Orbit Post Sitemap
The day we've been waiting for this week: FOMC decision on Wednesday, no rate hike is the real drama
From Tuesday to Wednesday tomorrow, the October FOMC meeting will take place, with the decision announced early Wednesday Beijing time. In a Reuters survey from October 16 to 20, 94 out of 100 economists believe the Fed will hold steady this time, keeping rates at 3.75% to 4.00%, which is basically the market consensus.
So the focus on Wednesday is not whether there will be a hike, but what the chair says. After the hot CPI in September, the market once priced the probability of an October hike above 50%, but the chair’s statement of "no forecast" pushed expectations back down. The real suspense now lies in December: futures markets price the December hike at about 80%, and the Reuters survey shows all 100 economists expect a hike in December.
For the crypto world, this script is very familiar: no rate hike does not mean a shift to easing, the boot of a December hike still hangs overhead. BTC just fell back below 80,000 today, and ETH is struggling around 2,300. At times like this, every word in macro commentary is magnified.
Will you stay up late to watch the decision early Wednesday, or wait until morning to act after seeing the results?
#OctoberFOMCRateDecisionComingThisWeek
$BTC$ETH $BTC $SOL This round of market pullback superficially looks like BTC losing the key 80,000 level, but essentially it is a chain reaction of institutional fund withdrawals combined with concentrated liquidation of long leverage positions.
The capital side had already weakened in advance: On October 22, the US stock BTC spot ETF saw a net outflow of $636 million in a single day, marking the largest single-day escape since June. In just two trading days, a total of $774 million was withdrawn, completely erasing all net inflows from early October.
The market then faced concentrated liquidation: Over 213,000 accounts were forcibly liquidated within 24 hours, with total network liquidation reaching $670 million, the vast majority being long positions.
Institutional buying has faded, and the market's long positions are highly crowded. Once the critical 80,000 support is broken, it naturally triggers a chain reaction of panic selling. Clearly, ETH's decline is much stronger than BTC's, indicating that market risk appetite is contracting rapidly and sentiment is collapsing faster than the price drop.
Approaching the Federal Reserve decision on October 28, although the market generally expects no change in the current stance, there remains divergence in December rate hike expectations and uncertainty about liquidity tightening.
The macro bearish tone remains unchanged. This round of correction is not a short-term emotional sell-off but the result of the resonance of capital, leverage, and macro factors.The next ten days might be the least suitable period in the year to chase rallies. Are you ready to face a series of intense macro events? Watching the market these past two days gives a very subtle feeling—the price is still fluctuating, but the rhythm is no longer a one-sided rally; it feels more like a phase of mixed game theory and position washing. The surface sentiment doesn't look bad, but underneath, risk appetite is tightening in advance. Let's lay out the calendar first. 10/27 Israel elections, 10/28 Federal Reserve meeting combined with Google and Microsoft earnings, 10/29 European Central Bank interest rate decision, 10/30 Bank of Japan interest rate decision, 11/2 Apple earnings, 11/3 U.S. midterm elections. Almost every day in these ten days has events that can change capital sentiment, and this density alone is worth caution. What I really care about is not which news will explode, but what the market has already priced in ahead of time. Having U.S. tech earnings and the Federal Reserve on the same day means risk appetite will be priced in all at once; the ECB and BOJ decisions on consecutive days mean exchange rate and carry trade volatility will spill over into the crypto market. In this environment, BTC is usually treated as a macro risk asset rather than a safe haven, while ETH and altcoins are more likely to lose their elasticity when sentiment contracts. There is also a bullish path: if earnings exceed expectations and interest rate paths turn dovish, suppressed risk appetite could quickly rebound, with BTC recovering first, ETH following, and altcoins taking over, making the pace very fast. But the risk is that when events are so dense, capital tends to deleverage first and hold cash rather than bet early, so rebounds can easily become windows for selling after a rally. $ETH now has an unavoidable source of selling pressure: Ethereum OGs are concentrating their sell-offs.
This group of early holders is very special; they sell during bull markets and continue to cash out in batches during bear market bottoming phases. They are almost always either selling coins or preparing to sell coins.
Their holding costs are extremely low, so even if the market drops significantly, they still have huge unrealized profits. As long as there is room for a rebound, they have the motivation to cash out.
As long as this portion of low-cost chips is not fully rotated or completely washed out, there will be a continuous selling pressure hanging above. Every rally can easily be crushed by sell orders from the OGs. One of the prerequisites for $ETH to achieve a smooth new high is to fully exchange this portion of cheap chips to newly entering funds.
Of course, the shakeout is not a matter of one or two days; this process will repeatedly tug and wear down patience. Short-term upward resistance will persist, so do not easily predict a one-sided breakout.
#微软Azure增长43%，AI商业化能否提速？ #谷歌云业务增长82%，AI投入回报如何？
The boss has something to say
Alphabet, Google's parent company, released its Q3 earnings after the US market closed on October 28. Cloud business grew by 82%, but the 2026 capital expenditure guidance was raised to $195 billion to $205 billion. This is an extremely large figure.
I believe the market's focus on Google's earnings report has shifted. It's no longer just about how much cloud revenue has increased, but whether this nearly $200 billion capital expenditure can bring actual returns. AI infrastructure is a money-consuming beast. Someone in the comments pointed out very accurately that Google is intensively signing long-term contracts for nuclear power fuel cells, and electricity might replace GPUs as the core bottleneck for AI expansion. Having computing power is useless without electricity.
The market reaction was lukewarm, with GOOGL down 0.09% and xGOOGL down 0.07%; funds are mostly waiting and watching.
For crypto, tech giants continue their arms race, funds remain in tech stocks, and BTC lacks external incremental support. The main factors to watch are macro conditions, oil prices, and the FOMC.
BTC took a waterfall drop last night; now it's a question of whether the 80,000 level can hold. If it doesn't, expect a deeper correction, with resistance above at 84,000 to 85,000. Don't chase the rally or sell in panic; set your stop losses properly.
The above analysis is time-sensitive; make sure to set your stop losses. Good luck. $BTC $ETH $ZEC The US stock market is a bit exciting this week, with the Federal Reserve and tech giants rushing to report earnings. It's really hard to say who will get hit first.
📅 October 27, HOOD reports earnings first.
📅 October 28, the Federal Reserve announces its interest rate decision, followed by Microsoft and Google after hours.
📅 October 29, US GDP and PCE data are released, and Amazon and COIN also have to report their results.
This time, I'm paying more attention to Microsoft.
Not because it will definitely rise, but because Wall Street's patience with AI is clearly not as good as before.
Meta's revenue exceeded expectations, yet its stock price still fell nearly 7%. Previously, the market got excited hearing about increased AI investment; now it’s starting to ask: when will the money invested pay off?
Microsoft and Google happen to be able to answer this question.
One looks at Azure, the other at Google Cloud. If cloud business growth can't meet market expectations, just talking about AI prospects probably won't be enough.
