Orbit Post Sitemap
#台积电Q3营收创新高，盈利表现待揭晓
这次财报季，台积电算是个很有意思的样本。提前放出来的三季度营收，实打实刷新历史纪录，同比直接涨了一半，数据摆在那里是真的亮眼。
但有意思的来了，消息出来股价反而往下掉了一截。说白了市场早就料到AI订单会把营收冲上去，大家现在根本不纠结卖了多少货，心里全悬在利润上面。
不是说营收高就万事大吉。现在大家就盯着毛利率能不能稳住65‑67%这个区间，还有四季度的业务展望。
短期股价下跌，更多属于利好落地后的资金避险，也不能直接就看空。AI芯片这块实实在在的需求还在手里。只不过现在不会再无脑吹，营收好看只是一张入场券，能不能赚到钱，今晚完整财报才会给出真正答案。后续能不能继续走强，就看利润端能不能跟上营收的脚步。#solana is scheduled for the Alpenglow upgrade on October 21
[Old Leek Observation] $SOL will have a core upgrade next week: confirmation time targeted at 150 milliseconds
Solana's official team has scheduled the Alpenglow mainnet switch for October 21. This is not an ordinary parameter adjustment but a replacement of the long-running TowerBFT with the new Votor consensus mechanism.
Currently, Solana's transaction finality takes about 12.8 seconds; the new mechanism aims to reduce this to around 150 milliseconds. Simply put, the goal is to make transactions reach an irreversible state faster, which has potential value for high-frequency trading, on-chain transaction applications, and scenarios requiring rapid settlement.
First, 150 milliseconds is the target, which does not mean the mainnet will immediately and stably achieve this under all network conditions upon launch.
Second, the new mechanism will move validator voting out of regular transaction blocks, so the number of transactions displayed on-chain may change after the upgrade. One should not judge Solana's user activity solely by the rise or fall in transaction counts.
Third, SOL's price will also depend on whether the market has already priced in expectations. If there was a price increase before the upgrade, profit-taking might occur at the official switch; if the upgrade is successfully implemented, actual performance improves, and capital continues to flow in, the technical advantage has a better chance to translate into sustained demand.
My judgment: this is a clear technical catalyst but not an unconditional bullish signal #TSMC Q3 Revenue Hits Record High, Profit Performance Yet to Be Revealed TSMC's Q3 revenue is about NT$1.49 trillion, up approximately 50% year-on-year, setting a historical record. Impressive revenue does not guarantee a winning financial report; tonight's focus is on gross margin, net profit, and Q4 guidance. AI computing power orders are the main reason, with companies like Nvidia relying on it for foundry services; however, high costs of building factories in the US and slow yield ramp-up pose hidden risks. Revenue is set, but profit performance is the real reveal tonight.#US CPI unexpectedly rises to 3.7%, PPI faces another test tonight
US stock market closing summary: Apple holds steady at $5 trillion, CoreWeave's stock and bonds both crash, scaring Musk
The three major indexes all closed lower, with the Dow down 0.35%, the Nasdaq down 0.15%, and the S&P 500 down 0.02%. On the surface, it seems calm, but there are undercurrents beneath
In terms of trading volume, Nvidia fell 3.16%, with $25.9 billion in turnover ranking first. Jensen Huang said something harsh: if customers develop unsafe AI software, he will stop supplying. Micron Technology fell 4.08%, ranking second, as storage chips continue to be pressured by technological advances from China's Changxin Memory.
CoreWeave is the biggest bomb tonight. Its stock price plummeted 15.46%, bond yields soared above 15%, and the market prices its default probability over the next five years at more than 50%. Musk responded directly with "Scary" after seeing this. This AI cloud service provider is called the "subprime loan segment in the AI field" by analysts; if the AI financing wave recedes, it could go straight to zero.
Apple, on the other hand, is very stable. It closed up 0.5%, with its market value surpassing $5 trillion for the first time, becoming the second company after Nvidia to achieve this. SpaceX has also been active, releasing the so-called "world's smartest" voice model Grok Voice 3.0, which can autonomously handle 70% of Starlink's customer service calls.
In summary: the AI sector is undergoing intense differentiation. Apple and Meta, which have real substance, are rising, while CoreWeave, built on high leverage, is being abandoned by the market. #TSMC Q3 revenue hits a new high, profit performance yet to be revealed
Core CPI is higher than expected, oil prices break 105 again, combined with the upcoming midterm elections. ASML's earnings exceeded expectations but still couldn't stop the downward trend. It is estimated that everything will settle after the midterm elections; historically, November and December have been good months for the US stock market. Currently, BTC is indeed still very strong, but from a mechanism perspective, if US and Korean stocks continue to fall, it will inevitably lead to liquidation and margin replenishment demands. Once this action starts and triggers a stampede, a chain reaction of declines will follow.
