海盗鸭

Another moment edging toward 5%: The U.S. 10-year Treasury yield is at a 20-month high, steadily moving closer to 5%. The last time it formally closed above 5% was in October 2023, when risk assets were under collective pressure. This current upswing is different: the driver isn’t rate hikes, but rather a surge in Treasury supply under fiscal deficits, rising inflation expectations, and the return of the term premium. At its core, the market is re-pricing national credit. Yields are the gravitational force behind pricing across all assets—each step higher adds another unit of pressure to the denominator of risk assets. The transmission to the crypto market is direct: risk-free returns are nearing 5%, the appeal of holding cash increases, and the opportunity cost of high-beta assets rises—assets like $BTC are hit first. Next, watch the subscription data from Treasury auctions and the Fed’s remarks. Any signals that keep interest rates high for longer will continue to weigh on valuations. In an environment that’s approaching 5%, will you de-lever and wait, or add on dips? #U.S. 10-year Treasury yield nearing 5%

NVIDIA’s CEO connects with Trump, publicly downplays AI risks—this news has a strong political flavor. But what ordinary people should extract from it are three actionable signals, not emotions.

First some context: Huang Renxun’s downplaying of “existential risks of AI” is not the first time. He previously said publicly that truly dangerous general AI is still decades away from reality. For now, the pressing issues are specific problems like bias and misuse. The AI policy backdrop of this White House is also one of looseness: in the early days of taking office, it revoked the prior administration’s AI regulatory executive order, with the emphasis on develop first, govern later. When the CEO calls the president to play down risks, it’s essentially another convergence of industrial capital and policy direction.

For ordinary users, I distill three actions. First, don’t use news like this as a reason to chase higher prices for $NVDAB —policy headwinds and tailwinds have long been priced in. What drives the stock price is the marginal change in data center capital expenditures; watching the capex guidance of hyperscale vendors is far more useful than focusing on the headline. Second, regulatory easing will likely accelerate the diffusion of open-source models. Toolchains keep getting cheaper—this is a concrete benefit that developers and everyday users can directly enjoy, something more tangible than gambling on a single stock. Third, if you hold tech or crypto positions, treat events like this as a source of volatility rather than a trend signal—repeated cycles of policy changes in domestic industry are the norm, so keep your position sizing with room for those repetitions.

My take: I’m not convinced by the “AI risk is still far off” judgment—it feels like an extension of a business position. But on the direction, I also admit that easing does accelerate rollout. These two things don’t conflict—you can be bullish on deployment while not blindly trusting the safety narrative.

Do you think AI should be regulated first for “runaway control,” or first for “misuse”?

#NVIDIA CEO connects with Trump downplays AI risks

Larry Ellison canceled a $7.5 billion Oracle stock sale plan. Insider activity at this level has long been one of the market’s most sensitive signals. Large executive sell-offs are often interpreted as “insiders believe the stock has peaked,” while an active pause is frequently read the other way as “the founder believes the stock is undervalued.” Given that Oracle’s share price has surged over the past two years on AI cloud infrastructure orders, and Ellison’s net worth has grown accordingly, stopping the cash-out at this point—whatever the true reason—would be an emotional positive for ORCLB bulls. In tech stocks, the votes a founder backs with real money are more persuasive than any roadshow. Still, one reminder: insider trading should be used only as reference, not as a basis. In history, there are plenty of examples where the stock price fell even after “insiders didn’t sell,” so don’t treat that as the only anchor. Would you include insider executives’ share increases or decreases in your decision-making?

About $MSTRB city value surpassing Ford, the fiercest clash isn’t really between bulls and bears—it’s between valuation methods. I lay out both sides’ arguments, and then I’ll say which side I’m on.

The critics’ representatives are JPMorgan’s analysts: they believe Strategy’s market-value premium far exceeds the value of the bitcoins it holds, and that there’s a clear bubble component. This criticism holds up mathematically—when mNAV is greater than 1, every share you buy is more expensive by $BTC than if you simply bought the corresponding proportion directly in the market. So what exactly are you buying with the premium? You’re buying Saylor’s ability to keep issuing equity and accumulating, the leverage structure of convertible bonds, and Strategy’s exposure to scarcity within the index and ETF ecosystem.