I'll also keep an eye on COIN, since its trading revenue is directly related to crypto market activity, which is also valuable reference for our crypto community.
I'm now more inclined to watch the real reaction after the earnings reports rather than betting in advance on who will beat expectations.
If I had to pick the most worth studying this week, I'd bet on Microsoft. $GOOGL $MSFT $AMZN #微软Azure增长43%，AI商业化能否提速？ $ZEC price is around 1190, with 24-hour fluctuations roughly between -0.3% and +4%, indicating high-level oscillation.
ZEC has recently outperformed BTC significantly. BTC dropped about 1%—1.6% in 24 hours, while ZEC's weekly chart shows an increase of about 7.5%, mainly driven by expectations for the NU7 upgrade and the privacy coin narrative.
From the chart perspective, the key battleground is near $1,200; resistance above is first seen at $1,230—$1,250, with stronger resistance at $1,300—$1,320; support is near $1,180, and if broken, attention should be paid to the $1,120—$1,100 area.
Fundamentally, the core catalyst for ZEC is the NU7 upgrade likely to activate around November 5: block time is planned to shorten from 75 seconds to 25 seconds, and 60% of the fees per block will be temporarily removed from circulation until February 2031, which is expected to tighten supply somewhat. However, the upgrade will also disable the old Sprout privacy pool, and a small amount of unmigrated funds may lose spendability, representing a low-probability but real technical risk.
On the capital side, the Grayscale Zcash ETF saw a significant inflow last week, but there were net outflows for three consecutive weeks prior, indicating institutional funds are still wavering. ZEC's volatility is much higher than BTC/ETH, so short-term chasing of highs should be cautious.
#本周将公布10月FOMC利率决议 With no new projections at the Oct. 28 decision, the Fed's statement and Chair's press conference will carry unusual weight. Unchanged rates may be broadly expected, but language on inflation and December matters because Q3 GDP and September PCE arrive the next day. That sequence could keep policy expectations fluid rather than settled.
#OctFOMCDecisionThisWeek In the futures market, contract delisting does not mean there is a problem with settlement; experienced traders often first look at the exit arrangements. I see Ducat's gradual protocol shutdown this time following the same logic: business operations need to be concluded, and how assets are returned requires a separate accounting.
The announcement states that bonds and Pass rewards that have been allocated, pDUCAT in pledge or cooling period, and circulating $DUCAT in wallets can all be redeemed without waiting for the cooling period. The snapshot of token-holding addresses has also been completed. The team claims that the liquidation certificate will fully support DUCAT with vault assets, but this is a claim to be verified and should not be taken as the liquidation result.
I value the exit mechanism more than simply labeling the shutdown as negative news: waiving the cooling period reduces exit restrictions; covering multiple holding states reduces rule fragmentation. Unlike traditional futures, here we need to look at the protocol's own liquidation arrangements. Are there friends who have also transitioned from traditional markets? How would you judge whether such an exit arrangement is orderly?#Robinhood交易业务增长能否延续？ I believe the growth of Robinhood's trading business is highly likely to continue, but the form of continuation is undergoing a structural shift—the growth engine is switching from the previously high-volatility crypto assets to a combination anchored by stocks and options, with prediction markets as the incremental breakout point. However, the sustainability of this shift depends on whether a key constraint can be broken: the increase in trading activity has not yet effectively translated into simultaneous growth in new users and net deposits. 86.83 is the highest point reached at 07:00, with only 80,977 contracts traded at that peak. To break it down, at 08:00, 213,609 contracts were traded, which is 1.3 times the combined volume of the previous two peaks.
The volume disparity is clear: no one is supporting the price above, but there are real sellers pushing it down below. However, it was bought back at 83.91 for 0.66 to close at 84.57, and the 24h low of 83.4 was not breached.
So currently, this is a box with money bet on both end$BLUR, $YGG, $SAND and these altcoins were once crazy too, all in the blockchain gaming sector or NFT sector, but now they are slowly dying, gradually going to zero. Similarly, many altcoins without real income are also slowly fading away. If a project cannot achieve breakeven after several years of development and has very few real customers, some are purely a Ponzi scheme, so relying solely on the coin to harvest retail investors is inevitable, and indeed it should have been delisted early. The current crypto space really doesn't need any new projects; I even feel that one or two new projects listed on exchanges per month are already oversupplied. Even the real trading volume of blue-chip coins like BTC, ETH, and SOL in the crypto space is pitifully low. If it weren't for futures propping them up, liquidity would be very poor. The crypto space needs to rest and recuperate; if this continues, there won't be many retail investors left.Elon Musk's post has 4.75 million views and 108,000 likes, but what it talks about is very straightforward: the same company, two systems, two sets of rules.
Starlink is a civilian commercial service with strict terms that prohibit its use in warfare. Starshield is a different system, managed by the U.S. government, and its usage is decided by them.
Last week, the Ukrainian military used the civilian Starlink link to control attack drones to strike targets, and this loophole was just closed. His point is that this is not news, just a redefinition of an existing boundary.
The latter part is even more worth noting: Starlink cannot be jammed or destroyed, and if it goes down, the front line immediately collapses. At the start of the 2022 war, when other communication means were destroyed or jammed, it held strong for those months.
This sounds like a threat, but essentially it’s about pricing power. Whoever can provide something irreplaceable has the leverage, even if they repeatedly claim they do not participate in decision-making.
The same applies in the market. Truly scarce infrastructure pricing is never just about cash flow; it also depends on whether things can function without it. There is not a single industry term in this text, but the logic is the same.
