Jealousy尼卡

Where is AI at now? Is there a bubble already?

My thoughts:

Investment is essentially about judging the cycle.

In the past, the internet, real estate, and new energy all went through the process from budding to explosion, then to capacity expansion, and finally to adjustment. AI will most likely follow the same pattern.

But I think it’s still too early to conclude that the “AI bubble has topped.” At least from the perspective of the industrial chain, we are still in the “shovels-for-sale” stage: foundational infrastructure such as GPUs, HBM, optical modules, PCBs, CCL, and data centers continues to expand. Meanwhile, truly “super applications” haven’t been realized at scale yet.

No one knows how far AI can go in the future, but at least for now, the main line of industrial growth hasn’t ended.

Of course, growth is never a straight line. Pullbacks, doubts, and valuation bubbles are all normal in the middle of the journey.

I’d rather think of today’s AI as a long-cycle trend, not a market that’s already nearing its end.
$NVDAB $SNDK #Aİ

SK Hynix $SKHY has seen a number of negative catalysts recently, and the stock price has been under pressure too. Let me share a few observations~

The news flow around SK Hynix lately really doesn’t look great.

Rumor has it that Nvidia may reduce HBM for Rubin Ultra, Hynix is carrying out a major capacity expansion, and the company’s earnings have also come in below expectations. Even in Korea’s investment community, stories have started circulating about “Hynix giving Nvidia a 50% discount in exchange for GPUs.” For a while, it almost looked like a meeting of bad news.

But if you break these items apart, I think many things are not as bleak as they seem. In fact, I believe a few points might be interpreted the opposite way by the market right now.

I still hold the same view: what’s truly worth worrying about isn’t short-term moves in the stock price, but whether the demand logic for AI memory has changed.

So far, I haven’t seen that logic being broken.

Of course, I don’t know whether the stock price will keep falling. But if it’s only because the market suddenly starts worrying that “HBM can’t be sold,” then I think that concern is still a bit premature—at least for now.

$SKHY #海力士

$AAOIB What’s most lacking right now may not be orders, but capacity.

Demand for 800G and 1.6T will, for quite a long time, exceed the company’s production capacity.

Even orders are already booked out to 2027 Q2.

What’s even more interesting is that each month the key customer’s demand for AAOI is about 20%–40% higher than the production capacity the company can currently provide.

And it’s not that the customer just occasionally asks when they can get the goods—it’s already at the level of weekly follow-ups to expedite shipments.

For a manufacturing business, this information is actually more valuable than, “Our pipeline is strong.”

Because you can boast about a pipeline, but customers usually don’t push for faster delivery without reason.

That also explains why the company still maintains its full-year 2026 revenue guidance of about $1.1 billion.
The issue isn’t whether the market can sell.

It’s whether the factory can produce it. So AAOI’s real growth formula going forward has gradually shifted from: demand × market share

to: capacity × yield × delivery speed × customer qualification.

That’s also why, in the end, almost all the questions in the call circled back to capacity expansion.

$AAOIB If capacity can be solved, then it really is something worth looking forward to.

#光模块订单排到2028年 #AAOI

#BTC走势分析

The number of whale wallets holding at least 10,000 BTC has rebounded to a 6-month high, reaching 90. Over the past 8 weeks, 6 net new such wallets were added, an increase of 7.1%. At the same time, small-wallet holdings in August fell~

Chip concentration is getting higher again—will this move be upward or downward this time?

$BTC

Why is the storage industry rallying so uniquely that SanDisk $SNDKB is rising?

I think there are several reasons:
1. The company’s latest quarterly report is still phenomenal. Revenue reached $8.97 billion, a new all-time high. The data center business, driven by AI demand, grew 103% quarter-over-quarter to $2.98 billion.
2. Real buybacks in black and white. SanDisk repurchased about $4.5 billion worth of stock in the previous quarter and still has $14.5 billion remaining in its buyback authorization.
3. The industry trend is clear. The NAND market is expected to grow from $300 billion in 2026 to $500 billion in 2027. The company has even already secured visibility into demand for more than four years.

In the end, the prior pullback in SanDisk’s share price ≠ weakening fundamentals—if anything, it might be the opposite~

An interesting piece of data:
Last week, MicroStrategy reduced 1,691 units, $BTC ;
Bitmine increased 7,391 units, $ETH , with staked ETH surpassing 5 million.
One sells BTC, while the other buys ETH.

Behind this isn’t just a simple shift between long and short positions, but different institutional understandings of how to position crypto assets for the future:
BTC is more like gold;
ETH is more like financial internet infrastructure.