The supporters’ logic is also strong: as long as the flywheel keeps turning (premium-issued tokens → more accumulation → higher bitcoin per share → premium maintained), the premium becomes self-fulfilling. In 2025, the company demonstrated through its $42 billion “21/21 plan” that its financing machine really can run—its holdings surpassed 600,000 BTC, representing about 3% of the total bitcoin supply. This isn’t a PPT narrative; it’s real inventory.

I tend to think both sides are right, just on different time horizons. In the short term, the premium is driven by liquidity and sentiment, and the volatility bubble critics talk about can happen at any time. In the long term, there’s only one decisive variable—the price trajectory of bitcoin itself. Strategy doesn’t have a second growth curve independent of bitcoin; all of its “operating leverage” comes from that one asset. That makes its fate both simple and brutal.

So my conclusion is: rather than arguing whether there’s a bubble, focus on the premium itself. The premium is this company’s only “product,” and it’s also the most honest emotional indicator—when it approaches 1, it means the market isn’t even willing to pay for the extra value of “trust in Saylor.”

What about you? If you had to choose one between “holding bitcoin directly” and “holding MSTR via the premium,” would you pay that premium? Why?

#Strategy market cap surpasses Ford

U.S. 30-year Treasury yields break above 5.40%. The real takeaway is this: it happens during the Federal Reserve’s easing cycle. Short-end yields are falling, while long-end yields are rising—not because of changes in policy rate expectations, but due to term premium, meaning the extra compensation the market demands for holding long-dated Treasuries.

The disagreement is precisely here. The bears’ logic chain is complete: persistent high fiscal deficits keep long-bond supply expanding; demand from key overseas holders is structurally weakening; and with repeated shocks to central bank independence, the term premium can only rise and rise further. This camp believes 5.40% is not a ceiling, but a signpost—meaning the long-end rate’s central tendency needs to shift upward systematically.

The bulls counter: risk-free coupons above 5.4% are rare configuration-level attractions for pension funds, insurance companies, and banks’ proprietary books—so supply will naturally find demand. And as long as inflation continues to cool, such a high real yield itself is the strongest counter-pressure: if bond prices are cheap enough, that is the biggest positive factor. In their view, this is a cyclical overshoot, not the new normal.

My leaning: I’m more convinced by the first half—that this up move is mainly driven by fiscal factors and term premium, not runaway inflation. This assessment directly affects what it means for the crypto market. If the term premium is being lifted slowly, the damage to long-duration risk assets like $BTC is a dull blade: it won’t cause a crash, but it will keep capping valuations. What you really need to watch is a chain reaction of liquidations from leveraged arbitrage positions triggered by a sudden surge in long-end yields.

More useful than picking sides: the tail spread and bid-to-cover ratio at the next long-bond auction—that’s where both bulls and bears cast their votes with real money.

Which side are you on: the new normal, or the overshoot?

#U.S. 30-year Treasury yields break above 5.40%

About Broadcom, what the market argues about is never demand—it’s supply. The AI business of $AVGOB —XPU chips custom-built for a few of the largest cloud providers—has orders already queued up to the absolute production capacity limit. The real bottleneck is not something you can immediately expand by just throwing money at it: it’s the advanced packaging stage. This leads to a counterintuitive conclusion: during an AI boom cycle, supply constraints are actually a good thing. With such constrained supply, order visibility is extremely high, and revenue outlook for the next few quarters is clearer than for most chip companies. But the other side of the coin is margin pressure and cash-flow stress: expanding capacity requires capital, and customer-specific customization discounting is also eroding profits. Existing analysis has pointed out that this is its biggest risk going forward. My view is that this model is a sweet kind of frustration in a bull market. But once market sentiment weakens, the same facts will be reframed just as quickly as a “growth ceiling.” When the stock price starts to drop then, it won’t be gentle. Which side are you on: treating the capacity bottleneck as a moat, or as a hidden risk? Share your reasoning.

