海盗鸭

From September 15 to 20, 6,678 wallets on the XRPL were emptied—about 11.7 million $XRP and nearly $20 million were transferred out, executed in six waves. First, someone manually drained eight wallets holding more than 99,999 XRP each; then a script scanned 1,682 wallets; after that, the next five waves reused the same set of scripts and private keys. Even harsher, the attacker sent 5,001 AccountDelete transactions to delete the accounts entirely, taking the remaining reserve funds as well.

Key point: This isn’t an XRPL vulnerability—every transfer was signed legitimately. What went wrong was the way the private keys were stored, mainly pointing to the D'CENT App wallet. Around 5.6 million XRP have already been bridged to Ethereum for cash-out.

My take is that self-custody has been hyped too much: importing a hardware wallet’s seed phrase into a phone app is like downgrading a cold wallet into a hot one. Once a seed phrase has appeared on a connected device, it shouldn’t be trusted anymore.

Have you ever imported your cold wallet’s seed phrase into a phone app?

A notable signal emerged last week in Citigroup’s profit revision index: the number of analysts cutting U.S. corporate earnings forecasts for the first time in 23 weeks surpassed those raising them, ending the longest cycle of upward revisions since September 2021.

The mechanism matters more than the numbers themselves. Stock prices roughly equal earnings divided by the discount rate, and this round of pressure is coming from both ends at once. On the numerator side, earnings were lowered for consumer staples, discretionary, materials, and financials. Kemper from BNP Paribas Wealth Management pointed directly to rising cost of living and energy prices as the cause. On the denominator side, the Federal Reserve raised rates by 25 basis points this month (the first increase in three years) and signaled that there could be another hike later this year; meanwhile, the 10-year Treasury yield has already moved close to the key threshold of 4.5%.

Historically, when “earnings downgrades + higher rates” occur together, what often follows is valuation compression rather than an earnings collapse. Helen Jewell of BlackRock reminded that the market currently still expects U.S. corporate earnings to grow at a rate of 15% to 18%, and that figure alone leaves plenty of room for downward adjustment. Michael Wilson of Morgan Stanley issued a quantitative warning: if valuations keep slipping and energy prices push policy tightening further, the S&P 500 could fall by as much as 7%.

My view is clear: this looks more like the end of a valuation-expansion phase than the beginning of an earnings downturn. The truly dangerous variable isn’t the pen in analysts’ hands—it’s oil prices. Brent returning above $100 is precisely the direct catalyst for rekindling the tightening expectations.

Let me leave you with a question: if the 10-year U.S. Treasury yield really does move above 4.5%, which type of asset would you cut first—long-duration bonds, growth stocks, or crypto?

#WallStreet profit forecasts turn bearish for the first time in 23 weeks

$ETH From around $1,900 to above $2,800, it has now pulled back to 2,678. The weekly chart is still up 8%. In this pullback, three signals have shown up on-chain.

First, the ETH balance on exchanges is declining—liquidity is being moved toward self-custody and staking. Second, the priority fee is rising—there are truly people competing for block space on-chain; it’s not just transfers. Third, exchange stablecoin reserves are being replenished—money comes first, and the orders follow.

I don’t buy the simplistic equation “on-chain indicators = bullish,” but I do recognize priority fee: it reflects real demand, not wallet shuffling. On Ethereum, those who can continuously pay high priority fees are basically DeFi, L2 settlements, and traders trying to front-run. When that activity returns, it matters more than the price itself.

Risks are also very straightforward: if fees are rising only because a small number of bots are front-running, then this move is fake. How do you usually tell these two situations apart—has real demand returned, or are bots propping up prices against each other?

#Ethereum Breaks Through $2700

September 25: Ethereum is trading above $2,700—at $2,708.77—up 2.55% over the past 24 hours. On the same day, $BTC was at $84,709.7, up 2.02%. Put those two numbers side by side: in this round, $ETH is the one that’s outperforming.

But there’s not much to say about the price itself. What I care about more is another piece of news that happened almost at the same time: Bitwise filed with the SEC an amended S-1 for its spot Ethereum ETF, laying out an entire section covering the staking mechanism, validator operations, slashing risk, and how staking rewards are accounted for. In the original filing, this issuer had explicitly written in black and white that it would not participate in any staking—this time, it effectively retracts that statement.

