Orbit Post Sitemap
Three head chefs sold the secret recipes from the kitchen by the pound to the neighboring street at the restaurant's peak popularity — that's my first reaction to watching CoreWeave's drama. Since going public, the co-founders have sold about $3.3 billion worth of shares, pulling out $2.4 billion this year alone. They claim it's to prepare ingredients in advance and operate transparently, yet they still hold onto 20% of the base recipe. But the front-of-house customers aren't fools: within a week, the stock price was cut by a quarter, and long-term bond yields soared past 15 points. This is no longer a plating issue; the stove is emitting black smoke.
I have a strict rule when cooking: never dump all the side dishes into the pot before the main course is fully cooked. The AI cloud business right now is like a huge tomahawk steak — crispy on the outside, but when cut open, the blood hasn't been fully drained. The order backlog is that beautiful crust, but cash flow is the core temperature of the meat. QCP smells the burnt scent of credit risk, and Bank of America warns that expanding too fast will turn the aftertaste bitter — it's like having twenty pots on the stove at once; no matter how strong the fire, the broth can't handle being short on fuel for each pot. Huang is still saying long-term AI demand looks good, like the head chef patting you on the shoulder saying "this dish has a future market," but the oil temperature in your pot is already fluctuating wildly.
When it comes to portfolio allocation, you have to think with your tongue. Mainstream coins are the main course — filling, satisfying hunger, and durable through turnover; altcoins are the seasoning — sprinkle a bit to enhance aroma, but half a jar can overpower all the original flavors; leveraged contracts are the extra spicy chili — too much burns your stomach, but just right warms you up. The AI cloud ingredient now smells like truffle, but when cut open, the texture hides an industrial-level spiciness. Bond yields over 15 points mean that spoonful of chili oil on the books has already burnt to the edge of the pot. Do you think it won't affect the neighboring AI cloud kitchen? Supply chain anxiety is never a single-store problem; the entire food street's stocking rhythm will tremble.
At the end of the day, founders selling shares is reducing portions, stock price drops are customers returning dishes, and widening credit spreads mean suppliers are demanding cash on delivery. When these three signals appear together, the recipe needs to be rebalanced. The biggest fear in the kitchen isn't too much salt, but the side dishes not being fully cooked before the appetizer chef runs off — and now the cutting board is still hot. The recipe isn't finished, and the stove is still burning. #coreweave3.3bsharesales"Countdown 3 days: SOL swaps to a 150-millisecond heart⚡️"
The hardest facts first: Around 22:50 Wednesday night (Beijing time), epoch 1064 begins, SOL swaps its consensus heart — finality slashed from 12.8 seconds to about 150 milliseconds.
Not a one-click process: The old TowerBFT runs another 5000 slots (about 18 minutes) handing over to Votor; validators pay a 0.8 SOL ticket, Firedancer absent, 82% stake ready to launch. Votes no longer go on-chain, don't interpret the transaction count crash on activation day as bearish.
Three days before the upgrade, SOL hovered at 103.97, down about 13.9% since October 6, 100 is the bull-bear line, positions cut 2.96% fastest — bulls get shaken out first.
BTC is the most stable: 81800 up 0.29%, positions 52.2 billion down 0.39%. ETH stuck at 2400: 2419 down 0.33%, 71.43 million contracts with 87.9% longs, retail long-short ratio 3.08 — stubborn bulls. Next week CPI, initial claims, PMI queue for release.
My personal view:
① Don't chase SOL: 100 is the life-or-death line, before upgrade it's waiting for the reveal, 103.97 no chasing longs or shorts;
② Watch two traps on activation day: don't treat the transaction count crash as bearish, no leverage moves 18 minutes before and after handover 🎯;
③ Use BTC as anchor: above 81800 low longs look to 82500, break 81200 run, currently I'm losing money 💰
Are you betting on upgrade bullish realization or betting the good news is priced in and will dump? See you in the comments 👀$BTC Macro: Conflicting data, the market is pricing in "stagflation," but most people are focusing on the wrong variables
US September nonfarm payrolls increased by only 29,000, and the probability of a rate hike in October dropped from 36% to 17% within a week; however, CPI year-over-year is 3.7%, PPI surged to 6.3%, and retail sales month-over-month rose 0.8%, far exceeding expectations. Employment is weak, inflation is hot, and the Federal Reserve is caught in the middle, with the market oscillating accordingly. Most people are still using the old framework of "rate hike = negative, no rate hike = positive" for BTC, but the real pricing variable now is not whether rates rise, but how long real interest rates can hold. Oil prices contributed half of the monthly CPI increase, and this part could fall at any time; however, retail spending was revised up to 1.3% in August and remains strong, indicating that consumer demand has not collapsed. If inflation falls slower than employment deteriorates, the Fed will be forced to choose between "protecting employment" and "controlling inflation," and whichever side it chooses, the loss of fiat currency credit is the long-term logic for crypto assets. In the short term, don't guess the direction; watching the marginal changes in real interest rates is more useful than watching the probability of rate hikes. Long-term holding is possible, but don't treat $STRK as a belief
STRK is the native token of Starknet, belonging to the Ethereum ZK-Rollup Layer 2 track. It's not completely without chance, but it's also not suitable to be considered a "buy and hold to win" asset. More accurately, STRK currently belongs to a high-risk watchlist: the technical narrative still exists, the ecosystem is recovering, but the unlocking pressure, user growth, and competitive landscape have not truly reversed.
🏆 From a fundamental perspective, STRK's biggest advantage is still its technical route.
Starknet uses STARK proofs, and by 2026 it has launched the STRK20 privacy framework and Shadow Accounts, with transaction fees gradually shifting to payment in STRK. After the v0.14.3 upgrade, if more transactions really use STRK to pay gas in the future, the token demand will be more genuine than now. There have also been discussions in the community about an independent Layer 1 and quantum-resistant routes; although these governance topics cannot immediately translate into performance, they do bring periodic sentiment fluctuations.
Ecosystem data is also slowly recovering. TVL, DEX trading volume, and stablecoin scale have all improved, indicating that capital and trading activity have not completely died out. On-chain transaction volume once returned to a relatively high level, and DEX trading volume noticeably expanded during certain periods. However, these improvements are still insufficient to indicate a full ecosystem recovery.A mine exploded on an oil tanker, causing oil prices to surge and then retreat
Two ultra-large oil tankers hit mines and caught fire in the Strait of Hormuz. Brent crude briefly surged to $110 but fell back to around $106 by the morning of the 18th, while WTI dropped below $99
The market is pricing in "no further escalation of risk"
They are very good at stirring things up—only controlling the tankers so they can't transport oil, not letting the tankers sink, and not allowing the conflict to escalate, just preventing oil transport
After liquidation, leverage remains unchanged
$BTC futures open interest still hangs at $52.6 billion, shrinking only 0.14% in 24 hours. Some long positions were wiped out, but the overall position size didn't decrease; just different players are holding on
Oil prices have retreated, but the mines remain. The position near 81,000 is a new position, not the batch that was already liquidated
#霍尔木兹再现水雷威胁，油价风险升温 The two flashes of fire in Hormuz are not just spikes on the oil price curve; they are two penetrating cracks torn open in the core of the global energy pipeline network—two supertankers hit mines and caught fire, one carrying over two million barrels of crude oil, instantly exceeding its load limit.