This is the mechanism at play. I really don't know if it will be triggered, but if it were me, I would probably take advantage of BTC still being around 80k to sell and then buy the tech stocks that are about to create a golden pit. From 80k to 100k is just a 25% increase, but if tech stocks fall into a deep pit, you might get a double return. This math problem should be clear to many. It's just uncertain whether the fragile situation will be broken.
Of course, if you say the strong get stronger and BTC is ready to double to 160k, that's also possible, but it's a low-probability event. $BTC #USCPI unexpectedly rose to 3.7%, and tonight PPI faces another test Record top-line growth is a fantastic headline, but in high-capital tech manufacturing, volume alone doesn't guarantee bottom-line efficiency.
With TSMC’s Q3 revenue reaching a record ~NT$1.49T (up ~50% YoY), the top line is clearly being driven by booming demand in AI and advanced semiconductor nodes. However, the core focus now shifts to profitability execution:
- Margin Compression Risks
- Execution over Volume
#TSMCOct15Earnings The busier the project, the token isn't necessarily more valuable
In the past, when evaluating projects, ecosystem growth was often seen as a positive for the token.
Recently, this judgment has started to show problems.
Past consensus:
$ARB, $LDO, and $ONDO all involve governance.
On-chain transactions, staking scale, and growth of financial products were easily regarded as positives.
Current changes:
ARB's fees are collected in ETH, and LDO holders do not receive automatic dividends.
Profits from ONDO's products do not necessarily mean profits for token holders.
Business growth and token returns have never been the same thing.
The most direct impact is that no matter how impressive the project data is, it cannot directly infer an increase in token value. What truly deserves scrutiny next is whose account the new revenue ultimately goes into.
#StandardCharteredOpensBTCandETHTrading, BankExpandsCryptoBusiness $ARB $LDO Although the South Korean stock market triggered a circuit breaker, there are actually two pieces of good news today: one is that the probability of an interest rate hike in October has dropped again, and the other is that both long-term and short-term U.S. Treasury yields have fallen.
For a long time, besides the issues of aging themes and high valuations leading the market to start focusing on earnings expectations, the problem troubling the U.S. stock market has been the competition for liquidity among U.S. stocks, U.S. Treasuries, and corporate bonds.
Now, the decline in U.S. stocks will drive liquidity into U.S. Treasuries to ease the Treasury crisis, while gold, which is directly linked to interest rates, remains generally stable. Without a larger liquidity crisis, gold should maintain stability. Even if there is a short-term sudden liquidity shortage causing a drop, the support from central bank gold purchases keeps its safety high, making it a suitable low-risk bottom-fishing asset.
Tonight's PPI data will most likely continue the trend of yesterday's CPI. Even if it unexpectedly comes in lower than expected, the positive impact will probably only trigger a small rebound and cannot change the downward trend. After this decline, it will likely take the FOMC decision combined with easing U.S.-Iran tensions and other positive factors to gradually calm the market, so patience is still required. #South Korean stocks fell nearly 8% intraday, global tech stocks collectively under pressure Even $BTC itself can't escape, 100x short, entered at 83265, marked at 81171.9, floating profit +186.52%. Hundred times leverage shorting BTC, this round is basically "god slaying," but profits are just on paper until realized.
The market has completely turned into a fan: BTC failed to hold 84100 twice, volume is still shrinking, now hovering around 81100; ETH broke 2580 early from 2626, the rebound is all paper tigers. Altcoin liquidity is fleeing, even BTC's center of gravity is pushed down, this is either a bottom or just exhaustion causing a pause.
Macros: oil prices falling and ETF inflows can't support 84000, CPI expectations are fully priced in for tomorrow night, good news is easily priced out, bad news will hit hard, high probability of a spike and then a pullback. Geopolitics is eerily quiet, don't be complacent.
I have locked in some profits in batches on this trade, leaving a base position to watch CPI. Do you dare to short now? Quote your price in the comments. #USCPIexceedsexpectationsto3.7%,PPItesttonight #TSMCQ3revenuehitsrecordhigh,profitperformanceawaits $BTC $ETH $SOL
#Orbit Post｜9.54 Billion Liquidated in 24 Hours! Epic Long Liquidation
Latest heatmap data: 9.54 billion USD liquidated across the entire network in 24 hours, with 178,000 liquidations. 92% were long liquidations, shorts only accounted for 8%. This is a large-scale long squeeze caused by rising macro interest rate expectations.
