Micron $MU fell to $858 at Thursday's low and closed at $931.24, an 8% round trip in a single session on 38.9 million shares, about 50% above its average volume. The flush came from Korea, where Samsung fell 9.5% and SK Hynix fell 11.1%. The recovery came from Micron's own calendar. Thursday Micron opened its second high-bandwidth memory packaging plant in Taichung. At the ceremony, COO Manish Bhatia said "the entire planned 2027 capacity is already sold out." That sits on top of guidance for $61.5 billion in revenue this quarter, plus or minus $1.5 billion, after a record $54.23 billion last quarter. A $35.16 billion buyback can start on Dec 9. The stock is still 26% below its $1,255 high. The open risk is labor. The Taoyuan union voted 99% to authorize a strike, mediation with the larger Taichung union is set for Oct 22, and the ask is 15% of operating profit, roughly $15 billion a year at last year's level. Is sold-out 2027 supply enough to hold the $858 low?

Health care makes new highs while last year's hot trades roll over. New lows: $AVAV $KTOS $RGTI $QBTS $DNN $UEC $JOBY $ACHR $OKLO $SMR New high: $PFE Capital is rotating from story stocks into cash generators. Bookmark this for your watchlist.

$GME CEO Ryan Cohen bought 844,130 shares Tuesday for about $22.8 million at an average near $27. That is his third open-market buy in a month, after $26.4 million in September and $17.1 million at the start of October. His direct holdings now sit at 42.5 million shares, roughly 9.1% of the company. The stock closed at $27.22, pennies from its $27.53 52-week high, and is up 35.6% this year. It added another 2% after hours on the filing. Now the calendar. GameStop's warrants carry a $32 strike and expire Oct 30. There are about 59 million of them, worth up to $1.9 billion in fresh cash if every one gets exercised. The stock needs to rise about 18% from here for that to make sense for holders. Insider buying this close to a strike deadline is not a guarantee of anything. It is a signal about what the CEO expects, and he is paying for it with his own money. TradeSmith's technical score on $GME reads 91 out of 100. I read that as money is flowing into the stock.

$NBIS closed Wednesday at $186.01, down 10%, and now sits 38% below its 52-week high of $299.86. Nebius itself released no news. The selling came from across the street. CoreWeave's 2032 bonds touched a 15% yield, Gartner said more than 80% of neocloud providers will stop operating independently by 2030, and every debt-funded data center name got repriced in one session. The one Nebius-specific headline was an insider sale, and it was smaller than it looked: the chief revenue officer exercised 7,000 options at $40 and sold at $211.23 under a trading plan filed in June, about $1.5 million. He still holds 336,000 options. So the question is not whether the contracts changed. They did not. The question is what the market will pay for contracted revenue when the cost of money for the whole group just jumped. Rosenblatt's $304 target, set six days before the drop, sits 63% above Wednesday's close. Same business, 10% cheaper, with a new fear attached. Is the credit scare a reason to wait, or the reason to look?

Palantir $PLTR is up 1,117% since the start of 2024, and it just recently hit another all-time high. Here is the bigger idea behind that number. More than $1 trillion has poured into AI infrastructure since ChatGPT launched in late 2022. That was phase one, the picks and shovels. Phase two is the companies that build applications on top of all that infrastructure. It mirrors the internet playbook. Apple, Amazon, and Google did not lay the fiber. They built the businesses that ran on it, and they compounded for decades. $PLTR is the clearest example today. It does not build data centers or frontier models. It builds software, Gotham, Foundry, and its AI Platform, on top of everyone else's. The growth is showing up in the revenue. Sales went from $1.09 billion in 2020 to $4.48 billion last year, and the company guides to roughly $8.15 billion this year, about 82% growth. TradeSmith backs the momentum. Its Quantum Score on $PLTR is 84 and its Technical Score is 88, both firmly in high-conviction territory. The application layer could produce the next decade of 10X winners, the same way the internet did. $PLTR will not be the only one. The real question is which names join it. Which AI application stock do you think runs next?

$NBIS carries a $37.5 billion backlog against revenue guidance of $3 billion to $3.4 billion for this year. The stock closed Tuesday at $206.86. Read that again. The contracted work on the books is roughly 11 times what the company expects to bill in all of 2026. The market is not pricing the backlog right now. It is pricing the financing. Cipher Digital $CIFR sold $850 million of zero-coupon converts Tuesday and fell 13% to a 52-week low, and the whole neocloud group took the hit. $NBIS is down about 11% for the week and sits roughly 19% below its recent peak near $255. Meanwhile the targets have not moved. The consensus price target is about $246, the high on the Street is $410, and management still points to a year-end run rate of $7 billion to $9 billion and 5 gigawatts of contracted power. One side is looking at the dilution risk across the group. The other is looking at a backlog that dwarfs this year's sales. Both are reading real numbers. Does a $37.5 billion backlog make the pullback a gift, or is group-wide dilution the bigger story?

