Crypto KOL Community & Market Trends

Big day

Bitcoin’s short-term holder supply is shrinking toward an extreme.
That means fewer coins are sitting with the most reactive holders, while more supply is moving into stronger hands.
Historically, similar STH contractions have appeared during periods of low interest before demand returned.
The key signal now is simple: supply is tightening, but demand hasn’t fully come back yet.
If that changes, $BTC could have a much stronger setup ahead.

Tom Lee: 2027–2028 could bring the biggest gains we see in our lifetimes
But there will still be another wave of pressure before the midterm elections.
$Anthropic is expected to go public in October and currently has a secondary-market valuation of $1.1 trillion
With an IPO valuation of $2 trillion, it could surpass $SPCX to become the largest IPO in history. Everyone is pulling funds out of $NVDA , cloud, storage, optical modules, and other sectors to buy pure-play AI. $OpenAI may follow later, bringing another two waves of major liquidity drains.
Then, before the end of the year, $SPCX will unlock another 25–30% of its shares.
Once all of this is over, 2027 will see a massive surge!

4 billion becomes 3 trillion, Harmony wants to roll back—do you support changing the ledger, or accepting the loss?

If your bank’s system were hacked, causing tens of trillions of banknotes to appear out of thin air, and the bank said, “We’ll change the ledger back and pretend nothing happened”—would you support it?

Most people might say: Of course—why should we let the hacker profit?

What if I told you this bank is called “blockchain”?

Would you still support it?

On August 12, the veteran public chain Harmony ran into trouble.

On-chain analyst Juiceberg was the first to discover that someone had exploited an “empty-block” vulnerability to mint approximately 4 billion ONE tokens out of thin air.

That was equivalent to 26% of the total supply.

Approximately 2.8 billion of them were quickly transferred to major exchanges.

**The price of ONE briefly plunged by more than 50%.**

Things were already explosive, right?

But what came next was even more explosive.

CertiK monitoring showed that the amount of abnormally minted ONE had exceeded 3 trillion, involving 6 abnormal blocks.

4 billion became 3 trillion.

26% became 2,000%.

You read that correctly.

Harmony’s total supply was originally approximately 15 billion tokens. Now 3 trillion have appeared out of thin air.

The entire tokenomics model collapsed overnight.

Harmony reacted quickly:

Emergency deployment of patch v2026.1.1 to prevent further minting

Suspension of the cross-chain bridge

Coordination with exchanges to freeze the funds in 4 wallet addresses

Most importantly: advancing an on-chain rollback plan

What does a rollback mean?

It means restoring the entire chain to its state before the attack occurred.

That amounts to wiping out all transactions after August 12—including normal transfers by innocent users, DeFi operations, and DEX trades.

Benefit: the hacker’s efforts come to nothing, and 3 trillion ONE disappear from the ledger.

Cost: blockchain’s “immutability” becomes a joke from that point on.

This reminds me of The DAO incident in 2016.

Hackers stole 3.6 million ETH from Ethereum, and Vitalik chose to roll back the chain through a hard fork, returning the funds to investors.

What happened next?

Ethereum split.

One side was the rolled-back Ethereum (ETH), while the other was Ethereum Classic (ETC), which insisted on immutability.

Ten years later, supporters of the two sides are still arguing.

And what Harmony faces today is even more extreme than The DAO—

The DAO was: “The money was stolen, so I’ll help you recover it.”

Harmony is: “The money was created out of thin air, so I’ll tear up the ledger and rewrite it.”

The former is recovering stolen assets; the latter is directly altering history.

So the question returns to the beginning—

Do you support a rollback?

Those who support a rollback would say:

3 trillion ONE appeared out of thin air, destroying the tokenomics

Without a rollback, holders’ assets would be diluted by 2,000%—is that fair?

The hacker gets incredibly rich—why should they?

The project team has a responsibility to protect users’ interests

Those who oppose a rollback would say:

The core value of blockchain is immutability

If it can be rolled back once today, it can be rolled back a second time tomorrow

Then I might as well keep my money in a bank; at least a bank needs approval to alter its ledger

This “decentralization” you keep talking about—is this it?

Both sides have a point.

But the most painful fact is that no matter what you choose, someone gets hurt.