In the U.S. AI chip sector, the controversy around $AMDB isn’t exactly low: Piper Sandler has issued an aggressive forecast—by 2030, AMD’s earnings per share could grow at a compound annual rate of about 65%. What does 65% compounding over five years really mean? $1 could turn into roughly $12, representing more than a tenfold increase in EPS. There are two assumptions underpinning this curve: first, AI accelerator business will capture a significant share from the current leader; and second, server CPUs will keep steadily advancing. I’m naturally wary of any super-long-term predictions: if, within five years, AMD’s MI-series misses even a single generation, or if cloud providers’ in-house chip development ramps up even faster, this curve would break. Still, looking at it another way, the fact that Wall Street is willing to put 65% into a formal report suggests that AMD has already moved from being merely “a shadow of the leader” to a company with its own independent narrative. Do you believe this number, or do you think it’s just another “perpetual-motion” report?

The signals coming from the U.S. Treasury market are becoming increasingly unsettling: long-term yields remain stubbornly high, and worries about the federal budget deficit and the scale of new debt issuance continue to intensify. Some analysts have already described this series of signs as a fairly significant warning. Whenever traditional safe-haven assets begin to feel “unsafe” on their own, discussions about Bitcoin tend to heat up—what Satoshi Nakamoto wrote years ago in the genesis block was that headline about bank bailouts. The very motive behind its creation was itself a response to the old system.

The logic isn’t complicated: if volatility in U.S. Treasuries increases and investors start to re-evaluate sovereign-debt risk, scarce assets and decentralized narratives may have an opportunity to absorb some of the safe-haven demand. Of course, Bitcoin has still mostly been treated as a risk asset; in the short term, it more often follows liquidity swings. It’s not realistic to expect it to immediately replace U.S. Treasuries. But this round of tension in the bond market is indeed adding new commentary to Bitcoin’s long-term story. Do you agree with the saying “U.S. Treasury risk equals Bitcoin opportunity”? Let’s talk in the comments. $BTC

Today is September 15, and the U.S. Senate is holding a procedural vote on the Clarity Act. First, let’s make the weight of this clear: at the core of this digital asset market bill is drawing the regulatory boundary for digital assets between the SEC and the CFTC—what counts as a security, what counts as a commodity, and who oversees the secondary market. It previously passed the House with a big bipartisan margin of 294 to 134; the real challenge has always been in the Senate.

What is being voted on today is not the bill itself, but the "motion to proceed"—the ticket that gets the bill onto the full chamber’s debate docket. This ticket requires 60 votes, meaning the minority party has an effective veto power, and every vote must be negotiated across party lines.

Why does this matter so much for the market segment? Because the four words "regulatory clarity" are the final missing piece for institutional capital to move in. Spot ETFs address the question of "can you buy"; Clarity addresses "what rules you play by after you buy." If the push goes smoothly, the most direct beneficiaries would be: mainstream assets with the strongest compliance expectations, $BTC and $ETH ; custody and RWA infrastructure; and the altcoin segment that has long been struggling under "securities classification."

Procedurally, even if it clears 60 votes today, there will still be a tug-of-war over amendments and a final vote afterward—possibly weeks more. But the directional signal will be priced in at the moment of the vote. Conversely, if it fails, a short-term sentiment shock is unavoidable, though the bill won’t die; Congress will bring it back during the fall agenda.

Do you think they can round up the 60 votes this time? Place your bet in the comments section.

#ClarityAct September 15 procedural vote

The Wall Street Journal reports that Tesla has regained most of its share of the U.S. electric-vehicle market, and the reason is quite interesting: it’s not that Tesla sells that much more aggressively itself—rather, its competitors have collectively pulled back. Several automakers have cut or scaled back their EV model lines, and the space that opened up was handed back to Tesla. Even more interesting is the timing: just as everyone was discussing slowing EV demand and companies were shrinking their investments in their earnings reports, market share was quietly re-concentrating. The EV market has shifted from a “bandwagon everyone wants in on” to a business where “only those who can do the numbers stay.” In between was nothing more than the reality of subsidy rollbacks and demand polarization over these past two years. For $TSLAB , market share has always been the anchor of the valuation narrative: when it moves back toward being the sole dominant player, the bearish argument that demand has peaked has to be weighed more carefully. But leading in market share and having healthy profit margins are two different things—the costs of a price war will ultimately show up in the financial statements. As for the next turning point in U.S. electric vehicles, who do you think will trigger it?