By coin structure, ETH is the hardest hit this round, with 287 million liquidated, 92.4% of which were long liquidations; BTC follows closely with 197 million, 90.6% long liquidations. Other altcoins and second-tier coins combined liquidated 149 million, also 92.3% longs. ETH leveraged long positions are more concentrated and sensitive to interest rates; once macro negative news hits, the stop-loss cascade is much stronger than BTC.
The largest single liquidation came from Binance BTC perpetual long, amounting to 12.0347 million USD, indicating that large leveraged longs buying at the top were directly wiped out.
Market sentiment interpretation:
Previously, leveraged longs betting on inflation easing and positioning for a rebound were almost entirely liquidated; high-leverage bottom-fishing funds were completely discouraged, and the market is reluctant to easily leverage long positions in the short term;
Short liquidations are very few, shorts have not collectively taken profits and exited, indicating the short side pressure is not yet fully released;
BTC is relatively more resilient than ETH, highlighting its capital flight-to-safety attribute, but overall market confidence is severely damaged, and bottom-fishing willingness has cooled significantly.
Tonight's PPI data is a key variable; if PPI continues to exceed expectations, there will be a new round of long liquidations; if data falls back, it will usher in a window for sentiment recovery. Tesla's deliveries exceeded expectations, but don't rush to include profit recovery just yet.
On October 2, the company announced third-quarter deliveries exceeding 486,000 vehicles, while also revealing energy storage deployments of 13.7 GWh. I believe this latter figure should not be treated as just incidental news. When assessing profitability, besides how well the cars sell, we also need to see how much gross profit per unit the energy storage business actually retains.
The second quarter serves as a ready reminder: energy storage deployments increased, energy business revenue grew year-over-year, but gross margin declined. Shipment volume tells us how many products reached customers but does not explain how prices, costs, and related expenses changed. Directly converting GWh into profit, just like estimating profit solely based on car sales, risks making the same mistake.
When the earnings report comes out on October 21, I will look at the automotive and energy businesses separately. If gross margins improve on both sides, the basis for profit recovery will be more solid; if only the delivery numbers look good but profits are eaten up by expenses, then we need to patiently find the reasons. The long-term story of robots cannot answer these questions about current quarter operations.
New businesses are certainly worth looking forward to. But every time actual profits are under pressure, pushing the discussion automatically to a more distant future makes me a bit anxious.
How much profit have the already delivered cars and energy storage products actually made? After the report is released, let's first answer this question.
#Tesla deliveries exceed expectations, profit recovery becomes the focus $SOON SOON This time, I reviewed it again using the four-cycle matrix.
From 0.17 to 0.56, more than doubled in a month, now back to 0.29.
Many people ask if it has bottomed out — my conclusion is: not yet.
The reason is not the candlestick, but the chips.
▸ 15-minute wick rebound, volume, OI, and price all failed to pass the "rebound closed loop" verification
▸ Retail investors' long-short ratio is 1.42, slightly bullish, but large holders' net long-short ratio is -15.6% — retail is bullish, smart money is shorting
▸ Most importantly: 20.24 million tokens unlock in 8 days (10/23), accounting for 3.4% of circulating market value
In the last 6 unlocks, the average 14-day change was -16.9%
So at this position, the probability of the rebound being a "bull trap" is greater than a "reversal."
My habit of reviewing the market is: first fix the direction of the large cycle, then let the signals of the smaller cycles obey it.
Not to be led by a single 15-minute bullish candle.
The only signal truly worth waiting for is:
A volume surge stabilizing above 0.3105, while OI begins to expand positively.
Missing any of these three conditions does not count as a reversal.
— A market review without emotion#Coinbase integrates Deribit, can Q3 revenue improve? ASML's Q3 revenue hits a new high, and whether future demand will materialize is testing risk appetite. UNI, as a high-beta altcoin, cannot remain unaffected; I tend to believe the short-term weakness is not over yet.
From the capital perspective, the current price at 7.028 is down 6.1%, with a volume of 17.454 million and open interest of 5.699 million. The positive fee rate of 0.01% indicates bulls are still holding positions, but the 4-hour retracement from the high is already 23.59%. Although buyers dominate the 1-hour by 1.38 times, the rebound looks more like short covering rather than new long entries.
If it rebounds to 7.135, consider light short positions with a stop loss at 7.318 and a target of 6.695; if it falls to around 6.648 and buying pressure increases, try going long with a stop loss at 6.512 and a target of 7.012. Do not exceed 5% position size per trade.
