Powerpei

The three major indexes were all up
Dow +0.53%, S&P +0.7%, Nasdaq +1%

Looks pretty good, right? Breaking it down tells a different story

Amazon surged 15.3% in a single day
This one stock is basically the engine driving the index higher
Strip out Amazon’s contribution and the gains in the S&P and Nasdaq shrink a lot

Now look at other sectors:
The Philadelphia Semiconductor Index jumped more than 5% at one point during the day, but closed up only 0.07%
Micron and Sandisk-type memory stocks spiked and then gave it all back
Russell 2000 small-caps fell directly by 0.5%

One number that explains the problem best:
In the S&P 500, there are more stocks declining than advancing—the ratio is 1.3:1

The indexes are rising, but most stocks are falling

This kind of “one-flower-show” rebound is something I’ve seen plenty of times

It usually happens in two situations: when sentiment repair is just starting out, or when hedge funds are forced to cover positions
It looks lively, but the durability is limited
A true trend needs more sectors to move together, broader participation from capital—not propping the whole thing up with just one stock.

My take:

This rebound in AI-related names looks more like a technical correction after an earlier oversold period. Amazon’s earnings report really did blow the doors off—but one earnings report can’t save the entire market. Going forward, if more companies’ results don’t follow through and there isn’t wider capital inflow, the market can’t sustain a picture propped up by a single stock.
Don’t just stare at the index levels—you’ll get fooled. Look at market breadth and sector confirmation, they’re more useful than specific price points.

Not investment advice; official data prevails.
#美股

The most interesting moment in today’s US stock market: Amazon and Apple released earnings reports on the same day, and their stock moves went completely in opposite directions.

Amazon blew past expectations and surged right after hours. Apple, on the other hand, fell.

The fact that capital picked a side so decisively shows that the market now only cares about one thing: whether your AI investment is actually making money.

Why did Amazon rise? AWS cloud business continues to expand, and AI workloads are being converted into real revenue.

What the market feared most before was “burning a lot of money with no visible returns.” This earnings report answered that directly: it can earn—and it’s accelerating.

Why did Apple drop? It wasn’t that the results were bad. The story of AI monetization just hasn’t been told yet.

How much incremental revenue will the Apple Intelligence subscription really bring? This earnings report didn’t provide an exciting answer.

Plus, the earlier news about price hikes has already been priced in—when the good news is realized, it turns into a negative.

The Nasdaq led the day, and all three major indexes moved higher overall.

The core driver is this: the return on AI capital expenditures has been reaffirmed. Not all money-burning is useless—the key is who can burn out real revenue.

My view:
The “Seven Giants” will continue to diverge. If they can prove that AI spending is turning into monetization (Amazon, Microsoft), capital will keep flowing in. If they can’t clearly explain the returns (Apple), even with a huge market cap, the stock will still be dumped.

This earnings season is basically wrapping up this week. The takeaway is clear: the market no longer buys the AI concept—it only buys AI revenue.

Not investment advice; data is based on official sources.
#美股

WLFI’s Super Node mechanism—after I finished reading the governance documents, I felt this is the toughest move in the USD1 ecosystem

The rules are simple: lock 50 million WLFI for at least 180 days.
At today’s price, that’s roughly $2.7 million.

So what do you get?
➢ You can integrate with authorized market makers to swap USDT/USDC 1:1 into USD1
➢ You can directly negotiate partnerships with the WLFI team
➢ You receive additional economic incentives
➢ Voting weight is weighted by the locked amount plus the remaining time

//

It sounds like an exclusive channel for the wealthy, but here’s the logic behind it:

USD1 expanded too fast before. Market makers made money through a mint-sell arbitrage cycle, taking a 15-basis-point spread each round—and the project had to subsidize the redemption cost.

Now this Super Node setup is essentially taking that profit back from market makers and distributing it to people willing to lock up long term.

There are too many partnership requests for the team to screen. So they simply use $2.7 million as the filter.

If you truly投了 (invested/participated), you have the right to sit down and talk.

//

What I find interesting is this:

It tightly links USD1 supply growth with the depth of WLFI locking.

Big holders who want to be super nodes naturally become promoters of USD1—and even act as mini distributors.

This is structurally much stronger than launching empty-airdrop programs or subsidizing trading volume.

Voting also requires at least two votes to obtain about a 2% staking reward—pure “sit back and do nothing” won’t work.

But I also won’t pretend I don’t see the problems:

A $2.7 million threshold means retail investors basically have no chance.

Even though voting power uses square-root weighting, people with more money still have louder voices.

————

My take:
Compared with governance designs where “everyone can vote” but nobody truly votes, this one is far more honest.

It directly admits one thing:
What can actually drive a stablecoin ecosystem is capital—resources—and people willing to bind long term.

USD1 needs sustained real demand and a distribution network, not short-term speculation.