It’s crucial to clarify first: as of now, the SEC has not approved any spot Ethereum ETF that allows staking. This is still only a proposal.

Why is this more important than $2,700? Because it changes the ETF holders’ revenue structure. Ethereum’s network-wide staking rate is about 34.7%, which corresponds to an individual staking yield of roughly 2.6%. The ETF would also need to charge management fees and set aside costs for custody and validator operations; after deductions, the net yield is likely to land somewhere around 1.9% to 2.2%. In other words, an ETH ETF that cannot stake will, in the long run, fail to outperform simply holding the coins and staking them yourself. Only by putting staking into the product does the ETF become eligible to compete with the “do it yourself” pathway. At present, cumulative net inflows into US spot Ethereum ETFs are about $13.3 billion, with total assets around $16.72 billion—roughly 5.2% of Ethereum’s market cap. This scale isn’t huge, so there’s actually more room for marginal changes, which makes it worth watching closely.

My view is that the pricing focus for $ETH in this round is shifting from “narrative” to “cash flow.” Spot buying can push the price up to $2,700 in a single day, but whether it can stay above $2,700 depends on whether this staking path can be truly unlocked—and how many new investors are willing to hand over their coins in exchange for that yield.

The downside is also clear: who bears slashing losses, validator concentration, the operational capacity of the custodian—until these details are finalized, everything remains in proposal form.

If staking for an Ethereum ETF is ultimately approved, would you swap your ETH for ETF shares, or would you keep staking it yourself?

#Ethereum breaks above $2,700

BCH rose about 28% within 24 hours, becoming the best-performing asset among the top 100 by market value. Its price, which hovered around $270, briefly surged above $340. The trigger was CME’s announcement that it would list Bitcoin Cash futures on October 19 (standard contracts for 250 BCH and micro contracts for 25 BCH), pending regulatory approval; on September 11, Grayscale also submitted an amended filing to convert its BCH trust into a spot ETF (ticker BCHG).

The bulls are looking at access. Trading volume increased from $1.47 billion to $2.13 billion, and open interest rose from roughly $350 million to $600 million—the highest since May of this year. The launch of futures means institutions can trade both directions, trade the basis, and market-making depth should follow; a spot ETF is another compliant channel. In the business of “compliance premium,” the biggest problem with an old coin is not the technology, but that there aren’t many incentives for people to trade it—CME happens to fill that gap.

The bears are looking at fundamentals. This rally wasn’t accompanied by any changes at the protocol level; what surged were positions and sentiment. Data from some platforms shows the sell-order ratio as high as 63%. Even more striking is the statistic: after 49 similar big breakouts over the past two years, the 30-day median return was -6.7%, and many cases fell back below the original breakout level.

I tend to believe this is an “access re-pricing” rather than a “value re-pricing”—its persistence doesn’t depend on the candlestick chart, but on whether the October 19 rollout can land on schedule. If it’s rejected or delayed, the gains will be quickly given back. By the way, during the same period $BTC briefly fell below $86,000; the money was moved from BTC, not new capital.

So here’s the question for you: is this old coin’s second compliant spring, or just a sentiment pulse worth only shorting implied volatility? Would you chase $BCH ?

#BCHrisesAbout28%AfterCMEfuturesListing

$ETH This rebound turned $2,700 into a real line in the sand.

Price action has been choppy: around September 21 it quickly surged from about $2,644, broke above $2,700, touched $2,701, then dropped back to around $2,650 shortly after. It then repeatedly probed $2,708 and $2,767, and when September 23 came around it extended to about $2,818 before being pushed back—falling more than $14 within 15 minutes. Above, there are dense sell walls around $2,708 and $2,780–$2,800. Below, step-like supports sit at $2,700, $2,650–$2,660, and $2,600.

The divergence is actually very clear: the fundamentals and the technicals are telling two different stories.

The bulls have hard data. The exchange balance of $ETH is about 14.6–14.8 million coins, the lowest since 2016. Staked/locked amounts are around 43 million coins, nearing 35% of total supply. In the U.S., spot ETH ETF net inflows in Q3 totaled about $10 billion, with total assets around $16.7 billion—roughly 5.2% of $ETH ’s total market cap. Leverage on the derivatives side also seems tilted toward the bulls; one statistic shows that if price can effectively hold above roughly $2,702, the mainstream platforms’ cumulative short-liquidation size could reach about $880 million.