To judge whether something is valuable, don’t first look at the story; first consider how many people would immediately be in trouble if it disappeared tomorrow. Last week's overall data seems to show a net inflow of ETF funds, appearing to favor the bulls, but the actual market is extremely divergent and very weak. The first three days of the week saw continuous accumulation, with a large amount of capital entering to prop up prices, but on Thursday and Friday there were consecutive large-scale exits, erasing more than half of the gains and chips accumulated in the first two days in just two days. Especially with leading institutional funds like IBIT and FBTC leading the retreat, this point best illustrates the issue: current buying is purely trial positioning, with no intention of firmly increasing long-term holdings. Although buying appears on the surface, the support is particularly fragile and cannot withstand selling pressure at all. Looking at the current candlestick structure: the price is stuck stubbornly around 80,000, oscillating repeatedly, with the 24-hour volatility range compressed very narrowly. The resistance between 81,000 and 81,300 is extremely stubborn; every time the price rebounds to this level, it is immediately hammered down with no incremental volume to support it, and the bulls have no strength to push higher. Market indicators have all flattened out, funding rates are approaching neutral, contract leverage has not expanded, and overall market sentiment is fully cautious, with no one daring to actively open directional positions. However, there is no need to be overly pessimistic; the market has not completely deteriorated. The short-term strong support lies between 79,800 and 80,000, with institutional long-term base positions still underpinning the market. 79,600 is an even more critical defensive bottom line; as long as this level holds, the overall oscillation structure will not collapse. Currently, the biggest uncertainty in the market is entirely concentrated on Federal Reserve news. At this stage, bulls and bears are evenly matched with no clear trend. All market rises and falls are contract funds washing each other out, switching back and forth between bull traps and bear traps, making this kind of market the easiest to lose money in.#SK海力士Q3财报在即，HBM4能否提振盈利？ 1. HBM4 mass production + AI long-term contract orders: HBM4 is a super high gross margin product, with long-term contracts locking prices for customers like Nvidia, with premiums far exceeding ordinary DRAM; this quarter, the shipment proportion of HBM continues to rise, which is the core driver for increasing profit margins
2. DRAM and NAND contract prices continue to rise: AI servers + ordinary PC and server demand resonate, with DRAM average prices rising significantly quarter-on-quarter
3. Product structure optimization: the proportion of high-margin HBM/high-end DRAM increases, diluting costs and raising overall net profit margin
❌ Negative factors (pressuring EPS, likely to result in No)
1. KRW to USD exchange rate fluctuations: KRW depreciation will compress USD EPS under the US-listed SKHY standard (this is the most easily overlooked risk)
2. HBM4 yield below expectations: if yield ramp-up is slow and HBM4 shipments fall short of expectations, gross margin will directly decline
3. Slowdown in spot price increases for memory, customer order cuts, or Samsung's volume increase causing price competition
4. Non-standard one-time gains are not counted as Normalized EPS, only core business profitability can be consideredNinety percent of people bet that the Federal Reserve won't move this week, but the market doesn't even have the strength to stand up.
The FOMC meets from Tuesday to Wednesday, with results released at 2 a.m. Beijing time on Thursday. The decision itself is no longer a variable.
What has changed is the latter part: a week ago, the bet was whether there would be a rate cut within the year; now it's whether there will be a hike in December. With oil prices in triple digits and sticky inflation, the option of a rate cut has been withdrawn. What’s lost is not just a rate change, but the hope that "it will loosen sooner or later."
Prices haven't broken through yet, but sentiment has already cracked. The fear and greed index dropped to 43, the lowest in three months, yet prices are still some distance from this week's lows—not because someone is selling off, but because no one is willing to buy.
A tougher challenge lies ahead: Google and Microsoft report after the market closes on Wednesday, Amazon after the market closes on Thursday, all squeezed between the decision and inflation data.
$BTC is stuck at 80,000, having touched 84,000 at its peak this week; $ETH was above 2,500 two weeks ago, but now can't hold above 2,300, with its relative price continuing downward.
The meeting only decides whether it jumps or not, not where it will jump.
#本周将公布10月FOMC利率决议 $BTC contracts fell to 79567 but bulls still hold strong: once the 79345 liquidity is lost, who would dare to treat oversold as a reversal?
What I see is not a bull counterattack, but the most familiar self-hypnosis in a weak range. BTC contracts are currently around 79567, down 0.68% since UTC midnight today, with a 24-hour low touching 79345. The price remains in a bearish moving average alignment, and the rebound has never managed to reclaim key resistance; the short-term RSI has entered oversold territory, which is precisely when people are most likely to mistakenly think "it can't fall further." But oversold is just a volatility state, not a pass to trend reversal.
What is more worrisome is that the recent funding rate is still positive, with top traders holding about 61.2% long positions by position value, and about 69.2% of all market accounts long. When most of the market is expecting a bullish candle to save positions, liquidity below 79345 is often more attractive. Open interest rose 0.31% in the past hour, but the price did not strengthen accordingly; this kind of combination looks more like increasing divergence rather than a healthy rally.
I always insist: you can choose not to be bullish, but if you are bearish, you should pay more attention to risk realization. However, this is not anyone's trading instruction. Temporary boosts from news cannot change the current structure. What bulls really need to prove is a volume-backed recovery, not to keep comforting themselves with "a rebound means reversal." Those who blindly bottom-fish, stubbornly hold longs, and fantasize that the market will inevitably recover—will they again treat 79345 as the last safety line this time? #Strategy Earnings Preview: BTC Book Gains Rebound Sharply
The Q3 book unrealized gain of about $21 billion is an actual accounting fact but is an "accounting echo" with limited signaling value;
What is more noteworthy is the capital allocation shift from "increasing BTC holdings" to "repairing capital structure."
Capital allocation shift: Q3 accumulation slowed significantly: only 334 BTC were added in the last week of the quarter, costing $28.7 million; during the same period, $176 million was spent on repurchasing STRC preferred shares, exceeding BTC purchase expenditure that week. STRC rebounded from a low of about $71 in June to about $99, showing initial success in support;
The consensus of "strong buying" indicates the market is pricing a race between Bitcoin's long-term trajectory and capital structure costs. If the coin price only fluctuates near the cost line, the book rebound is just a quarterly number swing.
1. Whether Bitcoin can stably operate above the $75,000 cost line, turning book gains from "quarterly swings" into "sustained positive spreads";
2. Whether the fund allocation between repurchasing STRC and increasing BTC holdings confirms the company’s shift from "unlimited BTC exposure increase" to "prioritizing capital structure repair."
The book rebound is real but signals are limited; the slowdown in BTC per share growth better defines the mid-term profile than the $21 billion unrealized gain. @OKX中文 @OKX星球 $BTC $xSTRC 主流链上分析（Santiment）把 ≥1000 BTC 的地址定义为BTC巨鲸（鲸鱼） - 最新统计：2044个地址 ≥1000 BTC，合计持有约717万枚BTC，占流通量接近36% 但是！地址 ≠ 真实独立的人/机构，这个是最容易踩坑的地方： 里面包含：交易所冷钱包、ETF托管地址、平台储备钱包，不是全部个人大户。比如一个交易所可能有好几个这种巨鲸地址，其实属于同一家机构。 更细的分层（另一种细分口径） - 鲨鱼：1000–5000 BTC - 鲸鱼：5000–10000 BTC（狭义whale）≈100多个地址 - 座头鲸（超级巨鲸）：＞10000 BTC，不到100个地址，这里面就是MicroStrategy、美国政府、早期巨富、大托管方这一档 两个关键误区 1. 没法精确知道真实“鲸鱼人数”：一个人可以控制几十个钱包；一个钱包也可能是很多用户的钱（交易所），链上只能数地址，分不清背后实体。 2. 不同平台门槛不一样：有些媒体宽松一点，把≥100 BTC就叫大户，这个口径数量会涨到接近2万个地址，不能混为一谈。 简单一句话总结 按圈内最常用的1The soup pot in the kitchen is bubbling, and the Q2 revenue shrank by nearly 20% year-on-year. The surface of this broth still floats with oil spots, but the bottom of the pot has already scorched. After the U.S. stock market closed on October 29, this U.S.-listed digital asset platform is set to release its Q3 earnings report. What I'm watching isn't the plating, but the stove: how far has the Deribit integration progressed, can the main pot of spot trading be reignited, and can the derivatives and USD stablecoin-related businesses add new flavors.
