To The Moon

$Broadcom(AVGO)$ isn’t just designing the chips anymore. It’s helping finance the customer buying the compute. Broadcom is providing Anthropic with a $42B convertible note facility. Here’s why that matters: 💰 Anthropic has committed $125.2B to lease $Alphabet(GOOG)$ TPU compute over five years. 🏦 Broadcom’s $42B facility could cover roughly one-third of that commitment. 📈 Broadcom can potentially convert the debt into Anthropic equity. ⚡ New TPU capacity is expected to start arriving in 2027, when Anthropic could become Broadcom’s largest compute customer. So the structure gets interesting: Broadcom designs the chips → helps finance the customer → locks in future demand → potentially receives equity upsi

🚨 The receivables argument against $Micron Technology(MU)$ is looking very different this quarter. Last quarter, Michael Burry and other bears pointed to Micron’s receivables growing faster than sales, suggesting customers might be taking longer to pay. This quarter tells a different story. 📈 Revenue: +31% Q/Q 📈 Receivables: +17% Q/Q 💰 Customer prepayments: $12.3B That $12.3B is the number I’m watching. Customers have effectively handed Micron $12.3B upfront through long-term contracts. That’s more than the $11.1B Micron spent on capex during the quarter. Some of those contracts extend through 2031, with the associated cash paid back over the second half of the contract periods. So the cash-flow dynamic looks very different from a customer-payment-

As expected, the panic around $Nu Holdings Ltd.(NU)$ ’s potential Monzo acquisition turned out to be a non-event. As I suggested two days ago, the deal is off, reportedly because Monzo’s VC investors wanted a valuation that was too high. But the two most important takeaways from $NU’s announcement are about its strategy and capital discipline. 1️⃣ $NU believes its existing growth opportunity is big enough. The company reaffirmed its focus on deepening its position in Brazil, scaling in Mexico and Colombia, and building its presence in the U.S. and globally through Nu Global. In other words, $NU doesn’t need to chase acquisitions just to keep growing. 2️⃣ Every acquisition has to clear the return hurdle. The company explicitly said investment decisio

In 2026, $Micron Technology(MU)$ went from a sleepy memory maker to one of the most profitable companies on the planet, with the stock up about 273% this year. Yet the market still prices it like a cyclical stock at the top of its cycle. After the Q3 report in June, MU jumped above $1,200, then fell roughly 40% over the following month as investors bet that memory prices were about to peak. Yet, Micron just posted its 6th revenue record in a row, and the numbers still look unreal. Revenue hit $54.2B, +379% Y/Y, above the high end of guidance. Gross margin climbed to 87.0%. Data center SSD revenue reached nearly $10B, more than 10x the year-ago quarter. Signed long-term contracts jumped from 16 to 26, with roughly $150B in RPO. Q1 FY2027 guidance o

$Take-Two(TTWO)$ is down more than 20% since July, even as the biggest release in gaming keeps getting more details. Instead of “buy the rumor, sell the news,” investors seem to be selling the rumors around potential delays or issues. Meanwhile, GTA 6 remains on track for November 19. 🎮 Rockstar just dropped a wave of new details: 🗺️ Map: roughly 2x the size of GTA 5’s 🌪️ Weather: hurricanes and tropical storms, with conditions changing across Leonida 🐊 Wildlife: 170+ animal species, including 60+ bird types and 30+ land mammals 🏄 Activities: scuba diving, base jumping, zoos, mini golf, date hikes and more 📸 12 new screenshots released alongside the coverage Game Informer’s 14-page GTA 6 cover story even caused its website and app to crash when it

$Nu Holdings Ltd.(NU)$ is now generating $1B+ in net income per quarter, growing at roughly a 30% CAGR — yet the stock is down 20%. That disconnect is what has my attention. 💰 $59B market cap 📉 ~13x forward P/E 🔄 $500M buyback launched in Q2 📈 30% CAGR The market is focused on macro volatility in Brazil and whether $NU can comfortably fund an ~$11B acquisition. But zoom out. $NU is expanding beyond its original customer base in Brazil, pushing further into the middle-income segment while continuing to grow in Mexico and Colombia. The business profile is evolving quickly. And now $NU has expanded into the U.S., opening another potential avenue for growth and a different narrative around the company. One thing that stands out: despite a ~$59B market

Most traders start with the wrong question: Are you bullish or bearish? The first question should be: Is this market environment even worth deploying capital into? You can have the right directional bias and still get chopped up if the market isn’t giving you clean conditions. My framework is simple: 1. In or out2. Then long or short Directional bias comes after the decision to participate. You don’t need 20 great trades a month. You may only need 2–3 big, high-conviction trades in a month or quarter to make your year. Everything else can be scratches. That’s fine. The goal isn’t to trade constantly. Participate selectively. Then press when the environment, direction, and catalyst all line up. And Q4 is starting to look like a period I want meaningful exposure to. No need to FOMO yet. But

Data centers get most of the attention around $Bloom Energy Corp(BE)$ . But factories have the same problem: they need a lot of reliable power, and they can't afford to wait years for the grid. A new chip fab can cost $20B+ to build. Leaving that facility idle for years because utility infrastructure isn't ready isn't an option. That’s where Bloom's on-site power systems become interesting. $BE is already deployed across major industrial sites, including: • $Intel(INTC)$ foundry expansions • $Ferrari NV(RACE)$ 's Maranello factory, cutting fuel use by roughly 20% • $Quanta(PWR)$ 's California facilities, with a $502M order

$Micron Technology(MU)$ reports earnings tomorrow, and the headline numbers may not tell the whole story. The biggest thing I’m watching is HBM demand tied to $NVIDIA(NVDA)$ ’s Rubin platform. Susquehanna analysts believe some HBM4 revenue originally expected this year could slip into 2027 as Rubin ramps. That could make Micron’s quarterly results look uneven, with some revenue shifting between quarters. But that’s timing, not necessarily a change in the demand trajectory. Rubin requires enormous amounts of high-bandwidth memory, and Micron is one of the few suppliers positioned to serve that demand. That’s why the market will be listening closely to management’s commentary on Rubin, HBM4 supply and demand

Michael Burry just moved up his AI crash bet, saying the “bubble in AI may burst sooner than later.” He closed out his short positions and moved into puts: • $Micron Technology(MU)$ — June $500 puts, stock around $1,080 • $NEBIUS(NBIS)$ — June puts below $100, stock around $230 • $iShares Semiconductor ETF(SOXX)$ — Sept 2027 puts in the low $400s • $Palantir Technologies Inc.(PLTR)$ — larger Sept 2027 put position in the low $100s His thesis points to research questioning how much AI revenue is actually proven, along with warnings that memory cyclicality could return. I see the setup differently. 1.Memory is cyclical. That