A number to feel how powerful Apple’s cash machine is: during Tim Cook’s 15 years as CEO, Apple has cumulatively repurchased more than $880 billion of its own shares. What does $880 billion even mean? It’s more than the full-year GDP of most countries, and higher than the total market value of many large tech companies. The repurchased shares are directly retired, which means the number of outstanding shares keeps shrinking—so even if the company’s overall value stands still, each share represents a larger slice. Add dividends on top, and it becomes one of the most stable shareholder-return engines of the past decade-plus. The confidence behind this is that the consumer electronics business generates steady cash flows, allowing Apple to avoid the painful trade-off between “keep investing” and “reward shareholders.” Some criticize buybacks as financial engineering rather than innovation, while others argue this is exactly the capital discipline a mature company should have. $AAPLB —this $880 billion: do you think it’s value creation or value moving?

Anthropic has set its IPO on Nasdaq rather than the New York Stock Exchange that most people had previously expected.

This choice itself is a signal. Nasdaq has long been the home base for tech and growth stocks—from chips to software to AI. It has greater trading depth, and market makers have the richest experience in pricing high-volatility growth companies. The NYSE, on the other hand, has won a number of large tech company IPOs over the past two years, and the market had largely been betting on that. By choosing Nasdaq, Anthropic has made its self-positioning very clear: “I’m a pure tech company—please value me as a tech company.”

Even more noteworthy is the scale. Reports say that in its most recent funding round, Anthropic’s valuation was already in the $100 billion range (about $183 billion). If it issues at the higher range that has circulated in the market, this would become one of the largest IPOs in history. For this Claude AI company, the meaning of going public is also very practical: spending on compute for training and inference is a cash-eating machine. Beyond equity financing, the public market is a deeper pool of capital.

For ordinary users, there’s really only one thing you can do right now: wait for the S-1. The prospectus will be the first to disclose real revenue, growth rate, profit margins, and cash burn—only then does the moment come to turn “AI belief” into numbers. Until then, all valuations are just narrative from the primary market.

Look one layer deeper: the competition among exchanges for AI companies is similar to the last decade’s battle for internet companies—Nasdaq’s win with Anthropic is essentially抢拿 this era’s core “asset label.”

With the AI company IPO wave here, will you pay attention to their prospectuses? Which number are you most eager to see?

#AnthropicChoosesNasdaqIPO

The number of holders of tokenized stocks grew by 619.1%—the figure sounds outrageous, but when you break it down, the logic isn’t mysterious.

First, let’s clarify the product: tokenized stocks are on-chain mappings of traditional shares. The underlying real stocks are held by a licensed institution, and tokens corresponding to those stocks are then issued on the blockchain. The token you hold represents a claim on that pool of shares, not direct shareholder status in the company. Starting around mid-2025, a batch of compliant issuers moved dozens of U.S. stocks and ETFs onto public blockchains. Everything from tech stocks like $NVDAB to index ETFs is included. Combine that with three things traditional markets can’t do—7×24 trading, fractionalized purchases, and instant settlement on-chain—and the number of holders naturally rises.

So the 619.1% doesn’t reflect a “profit-making effect.” What it really shows is the scale of a “try-it-out” effect: the base is small, so the growth rate looks dramatically scary.

If you want to participate, the four things that ordinary users should truly verify are:
1. Who is the issuer? Where are the underlying stocks custodied, and is there a 1:1 reserve proof?
2. The mechanism for handling when the token deviates from its net asset value—historically, tokenized assets have脱锚ed in extreme market conditions. What you buy on-chain might not be “stock,” but “sentiment.”
3. Your legal status—many products don’t come with shareholder rights like voting.
4. Whether your jurisdiction allows you to hold this kind of asset.

My view: tokenization is one of the most concrete tracks in the RWA (Real-World Assets) narrative, but within today’s growth rates, there’s definitely speculative air. Only issuers that survive after the tide goes out and any de-pegging events have already happened are worth trusting long-term.

Will you ever encounter tokenized stocks? What are you most worried about? Let’s discuss in the comments.

