Patrick

Why $AAOI is worth watching, especially after the drawdown. Every AI data center is a fiber optic network, and the thing that turns electricity into light inside it is a laser. Applied Optoelectronics is the only US listed optical company that makes its own lasers in America, it grows the indium phosphide crystals, fabs the laser chips in Sugar Land, Texas, and assembles the transceivers. In a buildout where hyperscalers are desperate for optical capacity, owning the laser fab is the moat. That's the thesis. And the market just cut the stock more then half in the past few months, on the guide, which is why it's interesting now. The quarter itself was a record. Q3 revenue was $269.9 million, up 127% year over year, a sixth consecutive quarterly record, and the company returned to GAAP profitability at $6.3 million. Datacenter revenue hit $155.6 million, up 254%, and CATV came in at $110.9 million, the top of guidance. What didn't land was the guide: Q4 revenue of $385–420 million came in below the $427 million Street bar, below the roughly $485 million implied by management's own full year math, and far below the $500 million plus the CEO talked about in August. Gross margin came in at 28.7%, missing the guided 29–30.5% range, dragged by a lower margin 400G mix. Here's what the selloff is missing imo. The demand is not the problem, supply is. 1.6T orders stand at roughly $400–450 million, and the Tier 1 hyperscale agreement signed in June is worth $1–1.5 billion over three years, covering at least 10% of that customer's 1.6T needs. A second agreement, possibly five times larger, is in final stages and could sign within weeks. The binding constraint is DSPs, the chips that drive the lasers, and management says supply improves substantially in the first quarter of 2027. Capacity is being built: the Sugar Land campus is expanding by 210,000 square feet, targeting over a million units a month by the end of 2027. The balance sheet holds $621.5 million in cash against minimal debt. Moat: the vertical integration is real, but it's partial. AAOI's fab makes the lasers for its CATV business and lower speed products. At the 1.6T tier, the 200G lasers and the DSPs come from the same merchant suppliers everyone else buys from. So the advantage is control over supply and cost at the speeds that pay the bills today, not a lock on the next generation. The customer concentration is extreme too: the top ten customers are 99% of revenue, one cable customer is 41%, and two datacenter customers are 28% and 23%. The Tier 1 deal's pricing is locked in, which caps the upside if market prices rise, and the customer committed no capital. My assumption is that DSP supply does improve in Q1 2027 as management says, that the second agreement signs, and that 1.6T revenue ramps from the $50–70 million guided for Q4 toward the $400–450 million order book. If that holds, a $4.2 billion company with a sixth consecutive revenue record, a $1–1.5 billion signed agreement, and a consensus target around $112, more than double the current price, is mispriced. But, Analysts did cut targets hard on Friday, Rosenblatt $220 to $115 and Needham $170 to $100, but both kept Buy ratings. If the DSP constraint drags on or the second agreement doesn't sign, the guide cut was the first of several, and the cheapness is deserved. That's the case. This isn’t a favourite, not a sure thing, a laser maker with real demand, a real capacity build, and a stock the market just more then halved on a timing problem. The next proof points are the second agreement and the Q1 DSP ramp. Until then, it's a trade on execution basically, not a bet on the AI buildout, the buildout is already priced into the orders. Now, do we trust management enough to execute?