The biggest fear in cooking is messing up the proportions. Major coins are the main dish, altcoins are the seasoning, and leveraged contracts are the extra spicy chili—too much burns the stomach, too little leaves it bland. Q2 subscription and service revenue reached $555 million, like a broth prepped in advance that can support the base; but the nearly 20% year-on-year revenue decline indicates that front-of-house orders are shrinking, and the heat under the spot trading wok hasn't returned to a strong flame. Deribit integration isn't just about moving a new stove into the kitchen; it depends on how the heat connects, the speed of serving dishes, and risk temperature control. Whether the spoonful of chili oil from derivatives can be fragrant without being harsh is the key to the Q3 menu's success.
USD stablecoin-related business is more like a fermenting sauce—usually not in the spotlight, but without it, the whole dish lacks a layer of aftertaste; derivatives are like a quick stir-fry over high heat—right timing brings aroma, overcooking burns the pot. Now the market has also put Microsoft's token-based side dish on the table; the linkage of tech stock tokens is like the aroma from the neighboring stove—lively to smell but can't replace the home broth. If spot trading revenue continues to weaken, relying solely on subscription services as the broth to hold the table, the Q3 financial report will feel thin; if Deribit integration can simmer derivatives and stablecoin businesses into a rich sauce, that will be the new flavor poured over the main dish.
My professional judgment is straightforward: this table isn't lacking plating now, but the sound of the stove reigniting and gripping the pot bottom. A nearly 20% drop in Q2 revenue means the menu's portions were cut; $555 million in subscription and service revenue is the broth the kitchen can still hold steady; if Deribit integration only stops at changing the stove, the main pot of spot trading remains lukewarm, and no matter how fragrant the spoonful of extra spicy chili from derivatives is, it can't cover the thinning front-of-house traffic. If what is served in Q3 is still a clear, bland soup, with no fire in the main dish and no thickness in the sauce, then no matter how busy the kitchen is, it will only be arranging cold dishes more neatly. #coinbaseq3deribitimpact$BTC Everyone is watching the dump; in two days, 770 million was dumped, scaring people into thinking it was going to crash.
Actually, BlackRock quietly net bought 566 million in a single week, so overall, big money still had a net inflow of 300 million for the whole week.
The main force is not selling off; they are clearly taking advantage of retail investors' panic to scoop up bloodied chips at the bottom.
The small dip created by the dump is for smart money to accumulate; once the chips are cleaned out, it will naturally go up.
#BTC
$BTC
👇Family, $BTC BTC is now stuck near 80,000, oil prices have surged to 111, AI stocks have pulled back, and risk appetite is being suppressed hard. In October, BTC ETFs have had a net outflow of 314 million, although last week's weekly data turned positive, the funding situation hasn't truly eased. The 80,000 level—holding it means consolidation, breaking below it means emotional collapse. The FOMC is the day after tomorrow, don't lick your wounds on the blade.
$ETH ETH is even more divided: October ETF net outflows are nearly 1.18 billion USD, with BlackRock's ETHA being the main selling pressure. But on the other hand, the total staked amount on the network has hit a record, with 43 to 44 million ETH locked, accounting for 36% of circulation. BitMine itself has staked 86% of its position. ETFs that can sell are running, chips that can't sell are locked up. 2,300-2,330 is ETH's psychological bottom line; holding daily closes above it means a false sell-off, breaking below it means a trend.
In short: BTC waits for the FOMC, ETH waits for the closing result at 2,300. Staying alive means having the next chance. #Circle财报在即，稳定币增长能否兑现收益？ Small dip is not a drop XRP stuck at 1.29, whales quietly withdrawing; DOGE dropped to 0.076, technicals only show "strong sell"; the market shakeout is still hitting those who mistake "a big drop" for "almost bottomed out".
XRP current price is about $1.29, down 1.24% in 24 hours. The key is not the price but on-chain data: the number of wallets holding 10 million to 100 million XRP dropped from 351 to 339, with 12 big holders exiting; meanwhile, XRP balance on exchanges increased by 50 million, supply is piling up on the order books. 1.26 is the recent defense line; if it holds, it may consolidate between 1.30 and 1.43; if it breaks, the next support is at 1.11, then 0.98. Whales are withdrawing, retail investors are buying; guessing the bottom in this structure requires more than just courage.
DOGE current price is about $0.0762, down nearly 2%. Technical indicators collectively signal a strong sell, down 4.39% in 7 days, and 21.22% in a month. Its problem has never been "has it dropped enough," but rather the lack of new buying reasons. If meme coin rebounds rely only on emotional pulses, every rally just opens the door for those ahead.AI investment is currently in a historic-level super capital expenditure cycle, likely to "hold up" in the short to medium term, but facing real ROI verification pressure in the long term (post-2028); Nvidia's moat remains wide but is shifting from a "chip monopoly" to a "full-stack + software ecosystem + financial leverage" model, which is not invincible. 1. How much longer can AI investment hold? Scale, momentum, and bubble risk The current investment scale is already extremely large. In 2026, the combined capital expenditure of major hyperscale cloud providers (Amazon, Alphabet, Microsoft, Meta, Oracle, etc.) and leading CSPs is expected to be in the range of $700 billion to $950 billion, with some institutions estimating the global top nine CSPs at about $957 billion; in 2027, it will further jump to $1.2–1.58 trillion or even higher. This is close to or exceeds the relative GDP share of historical cycles such as railroads and the internet, and some analysts call it "one of the largest capital projects in U.S. history." Driving forces for continuation: Demand side remains supply constrained. Nvidia's management clearly states that demand far exceeds their supply growth of 70%, with new platform orders like Vera Rubin covering major hyperscale cloud providers and AI clouds. Jensen Huang emphasizes "compute is revenue," and tokens have become "productive and profitable." Agentic AI and physical AI significantly increase computational intensity.The deadliest move on the chessboard is never the opponent sacrificing the queen, but rather realizing after their sacrifice that your king's flank has nowhere left to hide.
On October 28, Microsoft unveiled its new season's report card, with Azure posting nearly 40% year-over-year growth last quarter—that's not just a set of operating figures, it's an overt first move. Grandmasters never focus on where the current move lands; they look at whether it opens up the next twenty moves.
AI capital expenditure is a thoroughly positional sacrifice. Sacrificing a pawn or a knight is never about the piece itself, but about space, lines, and initiative. Yet chess logic is ruthless: the compensation for the sacrifice must be realized; if not, the sacrifice is merely a wasted loss of forces. Copilot's monetization is that compensation line. If it fails, the pawn structure will bear irreparable cracks, and as the midgame turns to the endgame, every isolated pawn will be picked off one by one.
So the real test isn't whether the 40% figure can be repeated, but whether it accelerates. A single first move doesn't constitute an advantage; only a series of consecutive first moves does. Once the tempo is lost, the position flips from "I'm attacking" to "I'm defending," and every defensive move demands extra time cost—once time pressure sets in, even the best piece positions will falter.