I've been in structural engineering for twenty years, and the two words I fear most are "to be confirmed." Until independently verified, all data are just renderings. Kpler reports transport volumes 43% lower than before the conflict; this is not a superficial color difference, but measured subsidence: the main building still stands, but the foundation beneath is sinking.
Brent closed at 109.86 on October 16, breaking through 110 intraday. That 3.7% is not an emotional premium; the market overnight recalculated the building's seismic rating. The original design assumed "the strait is always navigable," effectively placing all safety factors on a single column; now that column shows fatigue cracks, and the first reaction of the reviewer is never to repaint but to immediately add shear walls.
Iran's "conditional seven-day resumption of navigation" proposal via Qatar looks to me like a temporary scaffold. Scaffolds allow work to be done but are not included in structural calculations, do not bear long-term loads, and must be dismantled when the wind picks up. Without the other party's signature, this support system can't even enter the acceptance process; no matter how well the schedule is written, concrete cannot be poured.
As for the "alternative export," that's a makeshift bypass pipe. The diameter is insufficient, the joints don't match, and the pump station hasn't been expanded; hard connections will only transfer stress to the next elbow—the stress doesn't disappear, it just relocates, eventually bursting at an unmonitored node. True redundancy should have been reserved on day one, not patched after collapse.
Looking at linked instruments like $xIWM—they are like glass curtain walls hanging outside the main structure. Curtain walls are non-load-bearing but break first. In a real quake, the first to fall is never the core tube but the curtain wall; watching it shake violently and assuming it bears load is the most common mistake of outsiders: mistaking amplitude for load-bearing capacity. Its rise and fall are not fundamentals but wind pressure coefficients.
Designers know a building doesn't collapse from the top down; it starts with the foundation's moisture content. Seven days is not the construction period but the observation period; 110 is not the ceiling but the current stress reading of this building. #hormuzminethreatThe fastest scenario for storage is a rebound on the 21st; the worst case is that it will take until after the midterm elections. It's not exactly a rebound on the 21st, but the funds will start entering as the decline stops on the 21st. Meta's Muse will announce its revenue on the 21st, and the market will learn how much this AI agent can earn. Once the revenue is solid enough, the hardware will stabilize accordingly, since this means more computing power is needed. Similarly, the biggest shortage for such AI agents is CPUs, meaning INTC and AMD will see a wave of gains. Last time Muse came out, INTC rose from 100 to 140. The short-term king hesitated, waiting at 95, and ended up missing out. AMD also broke through 600, but personally, I prefer INTC. Americans want to make INTC the Nvidia of the CPU world. Based on previous situations, Muse's performance is very likely to be impressive and will significantly help hardware stabilize. Secondly, whenever the market faces such a big drop, big players always come out to speak, like this time it was Jensen Huang. Those in tech know how much Huang helps the sector, including in July and August when he often spoke up for tech, including Wash. Those currently stuck should not worry; it will basically recover. The AI bubble will not burst now. Huang wants the bubble to grow bigger. Right now, AI's biggest shortage is financing, and the market's biggest concern is profitability—whether the debt principal can be recovered. Corewave, as a central link, has a very high debt load. $INTC $SOXL $SNDK A $44 trillion pressure map has already been displayed on the forecast hall screen, but the warm and moist airflow supporting it—productivity measured on an all-economic scale—has not yet been observed at any ground station.
This is the most glaring contradiction in the current AI climate model: numerical forecasts are running two orders of magnitude faster than actual conditions. The $600 billion annual investment proposed by Goldman Sachs is equivalent to injecting a strong warm and moist advection accounting for 2% of the US GDP into the economy. Data centers and computing clusters are rapidly accumulating like cumulonimbus clouds on the plains, with cloud top heights continuously rising. Efficiency improvements at the enterprise level are localized showers—they have indeed occurred and have been measured by rain gauges—but extrapolating this to the entire basin’s runoff remains an extended forecast with low confidence.
We need to observe three layers of the weather system. The first layer: the capital expenditure front has passed, with thick clouds and ample lifting conditions. Giants treat computing power and data centers as moisture conveyors at the forefront of the cold front, with no signs of investment intensity weakening. The second layer: precipitation efficiency on the income side has yet to be verified, meaning there is much condensation in the clouds, but the proportion actually reaching the ground is questionable. The third layer: the valuation pressure gradient is firmly held by a low-pressure center; any productivity actuals falling short of expectations will cause a sudden wind shift, triggering short-term strong convection.
Next, look at the sounding curve of linked assets. Risk-hedging assets like $XAUT act as high-altitude temperature advectors; when the AI narrative pushes risk appetite to an extremely warm zone, their readings skew cold. Once the coverage issue of capital expenditure is exposed by actual data in earnings seasons, the temperature advection quickly reverses, and demand for hedging will rapidly descend like nocturnal radiative cooling. This is not a forecast for a single day’s weather but an assessment of the entire season’s circulation background—currently, the circulation pattern features a strong warm ridge and a weak cold vortex, but the ridge line position is extremely fragile.
My professional judgment is straightforward: $44 trillion is the ceiling scenario output by the model, not an actual forecast. What truly determines whether this AI front can bring effective precipitation is not how thick the clouds are, but whether the ground rain gauges can continuously record positive increments over the next few quarters. Until ground-level productivity actuals turn positive, all long-term growth precipitation forecasts remain within the dispersion of ensemble forecasts. #aicouldliftusgdpto44tLast December, he said BTC might drop to 65,000, and it later fell to 57,800. Now he says a new bull market has begun.
Fidelity Macro Director Timmer said on October 18 that a new four-year cycle bull market might be unfolding, adding: there seems to be no narrative at the moment, but bull markets rarely start with a narrative.
On December 18, 2025, he said (when BTC was at 88,000): the last halving cycle has ended in both price and time, it might pull back to 65,000 to 70,000, but is bullish in the long term.
BTC peaked at 126,200 last October and bottomed at 57,809 at the end of June this year, which is even deeper than his given range. Now at 81,798, it has recovered 41.5% from the low, still 35.2% below the high.
I think the power-law model gives a long-term slope, not turning points. 57,800 and 81,800 are only a few percentage points apart on the line, but for leveraged positions, it means life or death. So he wasn’t wrong both times; it’s just that this ruler can’t measure short cycles.
I don’t guess tops or bottoms, only look at ranges: September high 87,399, October low 79,775.
Only if it breaks above 87,399 is the new bull market confirmed; if it falls below 79,775, this wave is just a rebound.
Which side are you betting on: the start of a bull market or a bear market rebound?