With a rollback, innocent users’ normal transactions after August 12 would be wiped out. Without a rollback, the assets of all ONE holders would be diluted into worthless paper.

There are no winners.

What’s even more painful—

This is not the first time Harmony has run into trouble.

In 2022, Harmony’s cross-chain bridge was attacked by North Korean hackers, resulting in losses of approximately $100 million.

At the time, the community helped track the funds, but the project team did not pay a cent and only said, “Good job.”

So when trouble struck this time, well-known on-chain investigator ZachXBT directly refused to provide free assistance.

“Last time, I helped you track $100 million, and I didn’t even receive a thank-you payment. Forget it this time.”

Trust is something that disappears once it is overdrawn.

As of August 13, Harmony had reached agreement with validators and exchanges on the rollback path, and the patch had been activated.

The rollback will most likely be carried out.

But the real impact of this incident goes far beyond the price of ONE.

It is putting every crypto investor to the test:

Do you ultimately believe that “code is law,” or that “the project team has the final say”?

If the code can be rolled back and the law can be changed—

Then what is the difference between blockchain and traditional finance?

Do you support Harmony’s rollback?

If one day a project in which you have a large position encounters the same thing, how would you want it handled?#GateLaunchpool瓜分141万枚DOS #Gate多项交易指标全球Top4 #我的七夕交易分享 $BTC $ETH $ONE

The data is astonishing: the 451 S&P 500 companies that have already reported earnings saw profits rise by an average of 41.6%!
But if Micron and Google parent Alphabet are excluded, growth is cut in half to 21.5%. In other words, these two companies propped up earnings for the entire market.

Don’t just look at the 117% figure and stop there when reviewing Gate’s latest proof-of-reserves data.

Gate’s overall reserve ratio is 117%, covering nearly 500 types of user assets. User BTC holdings increased from 19,054 BTC in the previous period to 21,557 BTC, with excess reserves of 24.2%. User ETH holdings also increased from 344,935 ETH to 374,348 ETH, with excess reserves of 22.02%.

The stablecoin figures are also quite interesting. Total user holdings amounted to 1.336 billion stablecoins, while the platform held 1.59 billion in reserves, bringing the reserve ratio to 118.97%.

I personally keep a significant amount of funds on Gate. As the exchange continues expanding its features, from spot and futures trading to stocks, ETFs, prediction markets, and even the gradual integration of TradFi, the more funds and use cases involved, the more I pay attention to whether its reserves are sufficient to support the platform.

Of course, reserves exceeding 100% do not mean the platform is completely risk-free, but they do at least indicate that it has maintained a certain safety buffer. In particular, BTC, ETH, GT, and XRP all remain above 100%. I believe this kind of ongoing disclosure is more meaningful than simply claiming to be secure.

Now, all platforms are competing over who offers more products and higher yields. Personally, I think that the larger a platform becomes, the more essential it is for asset security and transparency to keep pace. Ensuring that users can keep their funds there safely is the most important thing. Expanding platform functionality on that foundation is what matters more to me!

Perks | GLM-5.2 at 40% off! An unbeatable value

I genuinely think this promotion is incredibly generous. GLM-5.2, a top-tier large language model, is now available to users at an exclusive limited-time price of 40% off.

Yesterday, one sentence from Musk completely overturned SpaceX’s valuation logic.

If one day, a rocket-building company tells you:

“Rockets are just a side business. AI is the main business.”

Would you believe it?

Whether you believe it or not, Musk said it.

And he said it more forcefully than anyone else.

On August 12, SpaceX released a 29-minute video of an internal all-hands meeting.

Standing in front of employees, Musk said three things that left investors around the world completely baffled—

First: “AI revenue will definitely exceed the combined revenue of all of SpaceX’s other businesses in September.”

Rockets, Starlink, Dragon spacecraft—all added together, beaten by AI in a single month.

Second: “By the end of next year, we will deploy 10 gigawatts of AI computing capacity, generating $300 billion to $500 billion in annual revenue.”

What does that mean? SpaceX’s total revenue last quarter was only $7.8 billion. $500 billion is 64 times its current revenue.

The third, and craziest, statement: “Five years from now, AI will account for 99% of SpaceX’s value.”

99%.

A rocket-building company deriving 99% of its value from AI.