#TokenizedStockHoldersGrew619.1%

Privacy DeFi has a new breakthrough: after Zama and Morpho’s first confidential vault TVL surpassed $40 million, the product line has officially expanded. Users can now access 12 existing Morpho vaults on Ethereum in a confidential way, along with 4 newly built fully confidential vaults. In addition, a private swap feature has been launched. The technical core is Fully Homomorphic Encryption (FHE): your deposit amount, positions, and yields are all encrypted, and computation is carried out directly on the ciphertext. Anyone on-chain cannot see how much you borrowed or how much you earned. This hits one of DeFi’s biggest pain points: full transparency. Whale positions are monitored, sniped, and even targeted for liquidation—these are the costs of transparent ledgers. If FHE can run at an acceptable cost, the final privacy piece for institutional capital to enter DeFi will be in place. $40 million is only the beginning—this direction for the $ETH ecosystem is worth ongoing attention. Would you be willing to accept the extra costs that FHE brings for privacy? $ETH

Tom Lee put it very bluntly this time: Ethereum’s killer app isn’t some breakout dApp; it’s becoming the settlement layer for Wall Street and artificial intelligence. The weight of this judgment lies in the fact that it comes from a traditional research institution like Fundstrat—when Wall Street starts treating $ETH as financial infrastructure rather than a speculative asset, the narrative has already changed. The logic isn’t complicated: institutions issue bonds, funds need settlement, and AI Agents need machine-to-machine payments—all of which require a neutral, permissionless clearing network, and Ethereum is currently the most mature option for these needs. Even more noteworthy is that he made this statement while the broader market was weakening, yet ETH held up against the trend. Money votes with its feet—more convincing than any research report. A killer app is never designed; it’s put to use. Do you think the Wall Street settlement narrative can support ETH’s next valuation upswing?

Liquidations worth $674 million across the entire web in 24 hours—let’s lay out the facts first: in this sell-off, $BTC was the first to break below a key support level. Long positions were wiped out in a coordinated way. In the past 24 hours, total liquidations across the whole market were about $674 million, and the vast majority were long positions—people betting on a rise got taken out in one fell swoop. $BTC and $ETH account for the two largest portions, but what’s truly brutal is the altcoins: mainstream coins fell roughly in the 3%–5% range, while some smaller alts and meme coins dropped 10%–20%. As a result, liquidation amounts relative to their market caps were dramatically amplified.

Why does this happen? Liquidation is never the cause—it’s the result. The mechanism is simple: price drops → the risk-control engine forces liquidations → liquidations themselves add further sell pressure → price keeps falling → the next wave of forced liquidations is triggered. This is what people call the “waterfall.” It doesn’t require any new news—only enough leverage in the market.

And leverage is exactly what had built up the most over the previous few weeks: funding rates staying persistently elevated and open interest rising continuously indicate that not only were more people going long, but doing it at a higher cost. At this point, any macro disturbance—hawkish remarks, economic data that dampens rate-cut expectations, or large holders moving coins to exchanges—could be enough to light the fuse.

One detail worth remembering: after the waterfall sweep, funding rates usually return to neutral or even turn negative. With leverage flushed out, market structure often becomes healthier. That’s why experienced traders don’t fear the liquidation event itself—they fear *their own* position taken with excessive leverage.

In this round, did you buy the dip, cut losses, or get washed out? Let’s talk about your response in the comments.

#全网爆仓6.74亿美元

Two population data sets simultaneously topped the trending searches. Total: 1.405 billion. New births: around 8 million. Taken separately, they’re just numbers. Put them together, and you can extract five things.

1. The total is still stable. The increment is gone. 1.405 billion suggests the base is basically holding. But an annual birth rate of 8 million means the growth engine is sputtering out. Where the total population goes in the future depends on the lifespan curve, not the fertility curve.

2. The baby-economy moves into early winter. Infant formula, maternal-and-child products, early education. What they face isn’t just a slowdown—it’s a shrinking customer pool. Industry shakeout will happen faster than you think.

3. The demographic dividend is still there, just in a different form. With a large total population, the scale effects of a single market still remain. Consumer-demand brands have far more room for error than small countries do.