$BTC $ETH $SOL
#机构资金持续流出，BTC能否企稳？ 【On-chain volume shrinks to a minimal level, what is the market waiting for?】
$BTC 24h trading volume is 11.89B, only 44% of the 10-day average volume of 26.72B; ETH is 5.92B, only 50% of the average volume. This minimal volume level has occurred a few times in the past year: early January 2026 (BTC 0.38x average volume, then rebounded 15%), mid-June 2026 (0.42x average volume, then continued to drop 8%). Minimal volume itself is not a signal; minimal volume plus key levels is. On 10/15, BTC volume surged to 42B (1.57x average volume), while hitting the monthly low of 79,736, which is a "testing the waters" signal — the first proof that someone is buying. But the second proof that it can't be pushed down hasn't appeared yet. Confirmation requires volume to break above 83,400.刚刚那一分钟，BTC又往下蹭了一格，盘口薄得像纸。 谁在真金白银地接这些货？ CoinShares最新数据挺扎心的：全球数字资产投资产品单周净流出4.34亿美元，美国现货ETF也在失血，BTC一周走了约2.25亿，ETH更夸张，直接撤了3.2亿。单日看，BTC净流出8240万，ETH连续三天净流出。SOL一周流出约2552万，XRP约212万，整体资金面还没掉头。 我盯着的不是这些数字本身，而是它们和美股风险偏好的错位。这周科技股其实没崩，纳指还在高位磨，但加密ETF却连续失血——说明机构不是在砍总仓位，而是在砍加密这个子仓位的优先级。跨市场联动看，美元指数没大动，美债收益率也没飙，外部流动性没紧到要命，钱是主动选择暂时不碰这个板块。这比"被宏观逼着卖"更微妙，因为一旦风险偏好回来，回补可以很快；但也意味着现在没有外部借口，纯粹是内部信心不足。 偏多的路径：BTC曾单日净流入1.76亿，随后又流出2.61亿，这种大幅摆动说明多空都在试探，不是单边溃败。如果下周BTC、ETH的ETF能连续转为净流入，哪怕金额不大，也代表新钱愿意接货，ETH被压得最狠，反弹弹性反而可能最大。偏空的风险：E$BZ believes that the 'Poison King' will want to secure the Senate in the midterm elections and also ensure that oil prices do not remain at such a high level. The extraction cost of US shale oil is around $68. As long as oil prices are controlled around $90, that is quite satisfactory for those oil giants. But now the price has reached about $106, which is clearly an overvalued price. Moreover, Trump will not take military action against Iran in the short term; if he really wants to act, it will be after the midterm elections, so the possibility of a second rise in oil prices is very low. Now is a relatively good time to short.There is an 84.5% probability of holding steady; this number is already set, but what’s truly worth pondering is whether the remaining 15.5% possibility will actually stir up waves in the market. According to the latest data from the CME FedWatch Tool, the market currently bets on an 84.5% chance that the Fed will keep interest rates unchanged in October, and a 15.5% chance of a 25 basis point hike (that is, raising the benchmark rate by 0.25 percentage points). Event-driven markets often move in two phases: the first phase is sentiment leading and price reacting first; the second phase is the fundamentals taking over, seeing if capital is willing to continue following. Looking at the current market coordinates: BTC is priced at 81,799.70, with a 24-hour increase of 0.27%; ETH is priced at 2,417.19, with a 24-hour decrease of 0.49%. Among spot USDT trading pairs with trading volumes above $100,000, 37 coins rose and 152 fell. Overall, the market has not shown a one-sided reaction to this news; bulls and bears are still probing. What to watch closely next are three lines. First, whether volume can pick up. In event-driven markets, the trading volume on the second day often carries more weight than the first day; price alone without volume means capital has not truly taken a stance. Second, whether other assets in the same sector move along. If only a single asset is rising, it usually concerns only itself and does not indicate consensus on the overall direction. Third, which way the funding rate is leaning. The funding rate is perpetualsand资金费开始叠了，叠到-1，每小时收取，看看后面会不会来极端逼空行情。US stocks are all taking hits, how far can the crypto rebound go? Looking at BTC and ETH these past two days, I think we still need to watch Wall Street's mood.
On October 16, the Nasdaq closed down about 2.1%, falling 4.5% for the week, marking its worst weekly performance since June. The S&P 500 also dropped about 1.1% that day. The pressure on tech stocks hasn't eased yet.
🔹 Why should the crypto market be cautious when tech stocks fall?
My understanding is that it comes down to how much risk capital is willing to take.
When investors start worrying whether AI investments can turn a profit or whether debt-fueled expansion can continue, they may reduce their positions in high-volatility assets. Even without new negative news, BTC and ETH could be dragged down by this sentiment.
🔹 What does the rotation of funds in US stocks mean for crypto?
On October 15, when the S&P rose, leading sectors included healthcare, indicating a shift in the market from AI to other sectors.
This reminds me: even if US stock indexes recover, it depends on who is rising. A shift of funds to defensive sectors doesn't mean everyone is ready to take risks again, nor does it directly imply an altcoin rally is coming.
🔹 What will I focus on next week?
Whether the Nasdaq can stop falling, whether selling pressure on tech stocks can ease, and whether there is spot buying support for BTC and ETH.
Today is Sunday, US markets are closed, but crypto is still trading. Weekend rebounds can be observed, but to confirm the strength of the recovery, we need to see if buying continues after US markets open next week.This meme coin is doomed to zero, nothing more to say! If the pumpers dare to push it back up, with so many people trapped, would they dare? 😎
【This short position is nailed tight】
A classic “A-shaped kill”!
From the peak at 0.12680, it plunged straight down to the current 0.06317, a direct 50% cut! A 7-day drop of -48.68%, those who chased longs at the 0.12 peak probably lost even their underwear now.
Finally caught this big trend! Shorted in at 0.07093, held all the way to now, floating profit soared to +32.69%!
Falling from 0.12, there was almost no decent rebound in between. How many retail investors are stuck up there? If the pumpers try to push the price up now, it would be charity to unlock the trapped positions above. The main players aren’t that stupid; they’ll just keep smashing it down, forcing retail to cut losses.
The price can’t even hold above the MA5, it’s a market fully dominated by bears.
Once market sentiment cools, these hype-driven meme coins are the first to be abandoned. Without funds to catch the fall, what would they pump with?
Pumpers, you were so arrogant when pushing the price up before, now you’re just as miserable smashing it down. I’m watching your show from the sidelines; if you dare to pump it back up now and free all the trapped retail investors above, I’ll salute you as a real man!
$STRK
#交易之声：你的经验值得被听到 Bullish brothers, open your eyes and take a look at this chart,
don't be deluded anymore.
Do you think 1130 is a solid bottom? Do you think it should rebound after dropping to 1140?
Let me tell you, a true bottom is never conjured up by willpower or fantasy.
Why is it falling? Look at what has happened on-chain in the past two days.
On-chain detective ZachXBT directly named suspects: North Korean hackers are allegedly exploiting Zcash's privacy features to transfer about $3.8 million ZEC stolen from a Bitget vulnerability.
Hackers are using it as a money laundering channel, and regulatory scrutiny will only get closer.
Meanwhile, Grayscale's ZCSH ETF funds continue to bleed, with a net outflow of $1.2197 million in a single day.
Smart money is withdrawing, dirty money is sneaking in; with such fundamentals, what can hold the thousand-dollar mark?
Look at the liquidation data. In the past 24 hours, ZEC has seen $8.8 million in liquidations across the network, with long positions liquidated at $6.7587 million and shorts only $2.0439 million.
