Bought MAG7 for a long time, but actually many people don't even know what these 7 companies really make money from. A table explains it directly: Company｜Real Cash Cow $AAPL｜iPhone + Services $MSFT｜Office/Azure/Enterprise Software $GOOGL｜Google Search Ads + YouTube $AMZN｜AWS + E-commerce Ads $META｜Facebook/Instagram Ads $NVDA｜AI GPUs + Data Centers $TSLA｜Cars + Energy Storage The three most easily misunderstood here are: Google is essentially still an advertising company; Amazon's real profit doesn't come from selling goods, but from AWS and advertising; Meta talks about AI every day, but in the end AI is still used to help improve ad efficiency and user engagement time. So although MAG7 often rise and fall together, they are fundamentally not the same kind of business. I think MAG7 should not be compared only by PE, but first look at: Where the money comes from → What drives growth → Whether AI is actually creating new revenue or just improving the efficiency of existing business. Understanding these three questions is much more useful than guessing stock price fluctuations every day.

When the 10-year US Treasury yield approaches 5%, many people's first reaction is: High interest rates = banks making money. But it's not that simple. The most comfortable environment for banks is not "the higher the interest rate, the better," but rather: High loan yields, but deposit costs don't rise too quickly, and the economy doesn't collapse. So now when I look at bank stocks, I categorize them into three types. For large commercial banks like $JPM and $BAC, I mainly look at net interest margin, loan growth, and bad debts. For investment banks like $GS and $MS, I pay more attention to when IPOs, mergers and acquisitions, and trading businesses truly come back. Then there are regional banks, which carry higher risks because they rely more on deposit costs, commercial real estate, and the quality of local loans. So if US Treasury yields continue to rise, I wouldn't directly call it a "bank bull market." The truly good scenario should be: Interest rates remain at a relatively high level → the economy does not enter a recession → companies continue to borrow → bad debts do not explode → capital market trading becomes active again. Only then are banks truly comfortable. The worst case is: Interest rates are very high, but the economy is dragged down by the high rates. In that case, the interest earned in the end might not even be enough to cover the bad debts.

gpt has changed the world here $NVDA gpu $MU $SKHY storage Two indispensable hardware components may take off here due to gpt's new model!

Recently, after reviewing the earnings reports of $DELL, $AVGO, and now $HPE, I am increasingly certain of one thing: The biggest problem with AI hardware right now is no longer whether there are orders, but whether there is enough product to sell. HPE's latest quarterly revenue was $12.21 billion, up 33.7% year-over-year, and they have once again raised their full-year and next year's performance forecasts. What really matters is not how much they beat expectations by. The management directly named the items with the tightest supply right now: Memory, NAND, CPU, Drives. This statement actually explains the entire AI hardware market. Servers continue to expand → DRAM/HBM remains in shortage → NAND/SSD demand continues to rise → CPU and networking equipment increase accordingly → Finally, this extends to optical modules, power, and liquid cooling. So, putting the recent earnings reports together: $DELL: AI server backlog of $95 billion $AVGO: AI semiconductor revenue +221% $HPE: Directly states demand far exceeds supply Three different companies, but the same answer: AI CapEx has not stopped at all. This is also why I no longer focus solely on $NVDA when looking at AI hardware. What’s truly interesting is the next layer: SK Hynix / $MU — Memory $SNDK — NAND/SSD $ANET — Networking $LITE / $COHR — Optics $VRT / $ETN — Power Previously, the market worried: Will AI servers have no buyers? Now it increasingly seems: Customers want to buy, but the supply chain can’t deliver. As long as this situation persists, it’s hard for me to be bearish on AI hardware. The next real opportunity to look for might not be the “next NVDA.” But rather: The next item that starts to become scarce.

Now that the Middle East conflict has escalated, people will immediately start buying $LMT, $RTX, $NOC on the timeline. But I think the biggest mistake military stocks often make is: Chasing after war. Because what truly determines the profits of defense companies in the coming years is not how many missiles are launched today, but whether countries will continue to replenish inventories, increase military budgets, and sign long-term contracts after the war. So I don’t just look at news headlines when it comes to defense stocks. I prefer to look at three things: Backlog, Book-to-Bill, and capacity expansion. For example, if a defense company has quarterly revenue of $10 billion but new orders of $20 billion, it means it’s not just living off old business, but the future revenue pool is still growing. This is exactly where $LMT, $RTX, and $NOC become truly interesting. Missiles, interception systems, air defense, fighter jets—once used, they are not replenished the next day but require years to reproduce. So the real investment logic left by war is not: "War is happening today, so defense stocks rise." But rather: After global weapons inventories decline, who will be responsible for refilling the warehouses in the coming years? In the short term, watch the war. In the long term, watch the orders. I’d rather study defense stocks when no one is talking about war than rush in to chase prices just as missiles land.