Why $BWEN is my top power infrastructure play The name is a lie, and the lie is the opportunity. Broadwind Energy doesn't make wind anything anymore. In May it sold its last wind tower plant and became a precision manufacturer for what's actually getting built in America right now: gas turbines, grid gear, and the heavy machines that dig the minerals. The market still prices it like a dying wind stock. The exit is what tells you management sees the board clearly. On May 5, Broadwind sold its Abilene, Texas wind tower facility to IES Holdings for up to $19.5 million. The business it sold was not a loser, it did $56.3 million of revenue and $9.7 million of adjusted EBITDA in 2025. They sold a profitable business on purpose, because the economics underneath it were about to break: federal wind tax incentives expire by the end of 2027, and the segment lives or dies on Washington. They got out while the asset still had a buyer basically. What's left is two businesses with nothing to do with wind subsidies. Brad Foote Gear Works has made precision gearing and gearboxes for a century, the gears inside power generation equipment, oil and gas drilling, mining, steel, marine and defense. Industrial Solutions kits and assembles the components for combined cycle natural gas turbines. The CEO called them "higher growth, more predictable, more profitable" than the wind business. The demand is the AI power buildout. The EIA expects U.S. power demand to hit records in 2026 and 2027, with natural gas holding about 39 to 40 percent of the mix. GE Vernova just posted a record $197 billion backlog. Broadwind is the micro cap supplier inside that same supply chain, and its numbers are moving the same direction. In the June quarter, continuing revenue rose 67% to $24.3 million, orders jumped to $35.2 million, Gearing orders surged 138%, and Industrial Solutions posted record orders of $17.2 million and record revenue of $13.2 million at roughly a 19% segment margin. Combined backlog hit $85 million, up 93% year over year, with a book to bill of 1.5 times. Then on November 2, Broadwind bought MCM, an Alabama precision manufacturer serving electric utilities, for $18 million plus up to $7 million in earnouts. MCM did about $16.5 million of revenue and $3.7 million of adjusted EBITDA, a 23% margin, so Broadwind paid 4.8 times EBITDA for a business more profitable than its own. It's a strategic entry into electrical transmission and distribution: Broadwind now touches the turbine, the kit that builds it, and the grid it feeds. The market cap is roughly $64 million, about 0.4 times trailing sales, and the consensus target is $6.25, more than double the current price. The bear case, honestly: it's a $64 million micro cap trading about 500,000 shares a day, barely profitable on an adjusted EBITDA basis, and the top line is lumpy, orders arrive in chunks from a handful of large OEMs. The MCM deal was funded partly with debt, guidance was withdrawn in May, and the whole thesis rides on gas turbine demand staying hot. The Q3 print on November 12 is the first clean look at the new company. My assumption is the backlog converts on the 6 to 18 month schedule management has described and MCM integrates cleanly. If that holds, a $90 million revenue business growing orders 20 to 40% with a 23% margin acquisition bolted on is worth more than 0.7 times sales. If orders stall, it's a $64 million machine shop with debt, and the cheapness is deserved. That's why it's my top power infrastructure play. The wind story is dead. The power story is just starting.

