美股投资young（求回本版）
Feed
Feed
If the market really enters a prolonged period of high interest rates, I think the easiest to be eliminated won't be tech companies, but three types of business models.
The first: those that always need financing to survive.
When interest rates are low, burning money for growth is fine; when interest rates are high, each round of financing becomes more expensive.
The second: profits always come five years later.
The premise for the market to pay for the future is that money is cheap enough. Once the risk-free rate remains high for a long time, long-term stories naturally get discounted.
The third: growth looks fast, but cash flow never comes out.
These companies are the most attractive in a bull market, but when funding costs rise, they are also the easiest to be repriced.
So if interest rates really stay high for a long time, I would increasingly favor:
Cash flow machines like $GOOGL, $META, $MSFT, $AMZN,
and companies like $NVDA, $AVGO that have truly turned AI demand into profits.
High interest rates won't kill growth.
What it really kills is:
Companies that only have growth stories but no ability to make money.
BTC surges back to 80,000, ETH starts to catch up, US stocks take off unevenly
$BTC is currently around $81,200, up over 5% intraday, quickly rebounding from about 77,000. Standard Chartered has opened institutional BTC and ETH spot trading in the UAE, continuing to expand incremental inflows; but after the big surge in August, high-level locked positions remain. Holding 80,000, next target is 82,000; if it falls back to 79,000, be cautious of a pullback after the rally.
$ETH is around $2,510, up nearly 5% intraday, finally catching up with BTC. Traditional financial inflows continue to expand, but 2,600 is the real resistance; if 2,450 holds, the catch-up rally can continue. After a volume breakout above 2,550, look towards 2,600; without volume, don’t chase after a long bullish candle.
$OKB is about $111, up nearly 5% intraday, with X Layer and fixed 21 million supply still the core logic. The 108–110 range is a support zone; holding above 112 targets 116; falling back below 106 indicates this rally is mostly just following the broader market recovery.
Looking at US stocks, $AVGO’s AI revenue surged 221% year-over-year in its earnings report, but the stock dropped about 4.6% as the market found the expectations insufficiently surprising; $DELL rose over 5%, with AI server orders at $60.9 billion and backlog at $95 billion, hardware demand remains strong; $SNOW rose nearly 20%, raising its product revenue forecast, indicating capital is starting to reward software companies that can turn AI into revenue again. Tonight’s watershed moment: good results are no longer enough, they must exceed market expectations.
#BTC加速拉升，资金还能继续接力吗？
Putting the latest three earnings reports of $SNOW, $DELL, and $AVGO together, my conclusion becomes even clearer:
AI software is starting to take off, but the strongest performance right now is still in hardware.
First, look at $SNOW:
Total revenue of $1.55 billion, +35%, product revenue $1.49 billion, +37%, full-year guidance continues to be raised, and management said about half of the recent growth acceleration comes from AI.
This earnings report is very good, proving that AI software is finally starting to charge fees and truly drive performance.
But looking at hardware, it's a completely different scale.
$DELL:
Quarterly revenue $47 billion, +58%
AI server revenue $16.4 billion
Single quarter AI server orders $60.9 billion
Backlog directly hits $95 billion
Then look at $AVGO:
Quarterly revenue $29.59 billion, +86%
AI semiconductor revenue $16.7 billion, +221%
Even more astonishing is management's direct outlook:
FY27 AI revenue about $115 billion
FY28 about $230 billion
So now there is a very obvious phenomenon in the AI industry:
Software is proving "AI can make money,"
but hardware has already entered a phase where:
Customers place orders worth tens of billions of dollars directly.
The reason is simple.
Software deployment can be tried and expanded gradually;
But once AI servers start to be built, GPU/XPU, HBM, networking, optical modules, storage, power—all must be purchased upfront.
So my current understanding of the AI market is:
Look at penetration rate for software, look at orders for hardware.
And what truly gives me certainty about performance in the next two to three years is still the latter.
$SNOW proves AI software is not a bubble.
But $DELL + $AVGO seem to be telling the market:
This round of AI CapEx is at least far from over right now.
Many people hear about AWS, Azure, and Google Cloud every day, but they actually don't understand the real differences between these three.
A table to understand the three cloud computing giants:
Company｜Core Strength
$AMZN AWS｜Largest scale, most comprehensive products, mature developer ecosystem
$MSFT Azure｜Strong enterprise clients, deeply integrated with Office/Windows
$GOOGL Google Cloud｜Strong in data analytics, AI, and developer tools
Simply put:
AWS is like the "Walmart" of cloud computing—has everything and the widest customer coverage.
Azure's biggest advantage is enterprise relationships.
Many companies already use Windows, Office, Teams, so they naturally put their cloud on Microsoft as well.
Google Cloud leans more towards data and AI.
Especially in large models, data analytics, and machine learning, Google naturally has technical accumulation.
So as AI develops further, the real beneficiaries are not just GPU sellers.
Because all AI applications ultimately need to:
Store data → Rent computing power → Tune models → Run services.
Behind all this, cloud is indispensable.
So when I look at the AI industry chain, cloud computing is actually the most easily overlooked layer.
Hardware determines whether AI can run; cloud determines how AI truly becomes a business.