Looking at $xMSTR's journey, it resembles a secondary variation hanging off the main line: if the main line goes astray, even the most exquisite tactics in the secondary variation lose their foundation. The funds on the board are the most sensitive light pieces, often detecting the opponent's responses earlier than the main board. Once credit spreads loosen, the financing lines of this grand AI game are constrained from the flanks; no matter how beautifully you arrange your pieces, a line from the opponent always hangs overhead.
Some ask how I calculate twenty moves ahead. It's not about predicting all twenty moves accurately, but about shutting down the opponent's three fiercest counterattack lines and leaving the rest to structure. Currently, the AI narrative is stuck in the most frustrating grid—piece advantage, open space, ample time, yet no winning move can be found. The investment is a real sacrifice of gold and silver, but the returns must be verified layer by layer; the valuation piece will sooner or later be cashed out in the endgame.
October 28 was not a decisive move; it was merely a response following the book. The true judgment boils down to one sentence: the sacrifice has been made, but is the initiative still in your hands? #microsoftq1azureandai$BTC BTC dropped from 84,500 to 79,600, and the more the price falls, the more people go long on OKX.
The long-short ratio of BTC contract accounts rose from 1.21 on 10/21 to 2.19 today, with long accounts now more than twice the short ones.
It has dropped nearly 6%, yet people are still bottom-fishing all the way.
The problem is, the more longs pile up, the price still can't rise, indicating that the bottom-fishing funds haven't been able to lift the market. The key focus below is the previous low at 79,168; once this is broken, it’s likely another batch of longs won’t be able to hold on.Trump is threatening another "epic rage operation," and Iran is reported to have escalated assassination plots.
This round of correction in the US stock market may not be over yet.
The most troublesome combination:
Rising conflicts + soaring oil prices + a sharp drop in gold.
Trump says "the war is almost over," but the market doesn't believe it at all—gold futures fell below $4000 intraday, and Brent crude oil stood at $116.
The real pressure is yet to come.
The longer high oil prices drag on, the harder it is for companies.
Rising transportation and production costs squeeze profits, making inflation harder to reduce. Even if US Treasury yields fall due to safe-haven buying, it doesn't mean the Fed will cut rates sooner, nor does it mean corporate financing will get cheaper.
AI tech stocks have their own accounts to settle.
Investment keeps increasing, but when will revenue and cash flow catch up? The market has already started to ask.
Adding energy costs and repeated geopolitical tensions, high valuations are even harder to sustain.
The key variable is:
If negotiations make progress and oil prices fall significantly, the judgment must change; but if fighting suddenly breaks out before the midterm elections, the market could plunge sharply.
Trump says the war is almost over while simultaneously threatening to fight.
The costs companies have to pay and the money AI needs to spend won't decrease just because of a statement.
At present, there is no sufficient reason to believe this round of correction is over.有很多人说现在是牛市来了，现在确实是已经是熊市的末期，目前第一波上穿站上50周均线，是形成了牛熊的转换，不过真正的转化成牛市，也会震荡一年多，没那么容易马上正真走出大牛市，第一，目前宏观市场，美债依然是加息状态，大方向还没有正真的利好，大牛市是需要多股资金合力推动的，单一的拉升无法走出真正的大牛市，现在只能确认是反弹，目前80000点，76000点聚集了很多吸货筹码，大资金想涨，也不会给散户抬轿子，会在这个位置不停震荡，随即往下砸，有些人会说越跌越买，除非你是现货做大周期长线，否则，市场有的是办法让你害怕，让你失去耐心，仅作为个人观点，不作为参考A $40 billion GPU debt financing is like pouring a massive raft foundation on quicksand, but with only one load-bearing column under vertical stress. SpaceX anchors its cash flow with Starlink and launch services; Musk is betting on the long-term load curve of the AI computing power skyscraper. But what really sends chills down my spine is his warning—that for over 200 emerging AI cloud companies without a second source of income, the debt market has already shut the credit gate.
In my view, this is a classic structural break between general contractors and subcontractors. The computing power leasing business looks like standard floor-by-floor replication on the surface, with costs amortizable and cash flow predictable. But if you flip to the structural cross-section of the blueprint, you’ll see: these companies’ foundations are all rented GPUs, and the load-bearing walls are single-client contracts. Once financing costs rise and risk control tightens, the whole building lacks even lateral stiffness and will enter plastic deformation at the slightest wind. Musk can get funding not because his design is prettier, but because he holds two completed and rent-collecting cash flow pipelines—Starlink and launches. This is true structural redundancy, a dual-core tube plus ring truss system with a complete lateral force resistance system.
As for those projects living solely on computing power rent, I’ve done too many similar real estate due diligences: gorgeous blueprints, attractive renderings, but actual foundation bearing capacity insufficient, rebar ratios cut corners, shear walls eccentrically arranged. When credit tightens, they don’t collapse but get acquired and restructured—known in the industry as unfinished buildings taken over by top developers, who revise plans, lower standards, and reapply for permits. Out of over 200, only a handful will retain independent structural systems. Consolidation isn’t bad; it’s structural optimization, removing redundant components and clarifying the main load paths.
Returning to the linkage with US stock token targets, the logic holds. The credit spread on computing power debt is widening, indicating the market is reassessing the seismic rating of the entire AI infrastructure. Those relying solely on external financing without self-operated income anchors are being deformed under deleveraging load; meanwhile, targets with cash flow foundations reveal true structural value amid bubble squeezing. This isn’t a valuation issue; it’s a load-bearing capacity issue.
I’ve seen too many projects die during construction, not design. Anyone can draw blueprints, but very few can drive foundations to bearing strata, cure concrete to design strength, and ensure sufficient rebar anchorage length at every node. The same applies to computing power expansion: financing is pouring, income is curing, and without proper curing, even the tallest building is a dangerous building.
The market is conducting a static load test; whoever’s structure buckles first will be forced out. #spacex40bgpufinancing#SK海力士Q3财报在即，HBM4能否提振盈利？
I roughly looked into it, and the hottest topic is still HBM4. Bulls believe the demand for AI servers is solid enough that the volume of HBM4 can directly explode profits, and it can also offset the weakness in traditional storage. Some big names in the community have posted that profits will hit a record high. As Nvidia's favored partner, SK Hynix's good performance is almost a given. But here's the problem: the market expectations are already so high that no matter how much profit is made, if it doesn't "greatly exceed expectations," could the earnings report actually become a case of all the good news being priced in and cause a sell-off? Also, the weakness in the traditional storage business is real, so whether HBM4 alone can fully cover it depends on the specific data tomorrow night.The isobars in the Strait of Hormuz have just relaxed by one notch, yet the radar echoes over eastern Saudi Arabia still show three isolated strong convective cells — this is the initial field this morning: oil flows are recovering, but the threat has not dissipated.