The bears have positioning and macro factors. The 1-hour and 4-hour RSI has entered the overbought zone, ADX is above 40, and after the spike the candles formed long upper wicks. The MACD red histogram has shortened, and momentum is clearly weakening. ETF flows saw continuous net outflows on September 16–17, only turning back with about $144 million in inflows on the 18th. More importantly, the macro picture has shifted: on September 16 the Fed raised rates by 25 bps to 3.75%–4.00%, and market pricing for another hike in October briefly rose to 53%–70%.

I tend to believe that the number 2,700 itself isn’t the key. What matters are two things: whether spot sell pressure around the $2,800 area can be absorbed with volume, and whether the rate-hike path will be derailed by inflation data. In an overbought state, chasing higher directly doesn’t offer a good risk/reward.

Which approach are you taking—adding on strength above $2,700, or waiting for a pullback to $2,650 before discussing?

#ETH breaks above $2,700

69.7%. This is the probability, as given by the CME “FedWatch” on September 24, that the Fed will raise rates by 25 basis points at the October policy meeting. The current target range for the federal funds rate is 3.75%–4.00%. Translated into plain terms, the market sees “one more hike” as nearly the default option, with the probability of holding rates steady down to just 30.3%.

First, make sure you understand how this number is derived—otherwise it’s easy to get led around by it. It comes from the implied pricing of interest-rate futures: consensus staked by a group of traders using real money, not a statement from the Fed. And it moves every day. Just a few days ago, this probability was still around 50%. The batch of data released on September 23 was the trigger: the composite PMI initial reading of 58.4, Brent crude oil returning to $100, and weak demand at the 5-year Treasury auction—pushing the probability to nearly seventy percent.

What should ordinary users pay attention to? I think it’s the following three things, not the news item itself.

First, don’t treat probability as fact. A 69.7% figure still implies roughly a 30% chance that the Fed will do nothing in October, while the market is pricing as if “a hike is certain.” That kind of one-way pricing is exactly what a position is most vulnerable to being blown up by a contrary data release.

Second, watch the bond market—don’t only listen to speeches. In this rise of long-end yields, the contribution from real rates is clearly greater than that from inflation compensation. In other words, the market is repricing “fiscal supply” and “real returns.” This isn’t something that will change direction because of a single meeting.

Third, lay out your scenarios first. An October hike followed by post-meeting remarks suggesting the job is already done is entirely different from a hike followed by continued emphasis that more hikes will come. Does your current positioning assume one of these outcomes—and does it leave you any room to get out if you’re wrong?

So, how would you judge the timing of the final rate hike in this cycle—will it land in October or December?

#Fed October Rate-Hike Probability Rises to 69.7%

The Federal Reserve on September 24th issued two proposed rules at once, translating the “pay stablecoins” line from the GENIUS Act into actionable provisions. Key points: for every $1 in tokens issued, the issuer must hold at least $1 in qualified reserves; qualified reserves are limited to U.S. dollar cash, Federal Reserve balances, demand deposits, U.S. Treasury securities with a remaining maturity of no more than 93 days, and overnight repos, among others; redemptions must generally be completed within two business days. Capital requirements follow a tiered structure—2% for the first $20 billion in outstanding circulation, 1.5% for the next $30 billion, and 1% for amounts above $50 billion. The public comment period is 60 days, and the bill’s effective date is anchored to January 18, 2027.

There’s almost no disagreement in the market on the “1:1” issue—it’s effectively settled. The real debate is about the other three items.

First, whether tokenized Treasuries count as qualified reserves. This determines whether the stablecoin’s underlying backing is T-bills resting in the traditional custody system, or tokenized Treasuries that can circulate natively on-chain 24/7. If it’s the latter, both the issuer’s asset side and liability side would run on-chain, and operational efficiency would be another order of magnitude; if it’s the former, banks’ advantages in custody and capital would be amplified.

Second, the tiered capital provisioning. The 2% rate for the first $20 billion and the drop to 1% above $50 billion is a clear scale benefit for large issuers—essentially using regulatory costs to build a moat for top players. For latecomers to catch up, they must first cross the most expensive tier. I’d argue this is the least discussed area in the proposal, yet with the deepest implications.