You might think Musk is hyping things up again.

But look at the numbers—

In the second quarter of this year, SpaceX’s AI business generated $2.6 billion in revenue, up 213% quarter over quarter and 247% year over year.

Starlink generated $4.29 billion, while space launches generated $960 million.

In just one quarter, AI went from zero to 60% of Starlink’s revenue.

At this growth rate, surpassing all other businesses in September is not nonsense—it is math.

Even more astonishing is capital expenditure—in the second quarter, SpaceX’s total capital expenditure was $18.37 billion, of which $15.83 billion was poured into AI computing infrastructure, accounting for **86%** of total capital expenditure.

In the first quarter, AI business capital expenditure was $7.72B, already more than twice the combined capital expenditure of the space and communications segments.

In other words, this company is nominally building rockets, but in reality it is frantically building data centers.

What about the rockets? What about Starlink?

Musk’s explanation is that they have all become “infrastructure” for AI.

Starlink: nearly 11,000 satellites currently in orbit, with plans to expand to 100k, eventually carrying more than 90% of global internet traffic.

Starship: increasing annual payload capacity from 2,500 tons to more than 1 million tons.

What are these for?

To move AI inference into space.

Musk has proposed a concept called “ground training, space inference”—AI training stays on Earth, while AI inference moves into space.

He calls it “Starmind”: a low-Earth-orbit computing network consisting of up to 1 million AI satellites.

Rockets are responsible for sending satellites into space, Starlink handles data transmission, and AI generates the money.

A complete closed loop.

But here’s the question—

Will the market buy this story?

SpaceX went public in June this year at an offering price of $135, reaching a valuation of $1.77 trillion.

Then its first earnings report came out, and the stock fell 8% after hours, briefly dropping from a high of $225 to around $118, nearly halving.

What is the market afraid of?

The cash burn.

Capital expenditure of $18.37 billion in a single quarter. The AI business posted a loss of $2.47 billion in the first quarter. Although revenue surged in the second quarter, net losses still came to $541 million.

The story is sexy, but the bill is brutal.

Morgan Stanley said the market’s valuation of SpaceX’s AI business is “extremely conservative.”

But there is a reason for that conservatism—10 gigawatts of computing capacity and $500 billion in annual revenue are still just projections.

Large-scale computing infrastructure construction and capital expenditure burn real cash, down to every last cent.

So, if you are following SpaceX—

What data would you most like to see to prove that its AI strategy is being delivered?

I have listed three. Think about them yourself:

First: the gross margin of the AI business.

Revenue can be propped up by scale, but profit is what really matters. The AI business lost $2.47 billion in the first quarter—when will it turn profitable?

Second: the pace at which 10 gigawatts of computing capacity is brought online.

Musk says it will reach 10 gigawatts by the end of the year. But the conversion rate of capital expenditure and the data-center deployment rate each quarter are the hard metrics that will test execution.

Third: renewals and expansion by major customers.

Anthropic pays $1.25 billion per month. Can customers of this caliber be retained? Can prices be raised? Can new business be expanded?

These three data points are more concrete than any “astronomical” forecast.

One final thought—

Musk has forcibly transformed SpaceX from an aerospace company into an AI computing company.

Rockets have become supporting actors, Starlink has become the pipeline, and AI is the true protagonist.

Five years from now, if AI really accounts for 99% of SpaceX’s value—

Then everyone mocking him for hyping things today will be a fool.

But if the 10 gigawatts of computing capacity are not built and the $500 billion in revenue fails to materialize—

Then everyone applauding him today will be the bagholders.

The story is expensive, but the truth is even more valuable.#GateLaunchpool瓜分141万枚DOS #Gate多项交易指标全球Top4 #我的七夕交易分享 $BTC $SPCX $ETH

Either they’re extremely careless, or something fishy is going on... Three years ago, $24 million in funds was stolen through phishing, but the address was kept in use without being changed. As a result, another $26 million was stolen today. However, this incident today was not phishing through an approval from a single address; it appears the wallets were compromised wholesale: funds were transferred out of 3 wallets, including a clean wallet that had never granted any approvals (0x8f3...914).
◎$24 million phished 3 years ago

$LIT Showing Strong Bullish Momentum!