There is an old saying: isolated cells can be bypassed, but the shear line cannot. Missiles are isolated convective cells; if intercepted successfully, it's just a short shower; but the oil pipelines, shipping schedules, and insurance costs form a shear line across the Persian Gulf. As long as it remains, any single outbreak will feed unstable energy back into the temperature field. Saudi Arabia has confirmed interceptions, but damage is unverified — in our terms, this is called "echo intensity uncalibrated by ground stations," which can be strong or just ground clutter, but it cannot be ignored during scheduling.
Brent is at 116.72, WTI at 103.27; these are not clear-sky prices but the sultry heat after a thunderstorm passes. Around 117 is already a dense band of isobars; pushing up one notch requires new energy injection, while the space for relaxation downward depends on when the moisture withdraws. Strait flow has improved, visibility has risen by one to two nautical miles; however, tanker rates and security costs — two humidity indicators — remain above 90%, and as long as moisture lingers, temperatures won't drop. The so-called "risk mitigation easing prices" essentially ask when the dew point will fall along with the temperature. So far, only the temperature has dropped slightly.
Trump said the conflict might end soon, Iran is opening its diplomatic doors, but no agreement is finalized — this set of signals is called model oscillation in numerical forecasting. Ensemble members at different times diverge greatly; it is taboo to use a single deterministic forecast to schedule the next seventy-two hours. Conclusions must wait for sounding data, not statements.
Switching perspective to the US stock market proxy. Its price behavior resembles coastal city visibility: oil flow on the sea surface recovers, visibility rises from 500 meters to two kilometers, but the near-surface inversion remains unbroken, and the fog ceiling still presses down. Risk appetite warms only in the lower layers, not reaching high altitudes. High oil prices create a persistent downdraft on transportation, insurance, and manufacturing costs, suppressing valuation ceilings; when geopolitical expectations ease, a warm advection arrives. Cold and warm air masses confront each other at the same grid point, resulting in sideways trading with large volatility — wind direction is uncertain, gusts are strong.
Judgment: this is not continuous clear weather under a high-pressure ridge but a brief calm inside a typhoon eye. Subsidence in the eye area and clearing skies are easily misread as "clearing up"; the eyewall remains on the periphery, and the next convective round only lacks a trigger. Flow recovery does not equal supply chain repair, easing rhetoric does not equal agreement signing, visibility improvement does not equal inversion removal. In ensemble forecasts, the probability of high volatility remains above 70% for the next three time steps, and the pressure gradient has not substantially converged.
First look at the sounding, then talk about the weather. #OilFlowsRecover For this opening position, I want to explain to newcomers the difference between "position amount" and "how much loss to prepare for." Core Trader 15 opened a long SNDK position with 10x leverage, nominal value about $35,900. This is just the initial position; the outcome is still unknown.
📌 My judgment is: when trading contracts, you should first set a loss budget, then work backward to determine the position size. No matter how eye-catching the amount in the monitor is, it can't answer whether this risk is appropriate because the data doesn't provide the account equity or stop-loss arrangement.
Let's do a simplified calculation based on this nominal value: with 10x leverage and ignoring fees, the initial margin is about $3,587; if the price moves against you by 2%, the unrealized loss is about $717, which is 20% of this margin. Leverage reduces the capital needed to enter but does not reduce the price risk of the same position size.
When I teach newcomers, I repeatedly remind them: margin is the capital occupied, not the preset stop-loss amount. The most common misconception is thinking "only about three thousand dollars used," which underestimates the volatility of the entire position. If your loss budget can't withstand normal fluctuations, placing the stop-loss too close to the entry price may cause it to trigger more easily.
Newcomers, if you are only willing to bear about $350 loss and plan to leave a 2% stop-loss space, how much nominal position would you control? If you get stuck at any calculation step, leave a message, and I'll help you break it down.This morning while scrolling through my phone, I saw a data post from Digital Today saying that the Dogecoin spot ETF had a net outflow of $538,000 last week, marking the largest single-week outflow since mid-August, with institutional participation nearly stalled. Looking at the 0.07639 price, I originally wanted to just close the app, but my finger slipped and I flipped to another on-chain snapshot.
In the past few days, large wallets holding over 100 million coins quietly added about 600 million $DOGE around the 0.08 mark, pushing total holdings directly to 56.21 billion coins. On one side, institutions going through brokerage channels are redeeming, while on the other, large on-chain holders are quietly accumulating.
What really caught my attention was that the long-short ratio of top traders' accounts on Binance stubbornly stayed around 2.64, not dropping at all. The price was pressed down from 0.0755 to 0.0764 to form a bottom, but the derivatives open interest (OI) actually increased by another 1.5% compared to a few days ago. Retail investors are hesitating, but the big players haven't let go. $DOGE's current shakeout hasn't scared off the smart money at all. Don't be fooled! Is this a last gasp or a real reversal? $ERA
era's recent surge looks fierce, rising from 0.071 to around 0.088, but those in the know understand that the first reaction after a sharp drop is often a setup for bulls to escape.
The 0.0900-0.0920 range above is an extremely heavy trapped position area; the previous decline started from here. This zone squeezes a large number of sell orders from those wanting to break even. Also, there is volume divergence: the half-hour trading volume is only 20% higher than the average of the previous three hours. Volume still hasn't appeared. A volume-less rise is just playing tricks. Without sustained expanding volume support, this kind of V-shaped reversal is unlikely to directly break through the strong resistance at 0.09.
Looking at on-chain large-scale capital outflows, despite a 20% price rebound, there is a short-term net outflow of 300,000. What does this indicate? It means that during the price increase, whales and main players are crazily distributing their chips while retail investors chase the rally. $BTC
Since the main players are running, why would you want to catch the falling knife? Chasing highs now is like grabbing fire. Lao Qi's strategy: the current rise is repairing indicators, not starting a new trend. If it gets resisted and falls back near 0.09, this is an excellent short entry point. You can gradually open short positions in the 0.088-0.09 range, set stop loss above 0.0925, and target a drop back to 0.08-0.075. $SOL #中东原油运输恢复，油价为何仍居高位？ The low touched 79,277, buying back above 80,000 but without continuation. Volume shrank from 53 million to 48 million, while open interest actually rose from 93,300 contracts back to 94,100 contracts. Price remains steady, positions are increasing, both longs and shorts are heavily betting around 80,000. There is a sell wall of about $8.53 million at 82,000, and Binance's sell orders at 82,500 are even thicker, around 33 million. Every step up in the rebound hits pre-arranged supply. Below, there are large buy orders at 79,000, but these orders represent intent, not commitment; when closing below, they withdraw faster than anyone else. At a deeper level is cost: corporate treasuries around 80,200, ETF holders' cost between 80,000–81,500. This is not technical support, but a collective stance of real money investors building positions here. If the price stays below for another day, they shift from break-even to unrealized losses, and their behavior becomes the variable. Macro conditions have not confirmed easing. AI credit pressure is rising, CoreWeave dropped 15.5% in one week; $BTC moves in line with risk appetite. On the 22nd, ETFs saw a single-day outflow of $635 million, with IBIT and FBTC withdrawing 254 million and 217 million respectively. Institutions are not bearish, but when risk budgets tighten, they exit first through the most liquid channels. Rules: do not guess direction between 79,000–80,000; only reassess if closing above 80,500 with volume support; single trade stop loss ≤ 2% of total position. The question is not where the bottom is, but under what conditions the cost base below 80,000 shifts from “holding” to “selling.” In this stalemate, the final move will be a short squeeze upwards.South Korea crypto tax: delayed again! 😮💨
Basically decided! South Korea's virtual asset income tax, originally scheduled to start on January 1, 2027, is now very likely to be postponed for the fourth time. The presidential office has indicated: discussions are "moving towards a delay" ✅.