Third, anti–money laundering (AML) thresholds. Board member Michael Barr publicly said he was dissatisfied with the standard that “violations must reach a material or systemic level to trigger enforcement,” calling it too lax. This kind of rare public split within regulation is telling.

My take: at its core, this proposal is about moving issuance power into the banking system—not opening it up to the outside. The dispute over reserve ratios is over; the next battleground is the form of assets on the asset side.

Which matters more to you—whether tokenized Treasuries can be included in the pool, or that 2% tiered capital? Whichever lands first largely determines who will be able to issue stablecoins over the next three years.

#Federal Reserve drafts rules for banks to issue payment stablecoins

XRP surged about 6% in 24 hours, back to $1.53–$1.56. It’s up more than 15% over the week and is leading among major coins. Behind it are three things stacking together: Spot XRP ETFs saw net inflows for three consecutive days ($20.02M, $18.04M, and $14.89M), with cumulative net inflows setting a new high of over $1.76B; analysts counted that whales bought 470M XRP in five days (about $724M); and in the 4-hour timeframe, short liquidations surged by over 45%.
However, the day before it just broke below $1.50. This move is more like a rebound after being suppressed around $1.65—the price action isn’t exactly clean. My take is fairly cautious: these ETF inflows are roughly only about 0.4% of XRP’s daily trading volume. What really drives the price is the short squeeze and giant whale accumulation; the ETFs are more like emotional endorsement. XRP is still down about 16% year-to-date. $XRP
For this move, do you think of it as a reversal, or just a bounce?

XRP ($XRP ) rose about 6.6% in a day, back to $1.55, after briefly dipping below $1.50 the day before. Three forces are stacking up: spot XRP ETF cumulative net inflows have surpassed $1.76 billion; over the past three trading days this week, another roughly $53 million flowed in; over the last five days, whales bought more than 470 million XRP, worth about $724 million at current prices; and on the 4-hour timeframe, the short liquidations surged by more than 45%.
But I don’t think those ETF inflows alone can support a 6% move. In total over three days, it’s only a bit over $50 million—on the order of tens of millions per day—nowhere near the daily trading volume scale of XRP. The real driver is the squeeze of leveraged shorts, with spot buying mainly acting as background noise. Whale accumulation is real—but it also means liquidity is more concentrated, making pullbacks more fragile.
So I’m inclined to label this move as a short-squeeze rally rather than the start of a new trend. Someone in the community has set a target of $1.80 for this week.
Do you think it will first catch up to $1.80, or does this wave end right here?

September 24: US Bitcoin spot ETFs saw a total net inflow of about $191 million, marking the 6th consecutive trading day of net inflows. Sounds like good news, but it’s interesting when you break it down.

In terms of structure, the inflows are highly concentrated: BlackRock’s IBIT had a single-day net inflow of about $163 million, accounting for roughly 85% of the day’s total; Fidelity’s FBTC had about $12.9 million; Morgan Stanley’s MSBT about $10.2 million; Franklin Templeton’s EZBC about $4.9 million; and Bitwise’s BITB about $4.1 million. The only net outflow came from WisdomTree’s BTCW, at around $40 million. Most of that day’s incremental demand came from just one product.

In size: total ETF net assets are about $108.9 billion, roughly 6.43% of Bitcoin’s total market cap, and the historical cumulative net inflow is about $57.4 billion.

What’s more worth watching is the marginal change. In the previous trading day (September 23), net inflows were $346.9 million. On September 24, the figure fell by about 45% month-over-month; going further back, September 21 was about $999 million, and September 22 about $714.7 million. From September 17 onward, over 6 consecutive days, cumulative inflows totaled about $2.84 billion. The absolute level of inflows is still there, but the slope has clearly flattened.

Meanwhile, on-chain data shows short-term holders moving about 47,600 $BTC to exchanges—worth more than $4 billion. This likely explains why large ETF inflows didn’t translate into sustained price gains: the ETF “is moving cargo into storage,” while exchanges are “unloading cargo out.”

My take is: ETF net inflows are shifting from being a “leading indicator for price” into a “speedometer for moving inventory.” Even with 6 straight days of inflows, the price around $88,000 started running into profit-taking—suggesting this portion of buying is more about absorbing existing supply rather than creating new incremental demand. For ordinary users, instead of obsessing over the daily positive or negative, it’s better to compare two figures: ETF net inflows minus exchange net inflows, which is the real net increase.