The Alerian MLP ETF $AMLP yields about 7.75%, and it sits right in the path of AI's exploding power demand. Here is the setup. AI data centers need enormous amounts of electricity, and they need it soon. Nuclear is the clean long-term answer, but new plants take 7 to 10 years to build. That timeline does not match the demand showing up right now. Natural gas fills the gap. It is abundant, it is cheap, and the infrastructure to move it already exists. That puts the pipeline operators in a powerful position. They get paid to transport the fuel that keeps the data centers running. Enterprise Products $EPD, Energy Transfer $ET, and Kinder Morgan $KMI are three of the largest names moving gas across the country. These are not speculative AI plays. They are toll-road businesses collecting fees on volume, with income streams that tend to hold up across cycles. For investors who want exposure to the AI power story without chasing expensive chip stocks, the pipelines offer a different angle entirely. You get a real yield today, backed by assets that grow more valuable as power demand climbs. The market is obsessed with who makes the chips. Fewer people are asking who delivers the power. That gap is the opportunity. Would you rather own the chipmakers or the pipelines here?

Over $1 trillion is already spent on AI infrastructure, and the biggest buyers are just getting started. Arista Networks $ANET, NetApp $NTAP, and Hewlett Packard Enterprise $HPE all pushed to new all-time highs this week. That is the market telling you the AI buildout is still in its early innings, not its final one. Alphabet $GOOG and Microsoft $MSFT alone have poured more than $1 trillion into AI infrastructure over roughly two years. This year they are on pace to spend over $700 billion combined, with trillions more already planned. Here is the part most people miss. The demand side has barely shown up yet. Less than 1% of the global population pays for a top-tier AI program today. By one estimate, fewer than 10% of large companies are seeing a meaningful impact from AI so far. That gap between what is being built and what is actually being used is the opportunity. The companies supplying the networking, storage, and servers behind this spend are the ones printing new highs while adoption is still ramping. Nvidia $NVDA and Taiwan Semiconductor $TSM remain the backbone powering every one of these buildouts. When the picks and shovels lead, the trend usually has room to run. Which AI infrastructure name are you holding into year end?

Rare: $CEG just lit up seven bullish signal studies on the same day. Constellation Energy closed at $300.40 yesterday after a run that tripped a MACD bullish zero cross, an Aroon up cross, an RSI(2) overbought thrust, a CCI(20) push above +150, a golden cross, a top 1% gap up, and a volume surge of 4.4x its average, all at once. Here is how each has played out before, roughly a year later, over the nearly 5 years $CEG has traded: RSI(2) above 95: 98 prior cases, higher 90% of the time, +60% average. MACD bullish cross: 33 prior, higher 91%, +65% average. CCI above +150: 42 prior, higher 79%, +54% average. Aroon up cross: 19 prior, higher 89%, +64% average. Gap up: 11 prior, higher 82%, +45% average. Volume surge: 6 prior, higher every time, +32% average. Golden cross: 4 prior, higher every time, +123% average (small sample). Seven green lights on one name in one day is rare. Would you chase $CEG here or wait for the pullback?

$RKLB grew revenue from $244 million in 2023 to $601 million in 2025, and the stock just hit a one-month high. After a sharp summer correction, the space trade may be waking back up. We laid out the bull case last September: the real opportunity is not just rockets, it is space-based communication, surveillance, GPS, internet, and cell service. The macro backing is strong. Washington sees a high-stakes race with China for space dominance, which means regulatory and financial support for the U.S. space industry. Our top space names were $RKLB, $BKSY, $PL, and $ASTS. Many soared 50% to 100% soon after, helped by anticipation of SpaceX's $SPCX IPO in June. Then came the correction. No sector climbs in a straight line, and the space trade needed to work off a lot of froth. After SpaceX itself, Rocket Lab is America's largest space economy stock, so its chart is a real-time read on the whole trade. The company builds rockets, satellites, and spacecraft components, and is increasingly able to design, launch, and operate entire satellite systems. $RKLB found a floor in the $60 to $65 range, bounced, and reached a one-month high this morning. That is one constructive step, not an all-clear. This high-potential trade just moved back near the top of our watch list.

$ARKG and $BUG both hit all-time highs today. The tankers $INSW $HAFN $TRMD $DHT all hit new highs too. $ZETA just hit a yearly high, up 83% since our call. Meanwhile $PEP, $MCD, $HD, and $LULU all hit one-year lows. Money is leaving consumer names for AI and resources. Bookmark this for your watchlist.