Why $SYM is my top robotics play Most robotics stocks sell you a robot. Symbotic sells you the warehouse, the whole thing, as one AI run system, and the market is pricing it like a hardware vendor while it quietly becomes a software and recurring revenue business. That mismatch is the thesis imo. It was built by Rick Cohen, a third generation grocery wholesaler who spent fifty years running C&S Wholesale Grocers, the largest grocery wholesaler in America. He automated warehouses because the margins in his own business were too thin to survive on labor alone. The system was proven inside his own network, then sold to Walmart, which didn't just buy it, it made Symbotic its automation arm. The backlog is what makes this great. Symbotic carries roughly $22.5 billion in contracted backlog, about ten times annual revenue, and on October 29 Walmart committed another roughly $6 billion across 25 regional distribution centers, a deal Symbotic says could add up to $6.92 billion to the backlog, and one Walmart cannot cut without Symbotic's consent. The whole company is worth about $22 billion. You're paying roughly one times contracted backlog, and only about 15% of it converts in the next twelve months, so the growth is backloaded and visible for years. There are three engines. The Walmart rollout: 42 regional centers, 25 more under the new deal, under a contract running to 2037. Exol, the warehouse as a service joint venture with SoftBank: at least $7.5 billion of Symbotic systems committed, $11.6 billion of the backlog, six sites totaling roughly six million square feet. And micro fulfillment: up to 400 Walmart stores, with Walmart funding $520 million of development. Fox Robotics and cellumation, bought this year, widen the surface Symbotic can automate. The inflection is already happening. In the June quarter Symbotic posted its first operating profit, $32.9 million on $720.8 million of revenue, up 22%, with gross margin at 22.3%, up from 13.7% in fiscal 2024. Adjusted EBITDA more than doubled to $95 million. Software revenue grew 57% and operations services 49%: the recurring layer compounding behind the hardware. The company guides Q4 to $760–780 million of revenue, with results November 23. So why is it cheap? The stock is $36.82, down 58% from its high, while the business grew 22%, turned profitable, and its biggest customer committed $6 billion more. The market is pricing the deceleration, growth has slowed from 134% to about 21%, and the concentration: Walmart is 85% of revenue, over 90% last quarter, on cost plus contracts that cap margin. Profitability is brand new and thin, free cash flow was negative $165 million in the quarter, and a new customer takes about 24 months to bring live. If Walmart ever slows the rollout, the thesis breaks. That's the single point of failure, and it's why the stock is cheap. My assumption is the backlog converts at roughly the guided pace, about 15% in the next twelve months, implying revenue around $3.4 billion against $2.65 billion trailing, with gross margin holding in the low twenties. If that holds, the operating leverage is real. If it doesn't, it's a backlog story that never converts, and the cheapness is deserved. That's why it's my top robotics play. Tesla's Optimus is a demo with no revenue. Amazon's robots aren't for sale. Symbotic is the only scaled, revenue generating, now profitable pure play robotics company in the public market. The robots already run the world's largest retailer's warehouses.

I didn’t really have much time to dig into that $FOUR report, and for some reason I’m at the gym right now and it’s on my mind. So here are my thoughts… They cut guidance again, yes.. Same quarter, it printed a record $200M of free cash flow and held full year free cash flow guidance at $465-475M. EBITDA guidance got trimmed about 6%. The cash guidance didn't move. I like that. Management cites travel disruption and soft same store sales. What actually grew was payments based revenue less network fees, the line they call their north star: $402M, up 18%, with 14% of that organic. Here's the part I find interesting tho. Tax free shopping is the piece most exposed to travel, and it's also the piece just getting started. Shift4 One puts payments, currency conversion, and tax refunds in one device that detects which transactions qualify. A foreign card swipe in a European shop can now earn three ways instead of one. It's live in 15 countries, ahead of schedule, and the Global Blue integration is nearly done, with $50M of annualized synergies expected by year end. Then look at what the company did with the news. It authorized a new $1B buyback right behind one that's almost used up, and said the stock price doesn't reflect the business. At roughly a $3.6B market cap, that's about 7x free cash flow, before counting the debt. The risk tho, is credibility. That's at least two guidance cuts this year. Growth is closer to 10% than the 34-51% headline rates the Global Blue deal produced, and the debt load is quite real. The forecast got cut. The cash didn't. The stock only reacted to one of them.

Just bought back into $KRKNF $PNG.V here, can’t believe it’s gotten this cheap…

OMG IM GREEN LOL $AAOI 1-3

No crying in the casino

0-3 Do not ask me for an earnings play 😂 $AAOI

I’m 0-2 with earnings plays this season BUT I’m in $AAOX for $AAOI tonight. Wish me luck