For most people, the Super Node is a mirror: it helps you see what kind of participants this project wants to attract.

Not investment advice; data should be verified against the official governance documents. #USD1

The US-South Korea $950 billion chip order has just landed
SK hynix has locked in Nvidia’s five-year HBM supply.

Taking a step back,
China’s memory makers’ window of opportunity actually looks clearer.

Changxin Technology makes DRAM, not HBM.

But the logic is the same: global memory demand is rising; overseas capacity is being tied up by the AI supply chain; and the share of domestic replacement will only keep increasing.

In the short term, the sentiment toward the tech sector is indeed hot.
This week, the earnings of the four biggest US stocks are coming in rapid succession. A-share and Hong Kong semiconductor shares are moving in sync as well—Changxin has already seen a run-up.

My view:

The direction is right—domestic DRAM replacement is the long-term logic.
But when chasing gains in the short term, be careful. It’s an earnings-week environment with big sentiment swings, and it’s completely normal to pull back after good news is priced in.

If you want to take a quick stab at short-term directional betting, the UP & DOWN on the Binance wallet has already been placed on Changxin Technology’s market. It expires tomorrow (July 30). Choose bullish or bearish yourself—the entry barrier isn’t high; treat it like a sentiment vote.

👉 [https://web3.binance.com/en/prediction/detail/cxmt-up-or-down-on-july-30th-2026]

Not investment advice; data is subject to official sources.

#长鑫科技 #币安钱包

Apple’s market value surges past $5 trillion, overtaking Nvidia

Financial media all align on one account: “Smart money” is returning to value—finally, the market understands the question of burning cash vs. not burning cash.

I don’t buy this.

Money flows to Apple not because Apple’s AI strategy is that brilliant.

Money flows to Apple not because Apple’s AI strategy is that brilliant. It’s because Nvidia, Microsoft, and Meta have been spending so aggressively—so recklessly—that even your parents, scrolling through the news, would ask: “Are these companies out of their minds?” When money panics, it runs to places that look stable. Apple just happens to be standing there.

But Apple’s so-called “cost-saving via collaboration” model—

Put simply, it means they don’t build their own power plants and keep selling air conditioners.

For iPhones to run AI features, they still rely on someone else’s computing power.

This isn’t dimensionality reduction—it’s a temporary move to avoid the spotlight.

My actual holdings:

Nvidia long-term position unchanged
No chasing Apple

The logic is simple:
Nvidia’s rise to the top is driven by real, tangible demand.
Apple’s outperformance is driven by other people’s fear.
Fear-driven rotation arrives fast—and exits just as quickly.

Once Microsoft, Meta, and Amazon’s earnings reports come out this week and capex numbers get spread out again, the money could turn around at any time.

Don’t let labels trap you.
Focus on why money moves from A to B—that’s the most honest part of the market.

Not investment advice; data should be based on official sources.
#美股 #Aİ

At the San Francisco AI Summit, South Korea and the U.S. signed a $950 billion chip deal
Five-year term

SK hynix supplies HBM (high-bandwidth memory) to NVIDIA, and Samsung handles foundry and packaging for Broadcom

It’s not an investment—it’s procurement commitment
NVIDIA has essentially welded the storage supply chain to South Korea

Feel the scale with one number: $950 billion—more than Apple’s annual revenue

My understanding is straightforward:

The AI arms race has changed
Back then, it was about who had the stronger model—now it’s about who can secure the supply chain
HBM has already been in short supply; this order pushes the shortage out another five years

Who the core supplier of HBM is—you don’t need me to say it, right

Data should be based on official disclosures, and does not constitute investment advice.#HBM

The Dark Side of the Moon has open-sourced Kimi K3’s weights—28 trillion parameters

I haven’t run it locally yet
Honestly, even for individuals, this scale is hard to run

But the arena data is right there: for programming and long-task directions, it ranks first—it’s not just talk

Its overall performance still has ground to cover compared with Fable 5 and GPT-5.6 Sol, and the official side has admitted it too—they didn’t hard-sell it

What I’m most interested in is something else:

In China, for the first time, someone has directly released weights at this level

The community can quantify, fine-tune, and do private deployment

For those companies that don’t want to feed their data into closed-source APIs, there’s now another path. OpenAI’s pricing is under pressure

Of course, the deployment threshold of 28 trillion parameters is there—ordinary people in the short term should still stick to adjusting via APIs. Once the quantized version comes out and the community’s first batch of real-world tests have finished, we can talk again then.

Jingcai Xuchuang (03308) HKEX IPO: applications close at noon on Sunday, July 27, and the listing date is Wednesday, July 30.

Maximum offer price: HK$1,010. One lot is 50 shares; the application fee is about HK$51,000.
Maximum fundraising: HK$55 billion, the largest IPO in Hong Kong shares this year.