The bulls are being ground down, one chain reaction liquidation after another. If you're still waiting for a rebound, all you'll get is the next batch of liquidations.
The next move is simple: hold your base position. If the pumpers really give face and it rebounds to 1180 to 1200 to lure longs, I'll take profits and roll them into short positions. Target? First watch the previous low at 1073; if it breaks, the abyss lies below.
Don't fight the trend
$BTC $ETH $ZEC
#特斯拉发布新一代Roadster，行情走高 #HBM4量产能否推动SK海力士盈利再创新高？
It can definitely reach new highs again because new technology often generates a high premium. And now, high-end memory itself already carries a premium, combined with the mass production of memory generated by new technology. They can further raise the price, so under this double-layered overlay, it will definitely bring more considerable revenue and profit when demand is high. Globally, there are not many manufacturers capable of producing high-end HBM4; it still belongs to a situation and pattern of a few companies jointly monopolizing. Although memory is a cyclical product with explosive demand and it is difficult to maintain long-term sustainability, unlike Nvidia, it will maintain a relatively high state before demand declines, possibly lasting one to two years. Theoretically, HBM demand normally wouldn't last this long, but indirectly, Jensen Huang previously collaborated with financial institutions like JPMorgan Chase, Morgan Stanley, Goldman Sachs, and AI companies to sign graphics card leasing agreements, extending the lifespan of these graphics card companies and allowing them to purchase more graphics cards with relatively reduced financial pressure. They can even come up with some money; for example, for a graphics card that costs 10,000 yuan upfront, through leasing, they only need to pay 100 yuan, and the remaining funds can be used to purchase memory modules and expand cloud data centers.Continuing with crude oil: short CL near 99 USD, first target at 95
Yesterday, my view on the crude oil rally over the week was just a sentiment premium from the Strait of Hormuz, not a real supply-demand gap
Today the market has already responded, CL in the after-hours session directly broke below 99, dropping to 98.68, high-level longs are starting to loosen, which from time to time means the sentiment premium is slowly dissipating
No need to focus on conflict news from the strait, as this blurs the distinction between friction and blockade
Although there was an incident with an oil tanker, the channel has not been completely closed, and original supplies have not been cut off
As long as the strait shipping is not cut off, the rally driven by panic is destined to fall back
The short-term longs piled up from last week's 10% surge are holding rich floating profits over the weekend
Now that the market has turned, profit-taking orders will gradually emerge, and concentrated long exits will accelerate price declines
The self-regulating logic of high oil prices still holds
As long as oil prices stay above 100, energy inflation will rise
After the Fed's rate cut expectations retreat, US Treasury yields rise, and the Nasdaq and semiconductors fall
The current oil price pullback is easing this inflation pressure
Once oil prices continue to decline, CPI and PPI inflation pressures will ease
The biggest pressure on tech stocks will be relieved
Next week we will focus on two levels:
First, 95 is our first target for shorting CL; after reaching it, we need to observe the strength of the stop in the decline
Second, watch if US Treasury yields and the dollar pull back, and if semiconductors and Nasdaq rebound
In this round of market moves, crude oil is the master switch of the market volatility
Oil price falls, inflation drops, and US stocks can rebound Recent Impact of the US Dollar Index Trend on Crypto 【Macro】
The US Dollar Index (DXY) has been oscillating around 103 recently, with no clear direction. For crypto, the negative correlation between DXY and BTC has weakened somewhat in 2026, but they still move together during extreme market conditions. If DXY breaks above 105, risk assets are likely to come under pressure; if it falls below 102, capital may flow back into risk assets.
Currently, both BTC and ETH are positioned 30% and 44% below their 52-week highs, respectively, pricing in quite a bit of negative sentiment. However, "falling a lot" does not mean "the fall is over"; the framework ensures small losses but does not guarantee every trade will be right. This is not the hitting zone now; playing dead is also a way to hold positions. 先对账，昨天夜盘的三句话。 第一，SOL 105 生死线，昨晚收 104.02，真跌破，我认。第二，BTC 反转确认线 82600，今天 24 小时最高只摸到 82480，连线都没碰上，反转这轮正式盖章：没戏。第三，81k 守得住就继续磨，今早最低扎到 80813.6，离昨天低点 80809.5 就差 4 刀，硬是没破，弹回 81897。 上午市场干嘛呢？一个字，磨。BTC 81,897，24 小时涨 0.44%，上下才 1.5 个点。ETH 2422，微跌 0.19%，但费率 0.0066% 是三兄弟里最高的，多头还愿意多掏利息，没完全躺平。SOL 104.27，跌 0.48%，费率 -0.0007% 还趴在负数里，空头没打算放手。 下午就看两件事：BTC 摸不摸得上 82072（上午高点），摸上去才有资格再去碰 82600，不站稳反转就是说说而已；往下盯 80813，破了今天低点，80000 心理关直接开战。 证伪写死：如果今天 BTC 收回并站稳 82600，我"反转没戏"的判断当场作废，今晚夜盘开篇第一个认错。 你觉得 82600 这条线今天有戏吗？评论区说。At first, I thought that after the crash at midnight, the leverage on ETH would lighten a bit.
Just checked OKEx ETH perpetual contracts, and the lowest point in that one hour at 5 AM was 2,355, with positions dropping by about 23,000. BTC also dipped a bit in the same hour, but positions basically stayed the same.
By noon, ETH positions had bounced back to about 587,000, a bit more than before the crash, and the price had climbed back up.
The wave of liquidations at midnight was replenished within seven hours, and leverage basically piled back up to last night's level. If by this time tomorrow OKEx ETH perpetual positions drop below 570,000, then I was wrong.
Were you liquidated in that midnight crash, or did you place orders early to catch the dip below?
$ETH #奈飞增长或放缓，广告业务成新看点
Looking back at ETF subscriptions and redemptions from October 12 to 16, the combined net outflow from the three spot channels BTC, ETH, and SOL was about $574 million, all moving in the same direction with no channel recording a weekly net inflow. This is not a fatal signal for the rebound, but it does indicate insufficient capital coordination.
$BTC: There was some inflow, but the continuity was lacking. This week saw a net outflow of $228.4 million. After an inflow of $175.7 million on Tuesday, it turned to an outflow of $261.4 million on Wednesday, followed by another outflow of $82.4 million on Friday. The buying side is not absent; the problem is the rapid switching between inflows and outflows, preventing continuous support.
$SOL: The outflow scale is small, but the trend signal is stronger. This week had a net outflow of $25.5 million, which is not significant in absolute terms, but there were net outflows for 10 consecutive trading days totaling $50.5 million. On Friday alone, $12.7 million flowed out, nearly half of the week's total. Continuous withdrawal is more noteworthy than a single large outflow, indicating that capital is steadily reducing exposure.
$ETH: Weekly outflow exceeded BTC. This week saw a net outflow of $320.1 million, surpassing BTC. There were outflows for three consecutive days from Wednesday to Friday, totaling $259.5 million. Looking only at the ETF channel, ETH has yet to see stable replenishment.