I genuinely don't get the "AI bubble" crowd. And at this point it's not even really about AI, it's a "big tech is a bubble". So let's remind ourselves what big tech actually just did. $MSFT Microsoft: $92.3B revenue (+19%), Azure up 48% and accelerating. Google Cloud more than doubled to $30.7B with operating income tripling. AWS: $46B (+39%, fastest in five years), and Jassy guided 40%+ growth for 2027 AND 2028. $NVDA NVIDIA: $96.2B revenue (+106%) at a 62% net margin. But forget AI for a second. $NFLX Netflix, a streaming company, nothing to do with AI, just did $12.87B in revenue with net income up 44% to $3.66B, a 34.6% operating margin, nearly $3B in free cash flow, and bought back a record $4.8B of its own stock. $META Meta's ad business did $63.1B, up 26%. $GOOG Google Search did $63.4B, up 12%, with query growth at its fastest in five years. $AMZN Amazon posted its highest operating margin ever. Microsoft's boring old Office cloud business did $27.8B, up 18%. Where exactly is the bubble? In the companies printing record profits? Here's the part the bubble callers never mention: The AI demand is contracted. Microsoft's sitting on $677B in remaining performance obligations (+73%). Google Cloud's backlog is $561B. AWS's is $578B. That's roughly $1.8 trillion in signed future revenue. Bubbles don't come with $1.8 trillion in purchase orders. "But the capex!", over $700B this year across the hyperscalers. Funded by what? Record operating income. Amazon's record margins. Microsoft still free cash flow positive while spending $45B a quarter. This isn't 2000, when telecoms borrowed money to lay fiber nobody used. These companies are spending cash they earned because customers are buying. Could stocks pull back 20%? Sure, that's markets. But a bubble is when prices detach from reality and the emperor (I just watched Star Wars) has no clothes. This emperor is wearing a $1.8 trillion suit of signed contracts, and the fitting room is full of companies doing record buybacks. I don't know what else people need to see.

$MU $DRAM $SKHY $SNDK

I had a lot of faith in $RDDT, I really did. But after that earnings call I decided to sell 80% of my position this morning. Unlucky, onto the next.

$RDDT The good: - Revenue $895M (+53%), beat the Street ($874–886M) and the top of their $860–870M guide - EPS $1.40 (+75%), beat $1.30–1.33 - Adj. EBITDA $402M at 44.9% margin, beat the $385–395M guide - Free cash flow $363M, nearly doubled (+98%); operating cash flow $364M (+96%) - Q4 guided to $1.0–1.03B revenue, their first billion dollar quarter guide, with EBITDA $480–510M - US ARPU $13.82 (+53%), gross margin 91.5%, $3.03B cash, added to the S&P 500 The bad - US users fell to 50.7Mc down 2% YoY and down 2.5M sequentially from 53.2M, versus 53.37M expected. - Global users flat at 130.3M sequentially, missed the 133.36M consensus. The great - The Q4 guide: $1B+ in a quarter - Cash generation exploding, 41% operating cash flow margin - They're extracting 53% more revenue per US user than a year ago. The monetization engine is the best in social Overall, the user number I was watching went the wrong way. The financials are GREAT tho, 895M revenue (+53%, a beat), $1.40 EPS (+75%), $402M of EBITDA at 44.9% margins, free cash flow nearly doubling to $363M, and Q4 guided to $1.0–1.03B, their first billion dollar quarter. US ARPU up 53% to $13.82. This is a monetization machine. But US daily users fell to 50.7M, down 2% year over year and down 2.5M sequentially. Logged in US growth flat at +1% for the second quarter running. That's the same metric that took the stock down 22% last quarter, and it's why shares are off 11% after hours to a new 52 week low. The lesson from Q2 stands: this market will not pay up for Reddit while US users shrink, no matter how good the P&L looks. unlucky.