How long will the delay be? The mainstream tendency is to postpone for about two more years. The US will only join the OECD "Crypto-Asset Reporting Framework" in 2029, so South Korea's actual tax implementation will likely be postponed to 2029. Some ruling party lawmakers have already proposed an amendment to set the date at 2029. The final decision still depends on the November National Assembly tax subcommittee or negotiations between the presidential office and the ruling party 📅.
Why the repeated delays? First, concerns about unfair tax burdens: before the US joins CARF, South Korea cannot obtain data on Korean users from US exchanges, so overseas exchange users might "slip through the cracks" 😅. Second, political pressure: many young investors, over 50,000 have petitioned, causing significant backlash 🙅♂️. Officials also say the tax infrastructure is insufficient, but analysts think this reason is weak.
$ZEC $ETH $IRYS
The core of the tax system remains unchanged: income from transferring or lending virtual assets is taxable; an annual tax exemption of 2.5 million KRW (about 1890 USD); a tax rate of 22%, including local income tax 💰.
In short: the direction to delay is clear, but the final date still needs confirmation by the National Assembly, so uncertainties remain. Don't rush to calculate taxes yet; wait for November 👀. #本周将公布10月FOMC利率决议 #Robinhood交易业务增长能否延续？ #BTC现货ETF资金反复，市场买盘能否企稳？
Taking a look at the ETF comparison of $BTC, $ETH, and $ZEC, I suddenly realized that Ethereum has already been shunned by institutional funds.
Ethereum ETFs are bleeding, but there is one coin quietly being picked up by funds.
Last week, Grayscale's ZEC ETF saw a net inflow of $41.1 million, marking the strongest weekly inflow since September. Just the week before, it was still experiencing a continuous three-week net outflow totaling $143 million.
What’s even more striking: ZEC is rising, BTC is fluctuating, ETH is flowing out, and funds are changing direction. $YFI
YFI's story is scarcity. It had a famously fair launch and a tiny supply that holders love to point at. But scarcity without growing vault usage is just a small number. Yield aggregation is crowded now, so fee generation and deposit trends matter more than the supply narrative ever will.Why is $BTC continuously falling? Don't just focus on the candlestick charts; the real variables lie in the macro environment!
Continuing from the previous discussion, what the market needs to pay attention to recently is not just whether BTC can hold 78,000, but two things: the policy uncertainty brought by the US midterm elections and the Federal Reserve's upcoming interest rate signals.
First, the midterm elections. The impact of this event on the crypto space is not about whether prices rise or fall on election day, but whether subsequent policy expectations will change. The Trump administration's relatively crypto-friendly policy stance may be reassessed by the market if the congressional power structure changes.
Next, interest rates. What really needs to be watched is not just whether there will be a rate hike this time, but what signals the Federal Reserve will release going forward.
The market usually trades on expectations in advance. If officials speak hawkishly, emphasizing that inflation pressure remains, the market may reassess the future interest rate path, and high-risk assets could be hit; conversely, if policy statements are not as tough as expected, it could give the market some breathing room.
Putting these two factors together, I remain cautiously bearish on BTC in the short term. 78,000 is the first key level to watch; if it breaks, focus on 75,000, and in extreme cases, look at 72,000. But these are risk scenarios based on conditions, not a premature assertion that the price will definitely reach these levels. #本周将公布10月FOMC利率决议 🔥Brothers, BTC and ETH might be brewing a critical turning point, but the real opportunity is not impulsive bottom-fishing, it's understanding the chips and the trend!
📉BTC previously pulled back near 79,500 before rebounding, oscillating steadily between 80,000 and 86,000, with 30-day volatility shrinking to about 6%. When volatility keeps narrowing, it often means the market is gathering strength to choose a direction, but whether it goes up or down still needs price confirmation.
📈Some traders see the 80,000–81,000 range as a long-term buy zone, targeting 94,000. If ETF funds keep flowing in and whales continue accumulating, it could indeed support the market, but funding data must be continuously verified, not just based on a single week's performance.
🔒ETH is also worth watching; around 2,300 is an important support zone. Rising staking ratios and declining exchange balances may reduce short-term tradable supply. However, supply contraction does not necessarily mean prices will rise; demand is the key.
⚠️My approach: observe in batches, control position size, consider following up only after breakout confirmation, and adjust promptly if key support breaks.
Do you think this is a bull market buildup or the calm before the storm? Let's chat in the comments! #本周将公布10月FOMC利率决议 #SK海力士Q3财报在即，HBM4能否提振盈利？ The Korean KOSPI's 4x is the forward PE (calculated using future expected profits), not the trailing PE, and it's an extreme figure hammered out by memory stocks.
Korean KOSPI forward PE ≈ 4x (Samsung + Hynix have very high weights!)
⚠️ Key point: It cannot be simply understood as "Korea is cheaper, guaranteed profit"; the discount comes with a cost.
1. Industry is extremely concentrated: Nearly half the KOSPI weight is tied to memory chips, which are highly cyclical products, with profits that can soar one year and plunge the next. A 4x PE means the market is pricing in a significant drop in memory profits next year, just as you saw, Wall Street can hit a round of selling anytime based on news. The cheapness is because risk pricing is fully factored in.
2. There is an inherent "Korean discount": Chaebol governance, dividend and buyback willingness, geopolitical position caught between China and the US, free flow of foreign capital, all require higher risk compensation for funds, so long-term valuations are naturally suppressed and not guaranteed to recover after a drop.
In summary:
Numerically, Korean stocks are much cheaper, but this cheapness already includes huge risks from geopolitics, cycles, and foreign capital withdrawal. This weekend, Kongshen did something that felt like "slapping himself in the face" again: he reduced most of his long positions. The image below shows what was left after the reductions.
Some are already ready to screenshot and shout, "Weren't you bullish on this wave? Why are you running away?" Don't rush. I haven't changed the direction, only the position size. Next Tuesday is Hynix's earnings report, Thursday has the Fed meeting combined with core PCE — a whole lineup of binary events in between. This kind of market is the worst to face with heavy positions and high leverage.