Of course, ETF subscriptions and redemptions are only part of the market's buying activity and cannot be used alone to conclude a decline. But to make this rebound more convincing, at least several consecutive trading days of net capital inflow need to be observed. 风险资产同步回落，$BTC 一度下探约81,000美元，$ETH 触及2,365美元，但关键防线并未失守：80,000美元与2,400美元分别构成短期支撑。真正值得盯住的不是跌幅本身，而是“价格扛住、资金却在撤”的背离——ETF单周净流出6.81亿美元，说明这一轮抛压并非散户情绪驱动，而是机构在主动降低风险敞口，短线定价权已经交到机构手里。这也解释了为何技术面依然偏弱：均线结构未修复，反弹缺乏成交量配合，价格只是被托住，而非被买起来。 从市场结构看，80,000美元与2,400美元之所以关键，是因为它们是本轮调整中多头最后的成本防线。守住，意味着抛压被逐步消化，空头需要新的催化剂才能继续施压；失守，则前期低位买盘将转为浮亏筹码，容易触发连锁减仓，调整空间随之打开。换句话说，这两个价位不是预测点位，而是多空力量的分界标尺。 多头要真正夺回主动权，需要两个条件同时成立：其一，FOMC释放不再偏鹰的信号，缓解风险资产的估值压力；其二，ETF资金流由负转正，证明机构从“撤退”切换为“回补”。缺一则反弹更可能只是区间内的技术性修复。空头的触发条件同样明确：$BTC 有效跌破80,000美元、$E"CPI/PPI Have Both Surged, the Last Piece of the Puzzle for a December Rate Hike"
After CPI at 3.7% and PPI at 6.3% both surged, the FOMC meeting on October 27-28 was basically a formality—the quiet period has begun, with an 87% probability of no action.
But December will be different, with bets already up to over 74%. The issue is: before the December decision, two key hard data points remain—October 29 GDP preliminary and PCE. PCE is the inflation indicator the Fed truly watches; even if CPI/PPI surge again, if PCE is moderate, the story could reverse.
So the script for the next two weeks is clear: the data vacuum will be supported by earnings reports (Netflix, Tesla, SpaceX all next week), and the real directional choice will come after October 29. BTC in the 80000-82000 range will likely consolidate until the end of the month.
The conclusion remains: no move unless 80000 breaks; if it breaks, watch 78000; if it holds, watch 83000.
$BTC
#美国CPI与PPI双超预期，年内能否再加息？ "One Car, Three Prices: What Is Musk Selling with the Roadster?"
This time, the information is more than what was shared on the night of the launch event. Three versions: Performance two-seater street-legal (deposit $50,000), Thrust two-seater track edition (deposit $250,000), Rocket single-seater by invitation only (price not disclosed). Three price points, three types of customers—Musk has clearly mastered scarcity marketing.
Technical details have also emerged: 10 thrusters, 600-liter compressed air tank, confirmed by engineering VP Lars Moravy himself. But aviation experts poured cold water: the demo was a hop, not a flight—just a few seconds of bouncing, which is a far cry from real flying. Musk also painted a big picture: the car will be able to perform all stunts autonomously, and the Texas headquarters will build a test track.
The most interesting part is the timing: six days before the Q3 earnings report on October 21, the Roadster isn’t just selling a car, it’s selling the sentiment that "Tesla engineering is still world-class." The AI narrative for Robotaxi + Optimus needs this sentiment as a foundation. It doesn’t matter that the car has been delayed for 8 years; the story must continue.
#特斯拉发布新一代Roadster，行情走高 Volume contraction at the bottom, BTC fiercely defends the 80,000 mark!
Current market shows intensified divergence between bulls and bears, overall capital flow is relatively cold, and the fear-greed index has cooled down to 49.
Market snapshot: BTC resists decline, altcoins bleed
BTC (Bitcoin): Firmly holding the $81,500 level, almost no change in 24h (-0.01%~+0.28%). However, the weekly chart shows two consecutive declines, with a 1.7% pullback this week, in a low-volume sideways consolidation.
ETH (Ethereum): Around $2,410, slight intraday rise but still weak, weekly drop reaches 3.8%, with a dip to $2,356 at early morning, clearly underperforming BTC.
Bitcoin's market dominance surpasses 60.0%, indicating funds are flowing into relatively safer top assets, and the market is in a defensive stance.
Total market capitalization is about $2.74 trillion, down approximately 3.0% in 24 hours, with altcoins under significant pressure. 🍜 Sunday noon, meme points, contract DEX, and on-chain gold — the three essentials
$BOME 0.0006958, the one in the meme circle talking about points gameplay 📉
BOOK OF MEME is not just about price fluctuations; it turns meme creation and referencing into an on-chain points system where users can earn tokens by posting and referencing. This design sets it apart from one-off memes, adding a retention mechanism. But this round of sentiment decline saw it drop from 0.000725 to 0.000695, relying purely on meme capital sentiment; the points gameplay can’t sustain the price.
$ASTER 0.7059, liquidity mining tied to contracts 🚀
Aster links token distribution to contract trading depth; users can mine tokens by providing liquidity and trading. This is how it competes with top DEXs for users. With Google Cloud growing 82% and AI investment returns, if enterprise AI adoption scales up, on-chain contract activity will pick up first, and second-tier DEXs will then get priced.
$XAUT 4154, the most active gold token on exchanges 🥇
XAUT is listed on major exchanges, tradable 24/7, allowing gold exposure without waiting for spot market openings. Stable at 4154 on Sunday noon, its volatility follows London gold spot prices, but on-chain premiums slightly widen during data release periods.
This article is for market analysis only and does not constitute investment advice.
#谷歌云业务增长82%，AI投入回报如何？ Chip index hits longest losing streak since 2008, Nvidia didn't escape either
Philadelphia Semiconductor Index SOX fell another 3.83% on Friday, marking its 8th consecutive trading day of decline, the longest downtrend since the 2008 financial crisis.
Nvidia $NVDA closed at $214.13, down 1.43%. The drop wasn't a crash, but the entire AI supply chain is dragged down by customer financing issues.
The trigger seems more related to balance sheets: CoreWeave's bond yield surged to 15%, dropping about 25% in a week. The market is questioning—how much longer can the model of borrowing money to stack computing power hold?
My view: This isn't the "end of the AI narrative," but more like a revaluation of financing costs. Demand from big tech companies is still expanding, but the highly leveraged GPU buyers are the first to get hit.
The failure conditions are simple: if SOX stops its losing streak next week and bond yields fall back, it means panic is subsiding; if NVDA breaks its previous low again with high volume, watch positions need to be further reduced.
I won't bottom-fish or catch a falling knife over the weekend. What signals will you be watching more closely at Monday's open?
#USStocks #AIChips $NVDA $AMD $AVGO#NetflixQ3AdsAndMargins The French National Assembly has passed a tax amendment targeting stablecoins. The focus is not on "raising taxes" per se, but on a shift in regulatory approach: previously, the emphasis was on issuers and reserve assets, but now taxation is being applied to the circulation of stablecoins.