— This is only a personal opinion and does not constitute investment advice. Wishing you successful trading. —
$UNI #TSMC Q3 revenue hits a new high, profitability performance to be revealed
#ASML Q3 revenue hits a new high, can future demand be realized?
$UNI 台积电又炸了。AI根本不是泡沫！连续第六个季度利润创新高。
三季度营收1.49万亿新台币，同比暴涨51%，净利润预计7463亿，同比飙65%。什么概念？一个季度赚的钱，比去年全年某些头部加密项目总市值还多。
但真正刺激的是：台积电现在等于AI世界的“印钞许可证”。
英伟达、AMD、苹果，管你谁家的AI芯片，最后都得排队进台积电的产线。更关键的是CoWoS先进封装——全球八成以上的AI加速器芯片全靠这条路线集成，而台积电几乎独占产能。这就叫卡脖子，只不过这次被卡的是硅谷巨头。
2nm制程从二季度的3%贡献直接拉到三季度11%，苹果和AMD抢着上。5nm以下先进制程营收占比冲到71%，四季度奔75%去。产品结构往高毛利方向死命迁移，定价权强到离谱。
台积电业绩是AI算力需求最硬的晴雨表。芯片代工端都在疯狂扩产，下游AI应用和算力网络的想象空间只会更大。链上AI、DePIN算力、去中心化推理这些叙事，底层逻辑正在被实打实的半导体数据一遍遍夯实。
台积电用财报证明了一件事：AI不是泡沫，是印钞机。问题是，你手里有沾边的筹码吗？#台积电Q3营收创新高，盈利表现待揭晓 $TSM CPI has been released, let's verify it against yesterday's projection. Last night at 20:30, the September CPI was published. First, let's present the actual data, then verify each item against yesterday's projection. Actual values (BLS official; expected values according to CNN report, listed separately): - Overall CPI: year-over-year +3.7% (expected 3.6%, previous 3.4%), month-over-month +0.6% - Core CPI: year-over-year +2.6% (expected 2.5%), month-over-month +0.3% (expected 0.2%) - Gasoline: month-over-month +9.1%, year-over-year +34.2%, contributing more than half of the month-over-month increase Conclusion: Above expectations, and core CPI also exceeded — this falls into the "above expectations" scenario projected yesterday. Verification against projection: 1. Did BTC come under pressure as projected? Yes. Before the release, BTC was around $83,860, and within 5 minutes after the release, it dropped to $82,731. Matt Hougan from Bitwise commented early after the data release that he originally thought it would drop 3-5%, but it actually only dropped about 1% — the decline was milder than some expected. 2. Was the support around 82000 effective? The 82000 level was breached, so this time it cannot simply be recorded as "support effective." Next time, the projection needs to clarify in advance: whether to judge support failure by intraday touches, closing position, or duration. 3. Reaction of US bonds and the dollar? After the release, US bond yields rose, and the dollarRevenue: Most likely to be flat, difficult to exceed expectations.
• The company guides revenue at about $12.86 billion, while the market consensus expectation is about $13.02 billion, approximately 1.2% higher than the guidance.
• Historically, Netflix's actual revenue usually fluctuates within ±0.5% of expectations (Q2 was $12.56 billion vs. expected $12.57 billion). If this expectation does not converge toward the guidance, revenue is more likely to slightly underperform.
Concerns:
• Viewing time in the first half of the year only grew about 2%, with growth mainly driven by price increases rather than user engagement.
• There was a one-time tax expense in Q3 last year, causing EPS to be significantly below expectations; such surprises have occurred before Q3.
• Profit margins are affected by the quarterly rhythm of content and marketing expenses.
My view: EPS is expected to beat by about 65–70%, revenue to beat by about 35–45%.
#Netflix growth may slow, advertising business becomes a new highlight PONS just said yesterday that the data rebounded, but I see the data shrank back a bit today.
Compared to October 12 and October 13, the latest daily ring ratios of platform fees and protocol income decreased by 30.83% and 35.98%, respectively.
With data performing so poorly, why did the coin price rise yesterday? I think it was more driven by news.
Yesterday, there was a rumor that Robinhood is preparing $100 million to support Robinhood Chain, which caused many ecosystem coins to take off. Whether it's true or not, I don't know, but quite a few. Crypto media picked it up again, which indeed stimulated the market.
But to be honest, improving core data is the key. You can't rely on news to drive buying every time; if it happens too often, the effect diminishes. It's like the boy who cried wolf, and the truth will eventually be revealed.