One more reminder that’s easy to overlook: intraday daily data is often revised, and different reporting methodologies may assign the date slightly differently—sometimes by a day. Don’t take any single day’s numbers as the basis for decisions.

When you look at the flow data in your day-to-day analysis, do you trust ETF net inflows more, or exchange net transfers more? Is there a third number you care about more?

#Bitcoin spot ETF net inflow of $191 million

September 15, the U.S. Senate failed to pass the procedural vote on the “CLARITY Act” by a 49–50 margin. The bill is stuck, but regulation didn’t stop—on September 24, CFTC staff updated their FAQs related to crypto assets and blockchain technology, adding four new questions and revising one. This matters more than it seems.

Three key updates:

First, customer funds can be invested in tokenized assets. Futures commission merchants (FCMs) and derivatives clearing organizations (DCOs) may invest customer funds in tokenized forms of products permitted under Regulation 1.25, such as tokenized money market fund shares. The requirement is that these tokens must carry legal and economic rights that are equal to or functionally equivalent to those of the underlying traditional assets, and must meet liquidity, concentration, maturity, and custody requirements.

Second, on-chain records can count as compliance records. Distributed ledger technology can be used to satisfy recordkeeping obligations under Regulations 1.31 and 45.2. When using private chains, it’s not required to keep an off-chain backup copy. However, for public blockchains, you must demonstrate that records can still be retrieved and produced in an emergency or during network disruption.

Third, eligible tokenized assets can be used as uncleared swap margin. The capital requirements are also specific: when an investment firm uses stablecoins it holds in-house to pay the remaining interest in a customer segregated account, it must hold at least 2% capital. For in-house $BTC and $ETH positions, it must hold 20% capital, aligning with the SEC’s discounted framework for broker-dealers.

Two mechanism-level points are worth noting. First, this is “staff guidance,” not a rulemaking. It neither creates nor amends rules, and it does not necessarily reflect the Commission’s overall position. The benefit is speed; the downside is fragility: one letter can be withdrawn by the next. Second, with the legislative route blocked, regulators can only push forward using their own tools. The SEC has also indicated it is prepared to propose rules on its own even if Congress does not act. That means in the future, the boundaries of compliance will be determined by when guidance and FAQs are published, not by the text of a bill.

I tend to think that in the short term this is a real positive for tokenized collateral and on-chain bookkeeping; but it also shifts risk from “whether Congress will legislate” to “whether the next piece of guidance will tighten the rules.” For institutions, the latter is less predictable.

Would you treat this level of guidance as enough certainty to add to positions, or would you rather wait for a real bill?

#CFTC updates guidance on tokenized assets for regulated entities

Last night I ordered a takeout delivery for 25 yuan by myself. I came across a post showing that the merchant’s take-home pay was 6.08 yuan, so I went ahead and checked the order details to make sure the numbers matched.

On Zhihu, this question has been pushed to the top these past two days: “For a single order of 25 yuan, the merchant only gets 6.08 yuan—where does the rest of the money go?” The comment section is heated. One side says the platform’s commission cuts are too harsh, while the other says riders and delivery cost money in the first place.

Looking at the line items, the parts that do add up are only a few yuan: platform commission, delivery service fees, the promotional subsidies the merchant pays for out of pocket, and then packing and ingredients. What makes merchants the most uncomfortable isn’t really those small commission percentages—it’s that “subsidies” line item. It often means merchants spend their own money to buy traffic. They pay the cost, but customers remember that there were platform discounts.

The 6.08 yuan doesn’t necessarily reflect the general situation—it varies wildly by city and by category. But the direction it points to is real: in low-priced orders, the merchant’s pricing power is taken away by the rules of the promotions. Who do you think is mainly to blame for this accounting?

On the night of September 23, the Central Bank of Brazil issued two more resolutions—No. 588 and No. 589—formally pulling self-custody wallets into the reporting network.