Cornerstone investors: Temasek, BlackRock, Hillhouse, Alibaba, Tencent.
A collective vote for the AI computing power supply chain.

Those who want to participate should already be moving—deadline is Sunday noon, and time is tight.
Data is based on the official prospectus only and does not constitute investment advice.

#港股打新

Google and Tesla released their Q2 earnings on the same night—I didn’t even look at revenue
I went straight to capex and free cash flow

Both are negative

→ → →

Over at Google: Revenue was $119.8B, up 24% year over year and above expectations

Google Cloud is even stronger—$24.8B, up a whopping 82% YoY
Looking purely at growth, nothing to complain about

But capital expenditures were $44.9B

That’s double year over year

Full-year guidance raised to $195–205B (from $180–190B)
Free cash flow? Negative $5.9B
Same period last year was positive $25.0B

One quarter: it went from +$25.0B to -$5.9B
Just that one number—down 3% after hours. I think the market is still being fairly restrained

Pichai said Gemini has 950 million monthly active users, and AI investment is “redefining every part of the business.”
Okay, I believe the demand is real
But the fact is the money is being spent faster than it’s being earned.

→ → →

Now Tesla’s is more interesting

Revenue was $28.24B, up 26% YoY and well above Wall Street’s $25.7B estimate. Autos +23%, energy storage +13%, services up 50%
The top line looks great

Then flip to profits: adjusted EPS was $0.33. The expectation was $0.51

That’s almost 40% off

Gross margin slid to 16.8%
Free cash flow also turned negative, -$1.1B. Capital expenditures were $5.79B, up 142% YoY, and full-year is expected to exceed $25B
Autonomous driving, Optimus, Cybercab, the AI chip factory—everything’s burning cash

Musk’s exact quote: “We will spend as quickly as possible, but not wastefully.” Investors clearly aren’t buying it—down 4% after hours, and premarket expands to 5–6%

→ → →

Put the two earnings reports side by side and the story is the same:
Growth is real, but so is the cash burn—and it’s burning faster than it’s earning.

The AI narrative isn’t dead. But the market’s attitude has changed
From “buy it if it touches AI”
To “how much did you spend, and how much did you get back?”

Next, Meta, Microsoft, and Amazon’s earnings—the capex line item will be under a microscope
Who can prove the money they spent eventually comes back can stay steady
If they can’t prove it, it’s the next after-hours -5%

Data should be based on official earnings reports; everything above is just my personal observations.#AI #美股

About six months ago building DeFi on Solana—what stablecoins can you choose?
USDC, or USDT
That's it

The data from 7/20 is out: Solana’s total stablecoin market cap has surpassed $15 billion

But what really made me stop isn’t that total
It’s the portion that’s not USDC/USDT—$4.81 billion, a new all-time high

The paradigm has cracked

—

Who’s eating this incremental growth? USD1

On Solana, the supply is $1.021 billion, accounting for 21% of its global total supply

Together with USDG, the two of them make up nearly half of the non USDC/USDT segment

There’s on-chain presence at the billion-dollar level—not numbers generated by airdrops. It’s users choosing with their feet
The first hard data showing the multi-chain expansion is working

All data is based on real-time on-chain data; the above is purely my personal observation.
#USD1 #WLFI

Two pieces of news hit at the same time today—both are related to AI infrastructure.

First, Nebius (NBIS)

After the close, the stock jumped nearly 19%.
The reason: Nvidia disclosed via a 13G filing that it holds about 9.3% of Nebius’s shares—22.25 million A-class ordinary shares. Included in this are warrants tied to the $2 billion strategic investment made earlier this March.

Nvidia didn’t invest casually.
It invested in Nebius—that is, AI cloud infrastructure.
This is the kind of thing that helps big enterprises build large-scale computing platforms.
This isn’t just a financial investment; it’s an industrial-chain linkage.

With NVDA and AMD moving higher too, market sentiment is very clear:
On the AI capex front, money is still flowing in.

//

Next, Supermicro (SMCI)

It also surged after hours.
The initial earnings report showed Q4 revenue near the low end of guidance—not exactly surprising. But two numbers are very eye-catching:

Gross margin jumped to 15–17%; previously guidance was only 8.2–8.4%.
That’s more than double.

The reason is that the customer and product mix are improving.
The share of higher-margin AI server orders is rising.

New orders exceeded $60 billion—an all-time record.
The backlog is equivalent to about 5 times the expected revenue for the current quarter.

$60 billion—what does that mean?
Even if it doesn’t sign a single new order, the existing backlog is enough to keep the business running for more than a year.

//

Taken together, these two developments point to the same conclusion:

AI computing demand isn’t cooling down—it’s accelerating.

Nvidia directly taking a stake in a cloud infrastructure company suggests it believes downstream demand is big enough and long-lasting enough to justify binding partners through equity.

