IWM +$642M (178 blocks, $0 SELLS) $GOOGL +$518M (129 blocks, $0 SELLS) $SOXX +$495M (133 blocks, $0 SELLS) $META +$377M (318 blocks) $MSFT +$334M (653 blocks) The sellers (3 names): $SPY -$1.45B (1,850 blocks, BOTH TAPES SELL) $AMZN -$574M (479 blocks) $TLT -$285M (212 blocks, DURATION SOLD) The broadest lit tape buy of Q2. 11 names buying. Five with zero sell blocks. The institutions bought tech, semis, small caps, and growth into Q3. Only $SPY, $AMZN, and $TLT sold. $QQQ +$4.36B. The largest $QQQ buy of the series. The institutions are buying the Nasdaq into July. $MU +$2.92B. Fourth consecutive post-earnings buy. The algorithm identified the accumulation. The earnings confirmed. The tape keeps buying. 4,081 blocks. The most blocks of any name. $SPY: -$1.45B. The lone index sell. $QQQ bought. $IWM bought. $SPY sold. The S&P distribution that began last week continues while the other indexes are accumulated. The institutions are not selling equities. They are selling the S&P specifically. $TSLA: +$1.06B with zero sells. Second consecutive session with near-zero or zero opposition. The $TSLA accumulation is new and aggressive. $TLT: -$285M. Duration sold. The bond trade that was bought into FOMC and PCE is being unwound. The institutions moved back from bonds to equities. $SOXX: +$495M zero sells. We flagged the $249M dark pool buy at 12:11 PM. The lit tape confirmed with zero opposition. Last day of Q2. The lit tape bought everything except $SPY. Q3 starts tomorrow. The tape doesn't have an opinion. It has a receipt.

@yasutaketinJun 30, 2026 · 8:18 PM

Etchedやセレブラスなどの「推論最適化」企業は、 $NVDA $AMD $AVGO （カスタムASIC）へのライバル。 Etchedの推論ラックはメモリー節約効果も高い。

@stocksavvyshayJun 30, 2026 · 7:39 PM

$TSM is targeting 2029 for panel-level CoPoS packaging as AI chip scaling shifts from transistor density to packaging capacity. CoPoS could fit more compute dies and up to 12 HBM4 stacks into one package widening TSMC’s moat behind next-gen $NVDA and $AMD chips. https://t.co/yvIpsgYYIS

@thevalueistJun 30, 2026 · 6:44 PM

$INTC $AMD $ARM $NVDA My updated CPU tracker is attached. These are names with a high concentration of CPU-driven revenues and earnings. https://t.co/9RhbLPIheN

@tradexwhispererJun 30, 2026 · 6:24 PM

$MU $SNDK $AMD $INTC $ARM $QCOM "ByteDance Reportedly Eyes Next-Gen In-House CPU for 2H27 Mass Production." The barrier for CPU? Clearly not as high as memory. Meanwhile at 86% margins, the memory competitive wall is so tall that nobody can even dream of producing advanced DRAM and HBM. The smarter play is locking in 5 year LTAs with the oligopoly. No shortcut exists. 'ByteDance currently relies on CPUs from Intel and AMD, both of which have significantly raised prices in recent months, with quarter over quarter increases of 10% to as much as 35%.' That kind of pricing power invites competition. And it just did. 'The initiative also comes as the role of high performance CPUs is being re evaluated in the era of agentic AI, with workloads evolving from pure matrix computation to more complex task orchestration.' 'Partnering with Qualcomm… could help ByteDance navigate an advanced chip supply chain constrained by bottlenecks spanning wafer fabrication to advanced packaging.' CPU moats shrink. Memory Oligopoly deepens.

@wliangJun 30, 2026 · 5:08 PM

@Lui_Y0 $NVDA is already at ~$5T. IMO $AMD will close that gap towards $5T MC over the next year or two. https://t.co/sFZSMU9sSB

+ 900 more