Let’s clarify the key figures first: any virtual-asset transaction involving transfers into or out of a self-custody wallet—if a single transaction is equal to or exceeds $10,000 (about R$51,700)—requires institutions to report to the Financial Activities Control Committee (Coaf). The deadline is the next business day after the transaction, and it is not conditioned on “suspicion”; if it meets the threshold, it must be reported. “Self-custody” means the wallet where the user holds the private keys themselves, with no intermediary custody of the assets. The central bank’s rationale is straightforward: these operations provide the least monitoring information available, so transparency must be added.

The accompanying Resolution No. 589 is the other side of the coin: it prohibits financial institutions, payment institutions, and authorized crypto businesses from conducting virtual-asset business with entities that have not been authorized by the central bank and are not included in the authorization application process.

Remember three timeline points: the new rules take effect on October 1, at the same time as the deadline for virtual-asset service providers to submit authorization applications; starting November 6, trading with unauthorized entities is officially banned (originally set for October 30, later extended by one week); and on January 1, 2027, partial data-reporting obligations take effect.

What best illustrates the current situation is a comparison: as of September 18, the central bank had received only five authorization applications (four under review and one rejected), while market estimates suggest Brazil has about 120 crypto service companies. The gap between the compliance window and the industry’s actual scale is the biggest source of uncertainty in the coming months.

These rules already had groundwork: Resolution No. 521 issued in May 2026 requires service providers to identify self-custody wallet owners, and in July, the Federal Tax Authority’s DeCripto reporting regime went live. Same action, but investors now face multi-layer reporting.

I don’t think this is “banning self-custody.” What it changes is the attribute of on-chain transfers—from anonymous displacement to identity-based reporting. But there’s an obvious gap: the text does not clearly specify that multiple small transfers split across batches should be aggregated and counted toward the $10,000 threshold, meaning there may still be room to split and evade in theory. So it’s more of a deterrent than an outright interception.

If the market you’re in copies these rules tomorrow, will you first change your withdrawal habits, or will you just split things up? Tell me what you’d do.

#Central Bank of Brazil requires reporting transfers of $10,000 self-custody wallets

Grayscale’s Zcash spot ETF (ZCSH) has seen its net asset value pass $1 billion—despite being listed for only a month. But the key thing to watch in this news is the denominator: cumulative net inflows are only $306 million, less than 30% of total assets.

The rest is driven by price. Since ZEC began trading on NYSE Arca on August 25, it has more than doubled. The ETF’s original holdings were revalued; on top of that, its parent company DCG swapped out 85,705 ZEC to receive about $100 million worth of ZCSH shares—this is a non-cash transaction. After stripping out those two items, the truly new money coming from outside is roughly $200 million.

This is basic ETF mechanics, and also where most people misread things: AUM = number of shares × NAV per share. When the underlying rises, the AUM rises by itself. “Breaking $1 billion” by itself doesn’t prove demand; it’s more like a confirmation after the fact.

So what counts as a real signal? I’d look at suppressed supply. ZEC’s shielded supply has returned to about 4.85 million coins, roughly 30% of circulating supply—back at the start of 2024 it was only 8%. The crucial point this time is that the shielded ratio climbed together with the price, rather than being passively boosted during a bear market. For a privacy coin, the shielding rate is the indicator of whether it’s truly being used—more reliable than AUM.

Another time marker: the 3-for-1 stock split became effective on September 30. The record date was September 28. The ticker code and CUSIP remain unchanged. The split doesn’t change total value; it just reduces NAV per share to about one-third, lowering the retail entry barrier—Grayscale did a 9-for-1 split for its Ethereum trust in 2020, using the same logic. If you treat the split as a bullish signal and chase, you’re buying the wrong story.

My inclination is that, in this $ZEC cycle, the ETF’s $1 billion is the result, not the cause. Going forward, watch for a divergence between futures positioning and funding rates. If positions are still hitting new highs while the price goes sideways, and the funding rate turns positive, that suggests leverage is piling up at elevated levels—not fresh demand entering.

Let me ask you something specific: if the shielded supply share falls back to below 20%, can you still hold $ZEC ? Or are you really just buying the privacy premium itself?

#GrayscaleZcashETFNetAssetValueReaches$1B

Microsoft raised its quarterly dividend from $0.91 to $0.98, up about 8%, annualized to $3.92. The record date is November 19 and the payout date is December 10. This is its 23rd consecutive year of increasing the dividend, but its 0.8% dividend yield is clearly lower than the tech sector’s average of about 1.37%, and it even fails to beat current U.S. Treasuries.