$LIT is regaining strength currently trading at $2.3960 (+2.49%) after recovering sharply from its local dip to $2.3134.

Bulls recently surged past $2.40 to touch $2.4091, showing solid buying interest on the lower timeframes.

Key 24H Metrics:
24h High: $2.4451
24h Low: $2.2125
24h Volume: 1.79M LIT ($4.17M USDT)

With price steadily consolidating near the top of its
recent move, a clean break above $2.41 could set up another test of the daily high at $2.445.

Are you holding $LIT for another leg up?#GateLaunchpool141MDOS #GateLaunchpool141MDOS #BigShortBurryBearsAI

Today’s U.S. AI Stock Information Gap:

1. The golden age of storage and optical modules is back! CPI data showed that inflation is easing, and $MU $SPCX $SMCI $NOK $LITE $SKHY all surged 6%–20%!

2. New Cloud skyrocketed! $CRWV surged 20% yesterday, with multiple institutions raising their price targets to 160, while $NBIS surged 35% today on strong earnings!

3. Is $NVDA changing direction? It no longer wants to just sell chips. According to insider sources, it is entering the race to build the world’s most powerful open-source AI model. It can’t beat Liang Sheng, so Huang Sheng has to step in.

4. However, $CSCO $CBRS $COHR also reported decent earnings, while their stock prices plunged 4%–14%! The market demands perfection: Cisco’s gross margin declined, Cerebras’ hardware sales fell and losses widened, and Coherent’s P/E ratio is too high.

5. DeepSeek-V4-Pro has been released! Its agent capabilities have improved significantly, rivaling Fable 5, while its price is only 2% of Fable 5’s.

The whale who used leverage to buy $30 million worth of ETH as it rebounded from its early-June bottom took profits and unwound the leverage today, successfully earning $4.3 million👍 1⃣ On 6/7, the whale borrowed 30 million USDS from Spark using ETH as collateral through recursive lending, then bought 18,212 ETH at an average price of $1,647. 2⃣ Today, the whale sold 15,993 ETH at $1,889 to repay the 30.2 million USDS loan, earning a profit on the leveraged position.

Up 22% in 10 days: Is this a valuation recovery or a “dead cat bounce”?

Did you sell at a loss on July 30?

If you did—congratulations, you successfully sold at the very bottom.

If you didn’t—congratulations, your paper losses have recovered by 22%.

But if you’re now hesitating over whether to chase the rally—

Stop for a moment.

On August 13, South Korea’s KOSPI Index surged as much as 4.8% intraday, rebounding more than 22% cumulatively from its July 30 low and officially entering a technical bull market.

10 days, 22%.

Samsung Electronics rose more than 5%, while SK hynix climbed more than 6%. SK hynix surged as much as 8% intraday.

How sharp was the rally? Even the temporary halt mechanism for programmatic buy orders was triggered.

Ten days ago, everyone was panic-selling.

Ten days later, everyone is asking the same question:

Can this rebound still be chased?

First, let’s look at the bulls’ logic—it is indeed compelling.

First, global cloud providers are pouring money into AI infrastructure at a furious pace.

The combined capital expenditure of the four major cloud providers could approach $750 billion in 2026. The combined capital expenditure of the nine major cloud providers is expected to grow 90% year over year and exceed $886.7 billion.

Amazon raised its estimate from $200 billion to $220 billion, while Google raised its figure from $180 billion to more than $195 billion.

All this money will ultimately become orders for Samsung and SK hynix.

Second, memory supply and demand continue to tighten, and HBM prices have gone crazy.

HBM3 prices have surged from $180–220 at their low in the second quarter of 2025 to $700–850 now.

SK hynix has already begun early mass shipments of HBM4, securing multi-year agreements with around 10 core customers. HBM supply is still expected to face a shortfall in 2027.

There are no signs of any slowdown in demand.

Third, Temasek has arrived.

Singapore’s sovereign wealth fund plans to make its first direct investment in South Korea’s stock market, targeting Samsung Electronics and SK hynix. Temasek has identified memory chips as the most undervalued segment of the AI value chain.

This is not retail money, nor is it hedge-fund capital—this is some of the world’s top long-term capital.

Once the news broke, Samsung and SK hynix both surged more than 8% intraday.

But the bears’ logic is—equally striking.