Stablecoins are the main on-chain carriers of the US dollar, and the $ETH chain is one of the networks with the highest concentration of stablecoin transfers. If actions like transfers and exchanges are included in the tax scope, the cost of using on-chain dollars will increase. For this chain, the short-term impact will be reflected in activity levels and fee expectations, rather than the price itself.
There are three things to watch next.
First, the scope of application: whether the tax applies only to transfers or also includes issuance and redemption.
Second, what legislative procedures remain to be completed, as the final version may still be modified.
Third, whether there are signs of transfer in the volume of on-chain stablecoin transfers, such as large settlements moving to other jurisdictions or other chains.
If the cost is truly implemented, the first to be affected will be high-frequency scenarios like market making, payments, and cross-border settlements, while long-term holders will feel little impact. From the perspective of the overall market, this type of policy is more like a slow variable; it will not immediately change direction but will influence whether funds are willing to keep on-chain dollars in Europe.
#法国众议院通过稳定币税修正案 $NFLX will disclose its Q3 earnings after the U.S. market closes on October 20. My attitude this time is cautiously optimistic.
The company has provided guidance, with a revenue target of $12.86 billion for the quarter, representing about a 12% year-over-year increase, and an expected operating margin of 33.2%.
My prediction: revenue will likely be close to the guidance, with a slight chance of profit exceeding expectations. Advertising monetization is the core focus of this earnings report.
There is a key point that cannot be overlooked.
Last year at the same time, there was a large tax expense that suppressed the profit margin for that period.
Therefore, the year-over-year increase in profit margin this year requires careful analysis to distinguish how much is due to actual business improvement and how much is due to the fading of last year's one-time factors.
Whether the stock price can rally after the earnings release does not solely depend on meeting the quarterly performance targets.
The real catalysts for the market will be two core factors: the actual growth of advertising revenue and whether the management's guidance for Q4 can exceed market expectations.
The potential for a positive surprise lies in these two areas. We await the post-market earnings release for verification.
#奈飞增长或放缓，广告业务成新看点 🚨🚨🚨 Bitcoin has shown another very typical local bottom signal!
The day before yesterday, BTC dropped back near $80,000, and the market once again experienced obvious realized loss selling pressure.
But this Glassnode on-chain data reveals something very interesting.
Looking back at the past three times:
📉 Early August: A large number of investors sold at a loss, followed by a price rebound.
📉 Mid-August: Loss selling pressure erupted again, and BTC then launched a strong rally.
📉 Mid-September: The market again saw concentrated loss-taking, followed by a price recovery.
All three clear panic loss-taking events happened just around local lows.
And now, the fourth signal has appeared.
What’s even more noteworthy:
Since the September high, the scale of profit-taking in the market has been gradually decreasing.
This means the profit-taking selling pressure that had been continuously suppressing the price may be weakening.
On one side, profit-taking is slowing down.
On the other side, panic loss-taking is appearing again.
This kind of chip structure often deserves special attention near market turning points.
If history repeats itself...
Then yesterday’s sell-off near $80,000 might not be the start of a new downtrend.
Instead, it could be the final clearing of the last batch of weak hands before the next upward move.
Of course, loss-taking selling pressure itself does not guarantee a bottom; price needs to stabilize above support and buying must continue to absorb supply.
$BTC $ETH $ZEC Giving points is not just a free giveaway; it's letting you make judgments for the platform
I received about three hundred points and casually bet that Tesla's earnings report would exceed expectations.
Only after placing the bet did I remember that this judgment originally had no value.
What is this event:
Points are issued by the platform, and you don't lose your principal if you bet wrong.
If you bet right, you can share $25,000.
Where does the money come from:
The platform uses this money to buy your attention.
Once you make your choice and want to wait for the result, the page opens several more times.
My guess:
People who bet real money will first check the earnings report date.
Those who bet with free points are mostly guessing.
When these two types of people mix in one pool, the results become interesting.
The $BTC market is the same; many watch, few calculate.
#特斯拉发布新一代Roadster，行情走高
#$BTC $TSLA The more pessimistic the market, the more attention should be paid to position structure.
According to this set of data, currently $BTC net short positions account for 62%, while net long positions only account for 38%.
What does this mean?
1️⃣ The bearish sentiment in the market may already be quite concentrated.
2️⃣ If negative factors continue to ferment, shorts may still profit.
3️⃣ If the price fails to drop for a long time, short covering could become the driving force for a rise.
This is an important logic in trading: beyond direction judgment, you also need to see how many people have already bet on the same direction.
However, the net short ratio alone cannot predict price movements; it must be combined with funding rates, open interest (OI), liquidation data, and spot buying to make a judgment.
When everyone is bearish, the real risk for $BTC may not only be a continued decline but also a sudden reverse rally.
Do you think it is currently a downtrend, or are the shorts starting to get crowded? From 1,000 USD to 100,000 USD in real trading — Day One
Let me explain why I buy Xiaomi $XIAOMI
Why treat Xiaomi as a target for regular investment? To be frank, I'm not betting on its robots going into mass production tomorrow; I'm betting on a sequence difference: factories start first, policies follow, financial reports gradually reflect it.
"Tieda" is now really working at Xiaomi's car factory. The success rate at the self-tapping nut station has improved from 90% to 98%, manual work is about 99%, just one point short. The goal is 99% by the end of the year. This is still far from "mass deployment," and the team themselves say complex tasks may take another two to three years. But the place where it interns is its own production line, no need to seek customers or compete for orders, which is different from most robot stories.
Lei Jun officially proposed at this year's Two Sessions to subsidize and give priority access to factories deploying humanoid robots. Subsidies haven't been implemented yet, but Xiaomi itself is the "early deployer" sample. I think this positioning is more practical than just marketing.
On the automotive side, monthly deliveries have consecutively exceeded 30,000, and revenue from chassis is holding up. But I don't want to hype the financial report as "definitely good." Storage price increases and pressure on automotive gross margins exist, and broker forecasts vary widely. CITIC Securities even predicts a 27% year-on-year adjusted net profit decline in 2026. So the real meaning of regular investment here is: exchanging time for robot options while enduring short-term profit noise from the automotive business.
Conclusion: Xiaomi with 5x leverage, buy on dips, just hold the spot, and let the bullets fly for a while
#特斯拉发布新一代Roadster，行情走高 An aggressive scenario placing US GDP at $44T by 2030 draws a useful distinction: AI investment is visible, while economy-wide productivity is not. Goldman Sachs' estimate of roughly $600B in US AI investment for 2026 shows commitment, not proof that returns will match capex.
The valuation test is whether efficiencies become durable revenue beyond a handful of firms. That evidence is still forming.
#AICouldLiftUSGDPTo44T 5:15 $BTC /$ETH /$SOL Early Morning Sharp Drop Analysis
The sharp drop at 5 AM superficially appears to be due to macroeconomic negative factors, but at its core, it is a chain liquidation of highly leveraged long positions.
The direct trigger was a whale liquidation. On-chain data shows that a whale who previously opened a $187 million long position with maximum leverage had 4,890 ETH longs liquidated at 5 AM at $2,356, with a position value of $11.74 million. The $2,356 price was exactly the lowest point ETH reached this morning. This liquidation was like toppling the first domino, triggering a cascade of programmed sell orders. In nearly one hour, over $100 million was liquidated across the network, with longs accounting for 84% of the liquidations.