The real story is on the other side: the prior fiscal quarter’s capital expenditures were $41.0 billion. The company’s guidance for the next quarter is $50.0 billion, and for the full year about $175.0 billion. As a result, free cash flow fell year over year by 23% to $19.6 billion, and for the full year about $67.0 billion, down 6.5%. In other words, the money spent on dividends is just a small fraction compared with AI infrastructure spending.

My take: this dividend increase from $MSFTB looks more like a confidence signal to the outside world than evidence of financial comfort. Exchanging dividend yield of 0.8% for the uncertainty around $175.0 billion in capital expenditures isn’t worth it.

Would you hold $MSFTB for a 0.8% dividend, or would you rather buy 5% U.S. Treasuries?

The US Dollar Index rose to 101 on September 23, its first time since July 30. It moved above an eight-week high, with an intraday gain of about 0.45%.

Most reports attribute the move to two things. First, the Fed raised rates by 25 basis points last week and signaled that at least one more hike is still possible later this year, keeping rate-hike expectations elevated. Second, energy prices in the European session have started to climb again, with Brent crude returning to above $100 per barrel. Since the US is a net exporter of oil, rising oil prices themselves are supportive for the dollar.

Standard Chartered strategist Steve Englander put it very plainly: the medium- to long-term dollar strength that had been predicted may finally be starting to show.

But there’s a detail that ordinary users should pay attention to: this time, precious metals were hit by a fierce selloff. Gold fell by more than $80 in a single day, dropping below 4,300, while silver plunged by 4%. Typically, a risk-off drive that strengthens the dollar would lift gold as well. This time, however, both the dollar and gold fell together—suggesting that the dominant variable isn’t risk aversion, but interest-rate differentials, and the opportunity cost of holding dollars starting to pay.

The transmission to the crypto market is direct: a stronger dollar plus higher rates means global liquidity is tightening at the margin. In risk assets, the segment with the highest beta gets hit first. Recently, $BTC has fallen back from an eight-month high and even dipped below $86,000; funds clearly rotated into assets like BCH and ZEC that have their own narratives. That is another side of the same story.

My view is that this round of dollar strength is not a one-off shock, but more like a “slow knife.” It won’t necessarily make $BTC drop 20% in a day, but it will keep raising the opportunity cost of holding non-yielding assets. For a market still in a rate-hike cycle, cash has become an option with a quote for the first time.

A concrete question: if the US Dollar Index rises above 102, would you cut altcoins first or cut BTC first?

#US Dollar Index returns to above 101 after two months

87,300 US dollars. Bitcoin touched that level twice in the past two days—both times it was pushed back. After breaking above 87,000 during the September 23 session, it turned and fell by about 3.85%, once even slipping below 84,000.

First, let’s talk about why this price level is special. This isn’t just some random round-number barrier—Bitcoin’s opening price in January 2026 was between 87,000 and 88,000, which is essentially the break-even line for the batch of buyers who entered at the start of this year. On the other side, the average cost basis of holdings for US spot Bitcoin ETFs is around 81,700 dollars. Once the price climbed above 85,900, ETF holders collectively first moved into net profit since January. With both trapped positions from earlier in the year and newly profitable positions stacked at the same level, it’s not surprising that the price can’t keep pushing higher.

Now, how did this rally start? On September 21, US spot Bitcoin ETFs saw net inflows of about 999 million US dollars in a single day, the largest daily figure this year. BlackRock’s IBIT alone contributed about 381 million. But the real accelerant was liquidation: within 24 hours, the entire market saw about 1.06 billion US dollars wiped out, of which roughly 844 million was short positions. Every time the price broke through a level, a wave of liquidations was triggered. Those forced buy-ins then pushed the price upward again, creating a stair-step pattern of “breakout—liquidation—another breakout.”

My view is that the slope from 80,000 to 87.3k was too steep; the main fuel was an ETF one-day pulse plus short covering, not sustained spot buying support. Every short position liquidated removes a future forced buyer from the market. That’s borrowed time. To check whether this is a trend, look at two things: whether ETF inflows can shift from “10 billion in a single day” to multiple consecutive days, and whether the 84,000 line can hold. On September 22, the Fear and Greed Index hit 78—entering “extreme greed” for the first time in 14 months. At positions like this, pullbacks often don’t need a specific reason.