gg $RDDT

My call for $RDDT US DAU: flat to slightly higher. U.S. DAU around 53.3-53.6 million, roughly in line with the 53.37M consensus. The bar is low tbh, and that's the first thing to understand. Zacks consensus has U.S. DAU at 53.37M versus 53.2M last quarter, so the Street is only asking for +0.17M sequentially, essentially flat. Year over year it's +3.4% (from 51.6M). Nobody is expecting a reacceleration… they just want the bleeding to stop basically. That means "flat" is a win here, and only a print below ~53.2M (another sequential drop) is a real miss. The best bullish evidence I found: Needham's traffic work. On October 9, Laura Martin reiterated her Buy and $300 target specifically citing *accelerating* traffic metrics in Q3, with app downloads up 19% quarter over quarter. That matters more than it sounds, app downloads are a leading indicator for logged in users, which is exactly the segment that was weak (+1% in Q2). If the top of the logged in funnel grew 19% in Q3, flat to up U.S. DAU is very achievable. But, the bearish evidence is also here: Wedbush cut its U.S. DAU estimates on October 23 while keeping its Buy rating, an analyst with channel checks seeing pressure, six days before the print. And Semrush's September data showed Reddit web visits down 8.2% month over month, with 57% of traffic coming from Google organic search. Two caveats on that: Semrush only measures web, not the app (where Reddit is ~73% mobile and actively pushing users), and September month over month has seasonal noise. But it confirms the search dependence risk is live I think. The Needham app download signal is the freshest and most directly tied to the weak spot (logged in users), while the Wedbush cut and Semrush data are softer/older signals. The Q2 dip was also tiny, 0.6%, and attributed to choppy referrals. Print below 53.2M and it's Q2 all over again, no matter what revenue does. The good news is the bar is so low that "flat" counts as good news in my eyes.

$AMZN $INTC $SNDK $RDDT all report after the close today.

$GOOG just told us why $RDDT ‘s human content just keeps getting more valuable… Google admitted on the call that rapid AI Mode adoption is weighing on Search monetization because those queries generate less revenue, even as query growth hit a five year high. That confirms the AI search pressure on referral traffic, Reddit's known headwind, but it also shows Google itself can't monetize AI search well yet, which is exactly why Reddit's human content keeps getting more valuable, not less.

$RDDT reports Q3 earnings tomorrow after the close, and I'm very bullish going into it. This is a management team that has not let shareholders down on execution. They guide conservatively and still deliver. Q1 revenue of $663M (+69%) beat estimates. Q2 revenue of $805M (+61%) beat by $74M, the eighth straight quarter of 60%+ growth. They guided Q3 to $860–870M, and the Street already expects them to beat that too. They've turned Reddit into a high margin cash machine: 91%+ gross margins, a 43% adjusted EBITDA margin in Q2, $262M in operating cash flow in the quarter, and over $1B in trailing operating cash flow for the first time ever. As Steve Huffman put it: "In an increasingly automated web, the value of real human perspective has never been higher." The ad platform keeps scaling, ad revenue up 64% in Q2 to $762M, advertiser count up 70%, Reddit Max revenue up over 150%, while international revenue (+84%) keeps offsetting any US user softness. And Huffman said it best on the last call: "We're not building for drive-by traffic. We're building a daily destination." Meanwhile the stock is down roughly 45% from its highs and closed today at $134~, below both its 50band 200 day averages. And remember, last quarter they beat on everything and the stock still fell 22% in a day. Now, the fun part… My call: Reddit beats, and the stock pops about 12% on Friday to roughly $150. The numbers I want to see: Revenue around $890 million (above the $875-886M consensus and well above their $860-870M guide), EPS around $1.40 (above the $1.30-1.33 consensus), and adjusted EBITDA margin near 44-45%. On users, I expect global DAU around 133 million and, critically, U.S. users flat to slightly up sequentially. For Q4 guidance, my own estimate is something around $950M-$1B, Why I land there: Eight straight quarters of 60%+ growth, a Q2 beat by $74M, Reddit Max revenue up over 150%, and international still compounding near 80%. Management has earned the benefit of the doubt on delivery. On the reaction side, the setup is there imo, the stock is down 45% from its highs, RSI in the mid 30s, 11.5% short interest, and the options market is pricing a ~12% move either way, so a beat into an oversold, heavily shorted name has potential. The risk to my prediction: last quarter is the tell for how this goes wrong. Q2 was a monster beat and the stock still fell 22% the next day because U.S. users slipped. If U.S. users slip again tomorrow, my prediction is wrong and you'll be looking below. That's the single variable that decides it imo.

Can someone tell me how $MRVL is in advanced talks into buying $LITE?? Math isn’t mathing here