It's the same with playing cards: no matter how good your hand is, you wouldn't push all your chips in before all the information is revealed. Reducing to a level where you can sleep well, keeping bullets ready to decide after the results come out, is much better than stubbornly holding onto an exam paper with unknown answers.
What about you this weekend, did you add or reduce?$JTO
Jito is really a bet on Solana's MEV economy. Validators earn extra through tips, and JTO sits on top as the governance layer for that machinery. The catch is obvious, though: if Solana activity cools, tip revenue thins out too. Is JTO a clean proxy for Solana network usage, or is governance alone too thin a story?$CVX
Convex only makes sense in the context of Curve. It pooled vote-locking power so users didn't have to lock CRV themselves, which put it at the center of the gauge-voting and bribe game. When that game cools, CVX feels it first. Can a second-layer governance token stay relevant when its base protocol's incentive race slows?#现货黄金跌破4000美元关口
Gold price breaks below the $4000 mark, can this wave really hit bottom and rebound?
Spot gold rebounded to around $3998. On October 28, the probability of a Fed rate hike is extremely low. If dovish guidance is delivered, a short-term pullback to $4300-$4500 is not difficult.
Additionally, with commodities generally under pressure and Brent crude oil retreating to oscillate near the $114 high, the market's risk aversion and inflation-hedging logic are intensely intertwined.
▶️ Liquidity squeeze risk
Although crude oil has pulled back in the short term, it still stands at the $114 high. If geopolitical backlash pushes oil prices higher, triggering secondary inflation anxiety, gold may not gain risk premium but could instead repeat concentrated sell-offs under liquidity tightening.
▶️ The dual nature of a lame-duck government
If a lame-duck government emerges after the election, a short-term shutdown could indeed be bullish for gold. But in the medium to long term, tax cuts being blocked and a slowdown in AI and traditional energy subsidies will slow debt expansion, indirectly strengthening the US dollar's credit and suppressing gold's long bull run.
▶️ Volatility cap suppression
Currently, option IV has dropped to about 20%, and this reduced volatility acts like a cap on the rebound, making a sharp surge unlikely.
Looking ahead, in the short term, gold prices are expected to test $4300-$4500, supported by rate discussion and election-related risk adjustments. However, before clear signals of rate cuts emerge, the conditions for a major bull market are not yet met, so caution is advised on any rally.
$XAUT $XAU $BZ
DYOR #SK海力士Q3财报在即，HBM4能否提振盈利？ Compared to historical memory cycles, this round indeed shows three new changes: First, multi-year LTAs combined with prepayment/deposit mechanisms transfer some expansion risks to customers, reducing the intensity of the classic "short-term price — capacity — surplus" transmission chain ODAILY; second, the competitive threshold for HBM extends from the chip itself to system architecture, 3D stacking, and advanced packaging, with customer co-development increasing replacement costs; third, demand has expanded from training to inference and enterprise-level storage, and AI model efficiency improvements may actually expand application scale and further boost infrastructure demand SK hynix Newsroom ODAILY.
However, risks similar to historical cycles have not disappeared: profits still heavily depend on contract price trends, market share is being eroded by competitors, massive capital expenditures will create rigid depreciation over the next two years, the company has not provided quarterly revenue and profit guidance, and growth pace still depends on three variables: DRAM/NAND prices, HBM4 ramp-up, and the proportion of high value-added products ODAILY. Third-party analysts expect Q3 revenue of about 98.2 trillion KRW and operating profit of about 7.7 trillion KRW; these figures represent market consensus expectations rather than company guidance and can serve as a reference benchmark for the next round of validation Caixin.#SK海力士Q3财报在即，HBM4能否提振盈利？ **At the market share level, the leading advantage is narrowing.** Counterpoint data shows the company's HBM revenue share dropped from 64% in 2025Q2 and 58% in 2026Q1 to 50% in 2026Q2; during the same period, Samsung's share rose from 21% to 33%, narrowing the gap from 37 percentage points to 17 percentage points. Samsung's HBM4 yield rate has risen to about 80% (Electronic Engineering Album Sina Finance).
**At the customer level, concentration and bargaining power risks coexist.** Hynix has long been Nvidia's main HBM supplier, historically accounting for 70% to 90% of Nvidia's procurement by purchase amount, currently estimated to have dropped to just over 50%; AMD is mainly supplied by Samsung and secondarily by Micron, with no effective supply evidence from Hynix yet (36Kr). Revenue depends on the procurement rhythm of a single major customer. Once the customer introduces a second supplier to press prices, both market share and gross margin will be under pressure. LTA price locks sacrifice spot price flexibility during price increases and may insufficiently buffer during price declines due to price terms favoring customers, making it a double-edged sword. Lying in bed unable to sleep, I went over this trade in my mind.
$PEPE
It has been fluctuating downward all the way, with a volume-reduced rebound and continuously lower lows; the sellers are indeed pushing. It even broke 4.3, I admit that.
But that's all—0.000000084, a 1.95% buffer, basically still standing at the edge, not yet in the safe zone. Meme coins have much bigger daily emotional swings than this.
So I have only two clear thoughts: Has it broken 4.2? No, still 2.33% short. Has it reclaimed 4.384? No. Neither happened, so do nothing.
I already pushed the stop-loss early; the worst case is a wasted effort, no loss, so I don't bother watching. How much profit or loss is up to it; how much I can take away is up to me. Put the phone down, sleep for two hours. $BTC $ETH
#10月FOMC临近，美联储暂停加息还有哪些变数？ #SK海力士Q3财报在即，HBM4能否提振盈利？ Seasonal storage mixed ASP rose about 30% quarter-on-quarter, while shipments only increased by a high single digit quarter-on-quarter — the primary driver for revenue and gross profit improvement is price increases, not volume expansion ODAILY. This is a typical late-cycle divergence between volume and price: prices push profits to extremes, while volume growth has already moderated. The company's own guidance for Q3 is also cautious: DRAM shipments about +10% quarter-on-quarter,
There are three layers to the bearish logic: first, the storage industry historically has no precedent for maintaining ultra-high gross margins long-term; the jump from 41% (same period last year) to 76% operating profit margin itself proves the cyclical volatility; second, starting in 2027, new capacities such as Cheongju M15X, Yongin Fab1, P&T7, M17 will be successively released, combined with Samsung and Micron's HBM4 capacity expansions and mainland Chinese manufacturers catching up in general DRAM/NAND, the supply side will turn loose within two years SK hynix Newsroom ODAILY; third, this round of demand is anchored on AI capital expenditures by leading tech companies, and if the commercialization returns of large models fall short of expectations causing capex marginal slowdown, customers may have overstocked in advance, overdrawing future demand ODAILY. After the earnings release, the stock price first rose 4%, then turned down over 11%, and the Korean and Japanese stocks had already fallen 14.65% before the earnings, indicating the market has begun pricing in the divergence over "whether peak profit margins can be sustained" OD