4. The story of labor costs needs to be rewritten. It’s not simply that “labor has gotten more expensive.” It’s a layered kind of expensive: scarcity-skill premiums accelerate. Replaceable jobs get eaten by automation first.

5. This one is the most important—and the most underestimated: 8 million is a slow-moving variable. But the capital market prices it as a fast variable. By the time population data “officially worsens” and hits the trending searches, relevant assets have already been priced in. The panic you see now may already be someone else cashing out.

Save this. Every year for the next decade, come back and check—verify it step by step. Which one do you think I got wrong? Or is there a sixth point? Add it in the comments.

Revolut’s user data was stolen by a “forged government email.” Many people’s first reaction is to laugh at a fintech company being so easy to fool—but what really needs to be understood is why these attacks specifically target financial institutions and often succeed.

This scam has a formal name: forged legal process. The attackers impersonate courts or law-enforcement agencies and send what appear to be legitimate subpoenas or “urgent disclosure requests” to the company’s legal and compliance teams. Its impact is in the design: legal teams receiving a “court order” must respond within a limited time window. Refusing may be construed as obstructing justice, but complying may not leave enough time to verify every detail of the document. The attack surface isn’t a system vulnerability—it’s the process and people constrained by institutional deadlines. For fintech firms with tens of millions of users, one mistaken judgment can lead to large-scale data leakage.

Revolut is a prime example of this kind of target: it has expanded extremely quickly, has a massive user base, and its compliance program has consistently lagged behind the pace of the business. Moreover, its licensed entities span multiple jurisdictions—so any “government” can be fabricated. It’s also not the first time data was leaked: in 2022, it previously exposed information of around 50,000 users due to issues with a third-party interface. The problem has never been a single point; it’s whether risk controls can keep up with the scale in a fast-growing company.

Next, there are three things to watch: how regulators classify this “forced handover”—whether it counts as force majeure or corporate negligence—which will determine fines and liability for compensation; whether the entire industry will build a unified channel to verify legal documents; and a question every user should ask customer support: is your financial app able to hand over your information based solely on a single email? #Revolut被骗交出用户数据被假政府邮件骗取

The Federal Reserve’s decision on interest rates is set for Wednesday, and the rate futures pricing has already pushed the probability of a rate hike to 87%—the pricing itself shows just how much the market is worried about inflation, and for $BTC this week is a real pressure test. What’s even more worth pondering is the view of Custodia Bank founder Caitlin Long: she believes that it’s no longer the Federal Reserve that truly controls the market liquidity switch; it’s the U.S. Treasury instead. The opening and closing of Treasury accounts determines how tight or loose the U.S. dollar is more directly than the federal funds rate. Following this logic, rate hikes themselves may not immediately wipe out risk assets, but if the Treasury simultaneously pulls liquidity out, then under this double squeeze, Bitcoin is unlikely to be spared. Instead of trying to guess the direction, it’s better to watch two things: the wording in the FOMC statement regarding inflation, and changes in the Treasury account balance. Also, don’t forget that 87% pricing leaves virtually no room for surprises—if there’s no rate hike, the rebound could be stronger than you might expect. How would you bet?

It’s down about 50% over the past year, but revenue is accelerating—$ORCLB is now one of the most controversial contradictions in the U.S. stock market. Evidence for the bulls: AI-related orders have pushed the backlog contract value (RPO) to record highs, and the growth engine is OCI, Oracle’s cloud infrastructure business. Enterprises have moved their AI training and inference budgets from on-prem data centers into the cloud. Evidence for the bears: To fulfill these orders, capital expenditures and debt have surged in tandem. The market has started to question the payback cycle of AI infrastructure, and the valuation logic has shifted from a growth-stock framework to a heavy-asset utility model. After the stock was cut in half, the question is actually very simple: do you believe Oracle can turn this massive backlog into real revenue and cash flow? If conversion goes smoothly, the current price is a gold mine; if not, the heavy-asset + high-leverage model will continue to be punished by valuation resets. A Buffett-style reminder: a 50% drop doesn’t automatically mean it’s cheap—the key is the certainty of future cash flows, not the magnitude of the decline itself. Would you buy $ORCLB on the dip?