If nothing unexpected happens, we should see new highs this month $NVDA honestly is really scary….. $GOOG probably a long way off $TSLA maybe next life 🤣

From selling shoes out of a trunk to becoming the world's strongest sports brand, and now being kicked out of the S&P 100. I still remember back in 2021 when I was in high school, everyone was damn paying a premium for Nike And it was several times over $NKE, in these 60+ years, is almost a history of American consumer brands. In 1964, Phil Knight and his track coach Bill Bowerman founded Blue Ribbon Sports. At first, they didn’t even have their own shoes; they imported Onitsuka Tiger from Japan, which is today’s ASICS. Knight even drove to track meets to sell shoes directly. In 1971, the company officially changed its name to: NIKE. That same year, the Swoosh logo, which would later be worth billions, was born. In 1984, the most critical event in Nike’s history happened: Signing Michael Jordan. Air Jordan not only saved the basketball shoe business but also fundamentally changed the entire business model of sneakers, celebrity endorsements, and street culture. In 1988: Just Do It. Nike began to transform from a company selling sports shoes into a global cultural icon. In the following decades, Air Max, Jordan, Dunk, Air Force 1—generation after generation of hit products pushed Nike to the top spot in global sneakers. By the end of 2021, Nike’s market value once exceeded $260 billion. But the problems started here. One of Nike’s biggest mistakes in recent years was putting too much faith in DTC (Direct to Consumer). They wanted to reduce reliance on retailers like Foot Locker and JD, pulling all consumers back to the Nike App, official website, and direct stores. In the short term, profit margins were higher. But the side effects were obvious: Worsening channel relationships, reduced exposure, and slower product innovation. In recent years, Hoka, On, Adidas, Anta, and Li Ning have aggressively taken market share, while Nike increasingly relied on older shoe models. This is especially evident in the Chinese market. In the latest quarter, Greater China revenue dropped directly by -17%. For fiscal year 2026, total revenue is only $46.4 billion, basically flat compared to the previous year, and even lower than $51.4 billion in 2024. Net profit also fell from $5.7 billion in 2024 to about $3.1 billion now. (U.S. Securities and Exchange Commission) The stock price is even worse. At the end of 2021, Nike’s market cap was about $264 billion. Now it’s only about: $57 billion. A shrinkage of nearly 80%. Then this week, something very symbolic happened: Nike was officially removed from the S&P 100. Replacing it was: $PANW. Also entering were: $DELL $ANET $SNDK A server company, an AI network, a cybersecurity firm, and a storage company. And the ones removed were: Nike, Simon Property, Colgate, Honeywell Aerospace. (News Release Archive) So I think the real interesting thing about this adjustment is not: "Nike is finished." But rather: The weighting of America’s top 100 core companies is shifting from traditional consumer brands toward AI and tech infrastructure. Nike once built a consumer empire through Jordan and brand marketing. Now the market prefers to value: Servers, networks, storage, security. Of course, Nike is far from dead. Since Elliott Hill returned, wholesale channels have been repaired, discounts reduced, and there’s a renewed focus on running and performance products. Signs of recovery are also appearing in North America. (https://t.co/UpMwU9WAm0) But the problem is: Nike no longer needs to prove whether it’s still a good brand. Instead, it needs to prove: Whether this 60+ year-old consumer empire can create a new Air Jordan for the next generation.

Many people say storage is just a cyclical stock, right now it’s rising only because of price hikes, and it will fall sooner or later. I used to think so too, until I spent a long time studying AI. Look, training a large model with GPUs requires tens of thousands of dollars in cards, but what really makes AI run is not just computing power, but also data — and data must rely on storage. You can choose not to buy NVIDIA, but it’s hard not to buy SSDs. In the future, AI won’t lack data, but will lack storage space. When the whole world is generating videos, generating 3D, generating Agents, storage will transform from a “cyclical product” into “infrastructure.” By the time you realize this, it might not just be a few times growth, but an entire era. “But I still think storage will fall sooner or later” 🤔 (except for my $SNDK 😘 seeing $10,000)

$ZEC has surged wildly again, but this rally isn't driven entirely by spot buying. This gain dwarfs $BTC's performance over a week or even a month! ZEC is currently trading around $1190, up nearly 18% in 24 hours, reaching a high of $1204. The first catalyst is Grayscale's ZCSH spot ETF: since its launch on August 25, assets under management have risen from about $305 million to $415 million, truly opening institutional access. The second catalyst is the return of the privacy coin narrative, with the market re-trading on-chain privacy demand. The third and most intense factor: after breaking $1000, a large number of shorts were forced to cover, with about $34.5 million in short positions liquidated in the previous round, and leveraged funds further amplifying the gains. But I want to remind you: there is a clear short squeeze component in this rally, with futures activity stronger than spot buying; the steeper the rise, the faster the potential pullback. In terms of levels, $1200–1205 is the immediate resistance; if volume supports a hold above this, then look to $1250. Only a break above $1250 opens the chance to push to $1300. On the downside, first watch $1120–1100 for support; if broken, a retest of $1050 is possible. The $1000–1020 range is the most critical boundary between strength and weakness in this move. My view is simple: don't chase at this level, nor short against the trend aggressively. Either wait for a low-volume stabilization near $1100 or wait for a breakout above $1205 followed by a pullback confirmation. There is no comfortable risk-reward ratio in the middle of this range. #ZEC现货ETF首日成交额1480万美元