In the next bull cycle, which will institutions choose: a “store-of-value asset,” or a “yield asset”?
What do you think?

$BTC

Ruisu lowers the Strategy stock target price from $213 to $165

All crypto-related assets are being lowered as well, including:

$STRC $CRCLB $COINB

From $2.4 Billion in Assets to a 30-Story Fall: Harry Yeh, the Crypto OG’s Last Day

Another heartbreaking piece of news has surfaced in the crypto world.

This time, it involves a veteran Chinese crypto OG—Harry Yeh.

Many people may not recognize his name, but in the early DeFi scene, he was a major figure.

Back in 2013, when the BTC price was still around $60, Harry Yeh had already entered the crypto market. Later, he founded Quantum Fintech Group. Starting with an initial scale of $2.5 million, it grew to the point where, at one time, it managed assets exceeding $2.4 billion.

In the early hours of August 7, a tragedy occurred at the Jade Park apartment complex in Asunción, the capital of Paraguay.

Around 4:30 a.m., police received the report. Harry Yeh was found having fallen from the 30th floor.

Some details at the scene also left the incident shrouded in uncertainty:

When he was found, he was completely naked, with a black plastic bag covering his body;

The door to his residence on the 30th floor was left open, and items inside were in disarray;

The police also investigated and collected evidence in another apartment he had on the 27th floor in the same building;

A 29-year-old Brazilian girlfriend told police she had no knowledge of the incident.

At present, the prosecution is still investigating three possibilities: accident, suicide, and homicide.

A crypto investor who once managed billions of dollars ultimately left the world in such a shocking way, far from home. $BTC

Recently, I’ve seen several big players selling Predict’s points—are they trying to recoup losses, or do they know something behind the scenes?

But from what I can see, the prediction market is still fairly optimistic about PRE’s FDV after TGE, mainly because it has the backing of Binance’s top-tier resources.

For @Polymarket, the pool of 100M–500M FDV for PRE, one day after its release, is all above 50%; @42space market speculation suggests the most concentrated range is between 300M and 400M.

Honestly, in this bear market, if Predict’s token issuance can reach an FDV of over 300M, that’s already quite strong.

Finally, let me say something controversial: in this market, 95% of people are here for the airdrop, not because they genuinely like prediction.

Tomorrow, Yushu Technology will have its IPO subscription—could it become one of the most outrageous wealth-effect stories of this year?

Yushu Technology’s final issue price has been set at 150.80 yuan per share.
For a Science and Innovation Board lot (500 shares), the payment required is about 75,400 yuan.

On the other side, on Hyperliquid, the pre-market price for Yushu Technology has already surged to 85.4U.
Using the previously applied exchange-rate convention:
85.4U ≈ 576.45 yuan per share

That means:
Theoretical price spread per share: about 424.98 yuan
Theoretical value of one lot: about 288,200 yuan
IPO subscription cost: about 75,400 yuan
Theoretical unrealized gain per lot: about 212,800 yuan

In other words, if the pre-market price can be realized into the trading price after listing on A-shares, the profit from one lot could be close to 3x the subscription cost.

Now looking at market cap:
Yushu’s offering market cap: about 61 billion yuan
Hyperliquid’s 85.4U implies a market cap of about 232.2 billion yuan

The pre-market market has already priced in a premium of about 3.81x the offering market cap.

Of course, the pre-market price isn’t the listing price—between the two, liquidity, sentiment, and capital games will all play a role.

But the biggest significance of this case is that:
The imagination around the robotics sector is now being re-priced by the capital markets.
In the past few years, the market chased AI computing power; in the next few years, embodied intelligence could become the new super-narrative.

Yushu is just the first batch of companies stepping into the spotlight.

Damn, Kaito's new regulations came out for a day and $KAITO dropped 20%

The official says now it's harder to cut the grass

Even those who drew up the collateralized pledge's “pie” don't believe it anymore

What does it mean?

Is $BTC a good teammate?
Or is $BTC just a backup?

Golden Retriever, don’t mess with people in the crypto circle like they’re idiots, okay?

This melon seems like nobody’s eating

UpShift’s apxUSD has been accrued, and the community is in an uproar. Honestly, this is all because previously $STRC collapsed, causing apxUSD’s price to crash and liquidity to dry up.

One comment made me laugh: “Every time I see your messages, it’s a pullback here and a pullback there—no good news almost ever. You guys can either do it properly, or if you can’t, just close the door.”