What I most want to ask is: if the fuel for the up move was shorts getting burned, after the shorts are fully burned, who’s going to buy?

$BTC

#Bitcoin hits resistance twice at 87,300

On September 23, the 10-year U.S. Treasury yield surged to 5.13% during intraday trading—its highest level since July 2007. The 5-year yield also broke above 5% in tandem, while the 30-year yield was around 5.4%. The single-day increase of more than 13 basis points was the largest daily jump since April 2025.

Most people read this as the Fed turning more hawkish. I think that’s a misinterpretation. Breaking it down, long-term yields equal expected future short-term rates plus a term premium. Calculations by Fed staff suggest that this round of long-end yield rise was almost entirely driven by an increase in the real risk premium, while inflation expectations remain anchored near the target. In other words, the market isn’t demanding compensation for “higher inflation”; it’s demanding compensation for “longer duration.” Some estimates say the term premium has jumped by about 200 basis points to the 85th percentile of the range since 1971.

So why has the term premium risen? Supply-side factors. Federal debt is approaching $40 trillion. The annual budget deficit is close to $2 trillion. Interest payments in the current fiscal year are about $1.37 trillion, already exceeding any spending item other than Social Security and Medicare. The Treasury has expanded its long-term bond buyback operations, but one transaction only purchased $5.2 billion with a target of $6.0 billion. After the buybacks, yields actually moved higher—markets clearly aren’t giving “any face” to that. On the same day, a $7.0 billion 5-year auction was also weak: the stopout yield was about 3 basis points above expectations.

So my view is that this looks more like a repricing of debt rather than the outcome of a single Fed policy meeting. Even if the Fed stays on hold, it’s unlikely the long end will return to below 4% on its own. What really matters to watch isn’t the dot plot—it’s the cadence of issuance and the participation rate of overseas buyers.

A question for you: if the power to set the pricing of long-end rates shifts from the Fed to the Treasury’s issuance desk, do you think Bitcoin is a “victim of liquidity being drained,” or a “beneficiary hedging against fiscal trust issues”? Which side are you on, and why?

#U.S. 10-year Treasury yield hits a 19-year high

21Shares listed Europe’s first physically backed Zcash ETP on September 22 on the Pan-European Exchanges in Paris and Amsterdam. The trading code is ZCASH, and the underlying $ZEC token is custodied by BitGo. The same batch also launched a product tracking the governance token ether.fi, ETHFI.

The product details are more interesting than the news headline. First, the annual management fee is 2.5%. By contrast, mainstream European Bitcoin and Ethereum products mostly fall in the 0.2% to 1% range—2.5% is clearly a “niche premium.” Second, the launch size is so small it’s a bit awkward: 5,000 units, each with a net asset value of $20.04. The initial day’s volume was about $100,000. This isn’t an allocation product—it’s a placeholder.

The market reaction was actually very hot. $ZEC is up more than 2,700% since the start of the year. After the ETP was listed, it surged to around $1,680 at one point (the first time since 2016), then pulled back to about $1,522, with a roughly 6.6% drop on the day. The market cap is close to $27.5 billion, ranking among the top ten crypto assets. There’s also an on-chain data point that’s even more telling: in the shielded pool there are about 4.91 million ZEC, or about 29% of total supply.

My take is this: the compliance channel for privacy coins really is opening. On August 25, the U.S. already had ZCSH listed on NYSE Arca; now Europe has followed. The path seems to be “first ETF, then ETP.” But the real signal that institutions are truly moving in isn’t that another ETP exists—it’s two things: the shielded-pool share continues to rise, and the custodian puts $ZEC into the standard service catalog. As for that 2.5% fee, in essence it’s pricing in “regulatory uncertainty.” As uncertainty declines, the fee should naturally come down. Conversely, if a year from now the product’s size is still only in the tens of millions of dollars, that would suggest institutions are buying the “privacy narrative,” not “privacy as an asset.”

Here’s a practical question for you: would you be willing to pay a 2.5% annual fee for a privacy narrative? Or would you rather hold the wallet yourself, take on custody of the private keys and compliance risk? How would you weigh the two?

#21Shares launches Europe’s first physically backed Zcash ETP