First, how much foreign capital has left South Korea this year?

Foreign investors’ cumulative net selling reached about $81 billion in the first half of the year. For the full year, it could reach $120 billion.

What does $120 billion mean?

The South Korean stock market has never experienced capital outflows on this scale.

Second, where is the KOSPI now?

Even after rebounding 22%, the KOSPI is still about 24% below its late-June peak.

It fell more than 40% from its high and then rebounded 22%—do you think this is a reversal or an oversold rebound?

Third, the geopolitical risks facing the memory industry.

China’s production capacity is expanding. U.S. controls over the semiconductor supply chain are tightening. South Korea’s memory-chip duo is caught in the middle, in a highly delicate position.

Fourth, and most importantly—

July’s plunge was a short-term bout of pain caused by deleveraging.

That wave of selling was not caused by a deterioration in fundamentals; leverage simply blew up.

And rebounds driven by forced deleveraging are often fast and fierce—but they are not necessarily sustainable.

My view is simple:

A rebound ≠ a reversal.

Temasek represents a long-term perspective spanning 5–10 years.

July’s plunge, meanwhile, was a short-term shock triggered by the unwinding of leveraged trades.

Both things can be true at the same time—

The long-term outlook can be positive while short-term volatility remains unavoidable.

Here’s the painful truth:

The market rose 22% in 10 days, but how many of you actually made money?

Those who sold at the bottom, those who were afraid to chase the rally, and those who chased it but feared a drop—

The biggest irony of this rebound is that the index came back, but your money didn’t.

Finally—

The bulls say: $750 billion in cloud-provider capital expenditure, undersupplied HBM, and Temasek buying the dip—this is a fundamental reversal.

The bears say: $120 billion in foreign capital has left, valuations are still 24% below their peak, and geopolitical risks remain unresolved—it’s merely a dead cat bounce.

Who is right and who is wrong?

I don’t know.

But there is one thing I am certain of:

At this level, the risk of chasing the rally and the pain of missing out are equally real.

Do you think the KOSPI can return to its June high? #GateLaunchpool瓜分141万枚DOS #7月CPI符合预期通胀继续降温 #我的七夕交易分享 $SAMSUNG $SKHY $BTC

Update: This whale, which participated in Ethereum’s ICO in 2015 at $0.311 per ETH, transferred all 2,000 ETH ($3.77 million) to Coinbase four hours ago. Over 11 years, $622 turned into $3.77 million—a 6,060x return.

Yesterday $LITE , today $COHR . While Maitong’s editor was off work, I went ahead and posted the update. $COHR ’s earnings report was equally explosive:

1. First, the earnings report BEAT expectations, with revenue exceeding forecasts (20.46/19.81), strong growth in data centers and communications, a sharp increase in gross margin (Non-GAAP gross margin: 40.2%), and net income exceeding expectations (2.41/2.19);

CPI at 3.4% was fully in line with expectations—why did BTC retreat after rising just 0.3%?

Did your heart start racing as you watched the CPI data last night?

U.S. July CPI came in at 3.4% year-on-year, while core CPI came in at 2.5%—both hitting expectations precisely.

Inflation is cooling. The probability of a September rate hike is falling. The negative catalyst is gone.

You thought: BTC is about to take off, right?

But what happened?

BTC rebounded from $63,200 to $64,400, gaining 1.9%—then turned lower and plunged back to around $63,500.

The full-day gain? 0.3%.

The Nasdaq rose 0.54%, while gold reversed in a V-shape to gain more than 1%.

Bitcoin acted like nothing had happened, lying flat in place.

Confused?

Where did things go wrong?

“Meeting expectations” is itself the biggest problem.

The market never pays for “what was expected.”

A month ago, the probability of a September rate hike was 30%. Everyone was anxious, panicked, and unable to sleep.

After yesterday’s CPI release, CME data showed that the probability of rates remaining unchanged in September rose to 59.9%.

The probability of a rate hike fell from 50% to 40%, while rate-cut expectations went from nonexistent to tangible—it sounds great, right?

But the problem is that the market had already priced in “cooling inflation” before the CPI was released.

Last week, spot Bitcoin ETFs recorded net inflows for five consecutive trading days, totaling approximately $854 million—the strongest momentum since May.

Smart money had already entered.