The deeper background is continuous capital outflow. This week, the US spot BTC ETF saw a net outflow of about $228 million, the ETH ETF outflow was even more severe at $299 million, and the SOL ETF has also experienced multiple weeks of outflows[citation:21]. Institutions are withdrawing, leverage is holding on, and once prices break key thresholds, a stampede is almost inevitable.
The differentiated declines among the three are also noteworthy. SOL’s 20-day drop is about 14%, ETH about 9.5%, and BTC only 3.4%[citation:21]. BTC’s market dominance is approaching 60%, with funds clearly contracting defensively into BTC, making altcoins the first victims of liquidity withdrawal.Today is Sunday, BTC is stuck between 81,000 and 82,000.
CoinXSight provided a clear weekend step chart: the first resistance is at 82,500 (Friday's battleground for bulls and bears, a low-volume fake breakout is a bull trap); above that 82,500 (Friday's battleground for bulls and bears, a low-volume fake breakout is a bull trap); the real action starts at 83,500 — where about 450 million dollars worth of short liquidations are clustered (82,800–82,800–83,500). Once a volume-backed breakout happens, short covering will cascade like dominoes toward 85,000; above that 85,000; and then 87,000 is the "short thesis invalidation line" — if it closes above this over the weekend, the trend will fully flip bullish.
On the downside, support is seen at 80,925 (the low point smashed out in three minutes on PPI day) and 80,925 (the low point smashed out in three minutes on PPI day) and the round number 80,000. Additionally, on 10/30, Deribit has 10 billion dollars in options expiring, and gamma positions will pull the price toward the magnet zone — this weekend is not a time to lie flat, but a prelude to position setting.
#美国CPI与PPI双超预期，年内能否再加息？ Oil, US stocks, and crypto are now tied together on the same rope, and the order is crucial.
Oil is the lifeline: boosting domestic energy not only suppresses inflation but also supports the dollar's floor. US stocks take the blame: tariffs push inflation up, rate cuts are stuck, valuations remain high, so the pain of deflating bubbles hits them first. Crypto is the ATM: no matter how good the policies are, when money really runs short, Wall Street will sell it first.
So don't just focus on positive announcements; watch where the money comes from. Wall Street won't lose, oil must be protected, and the easiest to get squeezed painfully is actually US stocks.
$SNDKMSTR closed at 136.83 on Friday, dropping about 9.56% in one day. I think this is index panic, not a crypto crash, so I'll observe first and not chase.
Observed: Opened at 145.95, high 146.83, low 134.11, volume about 31.97 million shares, previous close 151.29.
On October 15, MSCI finalized the non-operating company rules, the final list will be released on November 11, and will take effect after market close on November 30.
Strategy holds about 848,000 BTC, average price about 75,444; the company warned last year that a true removal could trigger about 2.8 billion in passive sell orders.
But current component stocks must fail screening for two consecutive annual reports to be kicked out, not out tonight.
On the same day, BTC on OKX is still around 81,900, showing a clear divergence between stock and crypto.
My view: What's being crushed is the index qualification premium, not BTC itself.
What to do: Observe and don't chase; wait for volume to push back above about 146.83 before reconsidering. If it falls below about 134.11, the low of this long bearish candle, then the strategy is invalid.
What do you think? Before the list comes out on November 11, will MSTR continue to kill the premium or slowly grind it back?
$MSTR $BTC $IBIT
#MSTR drops nearly 10% due to MSCI rules
#BTC spot ETF outflows about $82.4 million again on Friday盘面弹了一下又软下来，群里瞬间安静了几秒。 这波到底是真企稳，还是又被假尖刺骗了一次？ 我盯着那张现货 ETF 持续净流出的表，心里其实已经有答案了。价格没崩，但机构在减仓，这两件事同时发生，本身就是信号。大钱不是在抄底，是在等位置。这种姿态比砸盘更磨人，因为它意味着上方每次反弹都会遇到"回本就走"的抛压。 市场现在交易的其实不是"BTC 会不会归零"，而是"降息预期被提前计价了多少"。ETF 连续流出说明部分机构认为宽松节奏没那么快，先把风险敞口收一收。这个动作会传导到哪？先是 BTC 在高位反复洗盘，杠杆多头被来回扫；再是 ETH 跟涨乏力，山寨更明显，叙事疲劳叠加缺少新增资金，热点只能靠存量博弈维持，追高的人第二天就变成接盘的。 偏多的路径也不是没有：一旦 ETF 流出放缓、稳定币溢价回正、交易所库存出现下降，说明观望的钱开始回头，那时候的反弹才更有说服力。但风险点在于，现在很多人把"没跌"直接翻译成"要涨"，提前把仓位打满，结果每次假突破都被洗一遍。真正没被看见的，是时间成本，横盘越久，情绪越容易从犹豫滑向 FOMO，然后在最不该冲的时候冲进去。 我自己的节奏是：机构回来加上关Tesla's Q3 deliveries exceeded expectations, but the market's focus has quickly shifted to profit recovery. Deliveries are the "past tense," while profitability and cash flow are the "future tense" that determine stock price direction.
🚗 Deliveries beat expectations, but quality is questionable
Tesla delivered 486,532 vehicles globally in Q3, surpassing Wall Street's consensus estimate of about 461,000. However, there are concerns behind this "beat":
· Year-over-year contraction: Deliveries fell 2.1% from a record 497,099 vehicles in the same period last year, when U.S. buyers rushed to purchase before the federal tax credit expired.
· Reliance on inventory drawdown: Deliveries this quarter exceeded production by about 22,141 vehicles, marking the second consecutive quarter of inventory consumption. These sales may have involved discounts, putting pressure on gross margins.
· Deterioration in model mix: Model 3/Y accounted for 98% of total deliveries, while deliveries of "other models" like Cybertruck and Semi plunged about 48% to only 8,295 units.#特斯拉交付超预期后盈利修复成焦点 #HBM4 mass production: Can it drive SK Hynix's profits to new highs? SK Hynix is about to release its Q3 earnings report, and the market has high expectations for its HBM4 mass production. However, in my view, the resistance to pushing profits to new highs through this is not to be underestimated.
First, although HBM is hot, it is difficult to support the overall situation alone. If the price cycles of general DRAM and NAND decline, combined with industry expansion and competitive pressure, overall profits will inevitably be under pressure. Second, from a valuation perspective, the current stock price and P/B ratio have most likely already priced in the profit expectations of HBM4 in advance. Unless upcoming pricing or market share significantly exceeds expectations, the positive news of "new highs" is unlikely to further drive valuation upward.
Finally, the capital market also sends warnings. Recently, foreign investors have been net sellers continuously, coupled with the "draining" effect of Changxin news and ADR premium contraction, market selling pressure is obvious. With intensified earnings season battles, everyone should remain cautious of the risk of "all good news being priced in" and avoid blindly chasing highs. $SKHYNIX $SKHY #VanEck停止推进Lido质押以太坊ETF
🚨On October 16, 2025, VanEck filed the S-1 for the Lido Staked Ethereum ETF, becoming the first issuer to attempt to create a US ETF directly based on Lido liquid staking tokens. One year later, on October 14, 2026, it submitted Form RW to the SEC to voluntarily withdraw the registration.