That said, Saturn is still impressive in comparison—the product design is a success:
1. The split design of USDat and sUSDat is reasonable.
2. sUSDat’s redemption mechanism is strong.
3. High reserve transparency.
4. The overall product logic is simple and easy to understand.

In the end, I remembered something Guo Degang once said: “This industry—everything depends on peers propping you up.” You can only truly appreciate Saturn’s excellence through comparison~

$BTC

Oh no, the Digital Asset Market Clarity Act CLARITY still hasn’t passed!

Now it’s been pushed to mid-to-late September, but everyone doesn’t need to worry—passing it is just a matter of time.

Actually, many people don’t know that CLARITY has already covered two-fifths of the way through:

1.✅ The House has passed
2.✅ The Senate Banking Committee has passed (15-9), and moved to the full Senate for a vote
3.❌ The Senate passes the final vote
4. The President signs

As long as the Senate can pass it, Trump will definitely sign it. So the bill is basically just down to the last little push.

The odds of it passing this time were never that high. Just look at BTC—its price hasn’t really been affected at all.

Wait until September and see if we can get a boost to $BTC . If it can likely pass in September, then BTC should rise.

Be patient—hurrying won’t make the hot tofu cook faster. Keep doing your DCA (dollar-cost averaging).

Good news for $KAITO ?

$KAITO has announced a new round of upgrades to its staking mechanism, and the efficiency of reward distribution will be further improved:

1. The reward pool is expected to grow by 2x
By introducing a “voluntary participation mechanism,” unclaimed airdrop allocations will be redistributed to active stakers before claim, reducing wasted rewards.

2. The number of participating projects is expected to increase by 2x
Kaito will cancel token event (TGE) pre-token activity fees, encouraging more projects to join the staker airdrop system, so that ecosystem returns flow back to the community even more.

3. Long-term stakers gain higher weight
Under the new rules, participants must stake for the entire duration of the event. Temporary buyers who stake just for the airdrop will see reduced advantages, while true long-term holders receive more rewards.

Currently, most airdrops still require a minimum of 5000 $KAITO (basically excluding small holders). Yapybara NFT holders receive an additional 10% distribution bonus. Small holders can go to Pendle’s YT pool.

In the end, it feels like a positive development—at least Kaito is back, but not that big ~

Gold will definitely rise in the long run. $XAUT is no exception when it returns to around 4300.

Just look at the size of the U.S. Treasury bond market in the chart—it has already broken through 40 trillion.

The interest that needs to be paid every year is more than one trillion, which is even higher than the U.S. defense budget.

This money is basically impossible to repay through economic growth; the only way is to follow the old path of debt monetization.

So no matter how hawkish the Fed gets, in the end it will be forced to restart some form of bond buying.

A weakening U.S. dollar is basically obvious (in the long run).

One more point: regarding tonight’s Non-Farm Payrolls report, the market may already have priced in expectations, which could be why gold has risen.

Of course, in the long run this is also beneficial for $BTC ~~~

Let me drop a big hot take:

The coin most worth DCAing during this bear market is $HYPE .

I told you already: when $SPCXB rebounds, go short. With such a concrete negative catalyst like this unlock event, there’s basically nothing to be afraid of.

Even if the company’s Q2 earnings report looks great, it doesn’t solve the sell-pressure facing the market. For stocks with unlock events, in most cases they rally before the unlock (high-price dumping), and after the unlock they face pressure and need time to digest it. Unless there’s strong capital that firmly pushes upward, it’s rare to see sustained gains after a large-scale unlock.

Also, I noticed that many people don’t really grasp what this time’s 911.5 million shares being unlocked means.

Let’s calculate it simply: After $SPCXB was listed, it has been a typical low-float stock. Although the overall market cap is large, the shares actually available for trading in the secondary market are very few. The initial float ratio was only 4.9%. That’s why the price could be pushed up quickly earlier—the float was so small that with only a little capital pushing, the price had huge elasticity. But this time is different. This unlock involves 911.5 million shares, which is about 6.9% of the total outstanding shares—actually even more than the shares that were initially available to trade.

In other words:
Unlock before:
Float ratio about 4.9%
After unlock:
Float ratio jumps directly to about 11.8%
The tradable float more than doubles.

What does that mean? It’s like the market suddenly gets a batch of newly tradable shares, and the earlier advantage of scarce supply is clearly weakened. Of course, after the earnings report was released, $SPCXB did see a rebound—mainly because the AI business grew significantly.

But the problem is: good earnings ≠ no selling pressure. So you still have to look at the essence.

Black humor: the 4-hour candlestick chart drew a big rocket.