When the data actually came out, everyone realized—“Oh, it’s just as expected”—and then? Nothing.

The buying disappeared. Because those who were going to buy had already finished buying last week.

The disappearance of a negative catalyst does not mean a positive catalyst has appeared.

The market does not want “no rate hike.” It wants “certainty of rate cuts.”

CPI did not provide that. So BTC did not move either.

Going one level deeper: CPI at 3.4% is still 1.4 percentage points above the Federal Reserve’s 2% target.

Inflation has cooled, but it is nowhere near “victory.”

Oil prices are still hovering around $100. Housing costs rose 0.1% in July, accounting for two-thirds of the overall increase.

The Federal Reserve has no reason to cut rates. Not one.

So look at the CME data: the probability of rates remaining unchanged in September is 59.9%, while the probability of a rate hike is 40.1%.

No rate hike, but no rate cut either. This is what is known as “higher for longer.”

For BTC, this is the most frustrating state.

Rate cuts are the engine of a bull market. Not raising rates merely means “not dead,” not “alive.”

Speaking of this, I want to discuss something else that is happening—the mining companies are collectively fleeing Bitcoin.

You may not have noticed: Core Scientific’s AI data center hosting revenue surged from $8.6 million a year ago to $77.5 million, an increase of more than ninefold, making it the company’s largest business line and displacing Bitcoin mining.

MARA, the world’s largest publicly listed mining company, announced that it would transform from a “Bitcoin mining company” into an “energy infrastructure platform.” Riot Platforms just signed a $9.1 billion, 20-year AI data center partnership agreement with Anthropic.

The miners are voting with their feet.

After the halving, the block reward fell to 3.125 BTC. The price fell from a high of $126k to $64k, roughly halving.

Mining is no longer profitable. Renting out AI computing power, by contrast, has become a money-printing machine.

Even the most loyal Bitcoin miners are transforming—what do you think this means?

Back to the market.

BTC is currently stuck between $63,000 and $64,400, unable to move either up or down. The 50-day moving average is below the 200-day moving average, making the technical picture bearish.

Funding rates remain between +0.004% and +0.007%. The cost of holding long positions is low, with no overcrowding—but there is no enthusiasm either.

The market is waiting. Waiting for a real catalyst.

It could be tonight’s PPI. It could be the FOMC meeting at the end of the month. It could be a geopolitical black swan event.

But it definitely is not a CPI reading that “met expectations.”

Finally, three hard truths—

First, buy the rumor, sell the news—this game has been played for a century and has never changed.

Prices rise before the CPI data comes out and fall afterward. The script is clichéd, but people still fall for it every time.

Second, the market has already priced in “cooling inflation.” Now it wants “certainty of rate cuts.”

CPI did not provide that, so BTC did not move either.

Third, when even miners are transforming toward AI—are you still expecting BTC to take off on macroeconomic data?

Wake up.

This market needs something stronger than “meeting expectations.”

Do you think tonight’s PPI can give the market what it wants?#GateLaunchpool瓜分141万枚DOS #7月CPI符合预期 #我的七夕交易分享 $BTC $ETH $XAU

Once again, I became my own antidote
Look within, and life will flourish endlessly

Opportunity cost is precious
Some stocks are absolutely off-limits for me. Even if they go up, the money wouldn’t be easy for me to make—I’d need exceptionally good luck or have to wait an exceptionally long time. So why not put my time, energy, and attention into stocks that offer me better opportunities?
Stocks that don’t match my style and stocks that aren’t in the spotlight aren’t bad; we’re just not a good fit.

NBIS's surge after today's earnings report reminded me of a devilish detail. Previously, Big Short Burry chose to short the stock directly rather than buy put options because the implied volatility of the $NBIS put options exceeded 100%
So this time, by shorting directly, unlike options, losses cannot be effectively capped and are theoretically unlimited. NBIS has rebounded so much from the bottom, yet its market cap is only $70 billion—not exactly a large-cap stock
And is his counterparty at this very moment Citadel, which just happened to take down the new AI stock god Leopold?

Squirrel Bro, the strongest Chinese contender in the AI $Swarms space, has released his latest research project, $Blink . His article is far better than most of the low-quality content out there. If you support Squirrel or Blink, feel free to give it a read.