The registration statement never became effective, and no shares were ever issued or sold. The document only stated that there was "no intention to continue advancing," without providing specific regulatory, market, or technical reasons.
By 2026, the market already had ETH ETFs with direct staking. For example, BlackRock's ETHB stakes most of its ETH holdings and distributes rewards to investors; Grayscale and others have also started staking and distributing earnings.
Lido remains the largest liquid staking protocol in DeFi, but this withdrawal indicates that the Wall Street path of packaging it as a standalone ETF product is not smooth. Meanwhile, Lido's TVL has declined from nearly $38-40 billion at the time of application to about $23-24 billion, weakening the product narrative.
This signals an evolution in market structure—the native staking ETH ETFs can already meet the demand for "holding ETH + earning staking rewards," reducing the necessity of Lido as an intermediary layer in ETF packaging. This is not a rejection of staking or Lido, but a result of product design choices.In 5 trading days, hedge funds sold off all the AI stocks they had bought since 2025 — numbers from Goldman Sachs Prime ledger.
From October 8 to 14, long-short funds net sold AI stocks for 5 consecutive days, wiping out all net purchases since early 2025, including one day that was the second largest single-day net sell in the past year.
Long-short ratio 3.5 to 1: not shorting, but cutting longs. NVDA fell from 237.47 to 221.30 over these 5 days, closing at 214.13 on Friday, nearly a 10% pullback from the high; the Philadelphia Semiconductor Index fell for 8 consecutive days, the longest streak since 2008.
That week, global fundamental long-short funds lost 2.9%, Asian funds lost 6.8% in the first half of October. Goldman Sachs said this is the third time this year that "all strategies fell sharply on the same day," something that did not happen even once from the pandemic until the end of last year.
My judgment is that the market is not trading AI fundamentals, but "positioning" — crowded longs killing each other, AI demand itself has not been disproven.
The real verdict comes next week: about 70 S&P companies including Tesla and Meta will report earnings; if AI revenue and guidance hold up, this wave is just an emotional purge.
Don’t act before earnings, wait for positions and fundamentals to align.
If earnings exceed expectations and stabilize, the "deleveraging continuation" logic fails; if earnings continue to drift down and spread beyond the AI sector, the "golden pit" theory fails.
Which side do you believe more: a buying opportunity cleaned out by panic, or the start of a collapse from crowded trading?
#纳指100周跌5.5%，芯片股八连跌
$NVDA
#AI交易大撤退BTC broke below 80,000 this week but was bought back; I am "observing, not chasing," with the lifeline at 79,775.
On 10/15, the daily low hit 79,775 and closed back at 82,232 the same day.
Currently, OKX quotes around 81,900, with low volume sideways trading over the weekend.
The US spot BTC ETF saw a net outflow of 82.4 million USD on 10/16, with a cumulative outflow of about -615 million USD in October.
The total liquidation across the network dropped from nearly 1 billion USD around 10/15 to about 222 million USD on the most recent day.
I think holding 80,000 this time relies on spot buying after leverage clearing, not new money entering.
With CPI and PPI both exceeding expectations and ETFs still flowing out, it's hard to see an independent rally now.
My approach is to observe, not chase; first, see if it can hold above 80,400, and only a reclaim above 84,361 confirms a bullish trend.
From tonight until Monday 8 AM is the weekly close; if the daily or weekly closes break below 79,775, I admit being wrong and the 80,000 logic is invalid.
Which side are you on: holding 80,000 and pushing to 84,361, or breaking below on the weekly close and probing lower?
$BTC $ETH $MSTR
#US CPI and PPI both exceed expectations, will there be another rate hike this year? #BTC marketTesla's Q3 deliveries exceeded expectations, but the market's focus has quickly shifted to profit recovery. Deliveries are the "past tense," while profitability and cash flow are the "future tense" that determine stock price direction.
🚗 Deliveries beat expectations, but quality is questionable
Tesla delivered 486,532 vehicles globally in Q3, surpassing Wall Street's consensus estimate of about 461,000. However, there are concerns behind this "beat":
· Year-over-year contraction: Deliveries fell 2.1% from a record 497,099 vehicles in the same period last year, when U.S. buyers rushed to purchase before the federal tax credit expired.
· Reliance on inventory drawdown: Deliveries this quarter exceeded production by about 22,141 vehicles, marking the second consecutive quarter of inventory consumption. These sales may have involved discounts, putting pressure on gross margins.
· Deterioration in model mix: Model 3/Y accounted for 98% of total deliveries, while deliveries of "other models" like Cybertruck and Semi plunged about 48% to only 8,295 units.#特斯拉交付超预期后盈利修复成焦点 Anthropic, the developer of Claude, has been rated as the highest risk tier.
The rating agency gave the score in June, but it was only announced in October.
The community immediately started arguing. Because this is not just about them alone.
The problem is not limited to just one company.
The interest rate on U.S. Treasury bonds has also risen to the highest level in 24 years.
Lending money to the government is the most expensive it has been in 24 years.
The notion that this is the safest asset is now being questioned.
Why can Treasury bonds also lose value?
When interest rates rise, previously issued bonds lose value.
For example: you hold an IOU that promises to pay you 3 units in a year. Now the market offers 5 units.
Your IOU is unwanted. If you want to sell it, you have to discount it—selling at a loss.
U.S. Treasury bonds are the same.
They don’t owe you the principal, but they owe you interest. When interest rates are not high enough, you can only sell at a discount.
Even that old reliable allocation is no longer stable.
The traditional safe allocation is 60% stocks + 40% bonds. One goes up while the other goes down, theoretically offsetting each other.
It’s stable on the premise that the bond half never loses value.
That premise is now starting to loosen.
The case for Bitcoin
Adam Back says: before buying anything, compare it to Bitcoin—would buying it be worse than just holding Bitcoin?
The problem is it’s hard to compare. Bonds are backed by the government; Bitcoin is backed by nothing.
Moreover, bond ETFs are currently attracting massive inflows, with money still pouring into that old path.
So
Anthropic is being viewed pessimistically$115 million long liquidations in one hour, this drop really hurts.
First, what happened: On October 17, BTC once fell to about $81,000, with over $115 million long liquidations within an hour. Fortunately, it stabilized between $81,000 and $83,000 afterward, and weekend volatility was low, so it caught a breather. But the funding situation isn't great; on October 16, the US spot BTC ETF saw a net outflow of about $82.4 million, and sentiment clearly hasn't recovered yet.
My own judgment: The pressure was absorbed around $81,000, indicating panic hasn't spiraled out of control; however, with ETFs still flowing out, there's a lack of fuel for a decent short-term rebound.
What to watch next? During the week of October 19, over 200 US companies will report earnings, including Netflix, Tesla, and Meta. If these giants trigger big fluctuations in their results, sentiment can easily spill over into the crypto space. BTC is still following the mood of risk assets.
In short: Stabilizing doesn't mean a reversal. Let's see how earnings week plays out before setting the pace.