Supermicro’s order surge indicates that actual AI server purchasing is materializing—not just talk.

I’ve been saying that the main AI investment theme is shifting from chips toward infrastructure.

Power, data centers, servers.

Today’s two pieces of news add another layer of confirmation.

In the short term, these stocks may still be quite volatile—don’t chase the price.
But looking at the medium to long term, I think the direction is becoming clearer and clearer.

The above is purely my personal observation. Data should be based on official sources. It does not constitute investment advice. #英伟达 #美股

Spent a bit of time digging into $TBB The Bitcoin Bull

Dev @GoingParabolic has publicly revealed their identity, not anonymous, and has been interacting with the community

Fair launch; the narrative is simply: a Bitcoin bull market

Up 34% in 24h, market cap 17M, 1,700+ holders

Currently observing

There aren’t many memes where all three—narrative, execution, and community vibe—are present at the same time, but until the concentration issue is resolved, I won’t go heavy

CA: 42cXQvAAr7hcPBPWAS4ocVtDyeJ4Fa6gRR2uG4gppump

DYOR, use only spare funds—meme trades involve risk

After Netflix’s earnings report came out, the stock price dropped to a 52-week low.

Revenue was $12.56B, up 13% year over year, but it missed analysts’ expectations by a hair.
The Q3 guidance also left the market unhappy.
That’s why it fell.

This gave me a reminder: the market currently doesn’t have patience for high-growth tech stocks.
Just a little miss is enough.

────────────────

The most important thing this week is Wednesday.

Google and Tesla both release earnings that day.
These two results will directly shape the near-term direction of the entire Tech AI sector.

My current view is that, after the market opens, the three directions with the biggest volatility are:

▫️ Tech AI: especially earnings-related stocks—if they come out strong, they’ll rally; if they’re just a bit off, they’ll get sold off.

▫️ Oil & gas: geopolitics is still unfolding, and oil prices will move with it.

▫️ Defensive and value stocks: if risk sentiment keeps deteriorating, money may rotate into this area.

For now, I’m lightly positioned and waiting to see the data come out on Wednesday. This isn’t the time to race ahead.

For reference only and not investment advice. DYOR #美股 #财报季

Saw it from a post
Binance will list the SPCXUSD1 Perpetual Futures contract tomorrow (July 20)

Trading opens at 17:00 (Beijing time). Maximum leverage is 25x. Minimum order is 0.01 SPCX.

The key point is: the settlement asset is USD1, not USDT.

My first reaction:
This isn’t just a matter of adding a contract.

//

First, let’s talk about the underlying asset itself—SpaceX.

IPO on June 12. Opening price was $135. It closed at $160 on day one.

Then it kept falling. It has already broken below its issue price—down nearly 45% from its peak.
There will be another lock-up release in early August, with 123 billion dollars’ worth of shares set to be released.

So at this timing, SpaceX contracts won’t lack volatility.

With 25x leverage, you can make money fast—and lose money fast too. Don’t get carried away.

————

But what I care about more is the USD1 layer.

I’ve been tracking USD1.

Previously, it mainly relied on Binance and Bybit’s holding rewards to attract volume, and there were also on-chain incentives.

But honestly, rewards can bring people in, yet they can’t keep them.

This time is different.

It’s directly used as the futures settlement asset for the period.

If you want to open a SpaceX contract, your margin has to be USD1.
Funding is settled once every 8 hours, and the funding is also in USD1.
Not “you can choose to use it”—you can only use it.

The rewards campaign is basically an invitation for you to come sit at the table.
This is a completely different situation—here you must hold it to be allowed at the table.

//

Think one step further:
Anyone who wants to trade this contract has to go swap in USD1 first.

That’s short-term buy-side demand.
Funding is settled every 8 hours, and USD1 will keep circulating inside the system.

That’s ongoing liquidity.

And because SpaceX is itself such a hot topic, the exposure is far stronger than those previous on-chain reward campaigns.

Of course, risks are there too.
SpaceX is currently trading below its issue price, volatility is high, and there won’t be few people who get wrecked with 25x leverage.
USD1’s connection to the Trump family is also being watched—regulators have had eyes on it for a long time.
Even previously, a Democratic senator requested a hearing.

So my view is:
In the short term, it’s driven more by sentiment and flow; in the long term, we need to see whether USD1 can keep maintaining settlement use cases like this.

//

After it goes live tomorrow, pay attention to these details:

The actual trading volume of SPCXUSD1, and the changes in USD1 holdings on Binance.
The former tells you whether the contract itself can really run.
The latter tells you whether new money is coming in because of this contract.

For anyone who wants to participate: manage leverage well.
At this point, SpaceX’s direction is unclear—don’t bet with a big position.

Not investment advice. DORY #USD1