The effect was so good yesterday that I guess the data on the 14th might improve a bit. The market is too bad now, the overall market is like this, let alone any meme coins. #Circle earnings report is coming, can stablecoin growth deliver returns? Maji went from a whale to a small fish in ten days, and the cost was too high!
Once a big player in the crypto circle, he was forced to the edge of a cliff by high leverage. Data shows a total loss exceeding $36.57 million, with another $4.43 million lost in the past 7 days, leaving only $279,800 in the account. Currently, he is holding a full position of 2,100 $ETH long, with a position value of $5.07 million, opened at 2,515, actual leverage over 36x, unrealized loss of $200,000, and funding fees accumulated to $1.36 million. The most fatal is the liquidation price at 2,397, with available margin already zero, purely relying on market movement to survive.
Combined with the macro pressure of the US September CPI rising to 3.7%, ETH is fluctuating with the market. This kind of extreme naked long is like licking blood on the edge of a knife. From whale to small fish in ten days, high leverage amplified greed and accelerated destruction. If a whale ignores position management when bottom fishing, it can also instantly go to zero. Don't blindly trust big players' games; the survival rule is always light positions and risk control. $BTC $ETH $ZEC $57 million outflow in one week, are DOGE whales withdrawing or rotating?
Scene of the case: two conflicting clues.
Clue one: In the past week, $DOGE spot capital flow was a net outflow of $57 million, with the price dropping 8.9%, hovering around 0.085. Clue two: Around October 6, reports tracked whale addresses accumulating about $112 million in purchases.
One is going out, the other coming in. Who is telling the truth?
With over a decade of auditing experience, my take is: both are true, but they refer to different things. The $57 million outflow, scattered and rapid, is mostly small to mid-sized positions scared off by the market; the $112 million inflow, slow and heavy, is usually institutions or whales accumulating in batches at low levels.
This is called "rotation."
"Big cleanup" of chips
In financial markets, the healthiest rallies are rarely one-off; they come with intense volatility and cleansing. When the price hits the key support at 0.085, panic causes the less resolute holders to surrender their chips. This $57 million outflow is like removing sand from the foundation.
Meanwhile, the $112 million inflow is the cement quietly entering at this low level. Whales are not rushing to push prices up; they need enough space and time to gather sufficient chips. This "low-level absorption" behavior is often more noteworthy than a simple price rise.
0.085: The dividing line between bulls and bears #闪迪存储业务高增长能否延续？ 溜达鹅今天翻财报，被闪迪的数据惊到了。 2026财年第四季度（截至7月3日），闪迪营收$89.65亿，环比增长51%。GAAP净利润$69亿，稀释每股收益$43.97。Non-GAAP每股收益$39.25。 一个季度赚$69亿净利润，什么概念？相当于每天赚$7,600万。 更夸张的是增长速度。 环比增长51%，其中约1/3来自销量增长，2/3来自价格上涨。也就是说，闪迪不光卖得多了，还卖得贵了——而且客户还在抢着买。 为什么卖得贵还有人抢？因为AI。 AI数据中心建设正在创造对企业级NAND闪存存储的结构性需求。训练一个大模型需要海量数据存储，推理阶段需要更快的存储读取速度。AI模型越大，需要的存储越多。而NAND闪存，就是AI时代的"粮仓"。 闪迪的数据中心业务就是最好的证明。上一季度（Q3），数据中心收入$14.67亿，同比暴增645%。这不是增长，是爆炸。 而且这已经是闪迪连续第三个季度大幅超预期。内存短缺和供应紧张持续，价格一直在涨。客户不是在选哪家便宜，是在抢哪家有货。 闪迪自己对未来的预期更激进。 公司预计NAND市场规模在2027年将达到The kitchen stove hasn't been turned off yet, but the ledger is already on the table—168 billion in single-quarter revenue, a sharp 46% quarter-on-quarter increase. This pot of broth is boiling fiercer than any I've seen in my 20 years. But as a five-star veteran chef, I only ask one question: The broth is richer, but have the firewood costs been accounted for?
For the second consecutive quarter, the books have turned from red to green, with a gross margin just over 10% and positive quarterly cash flow. What does this mean in the kitchen? It means a dish has finally been made without losses, leaving a bit of sauce at the bottom of the plate. But any cook knows, a 10% gross margin is like balancing on a tightrope: rent, labor, gas—if any of these rise even slightly, this dish instantly turns into a loss.
More glaring is the turnover rate—its growth rate has slowed compared to the previous quarter. The storefronts keep expanding, but customers are eating slower, indicating the front of house is growing while the kitchen is getting clogged. Signature dishes can hold the scene, but what truly eats into profits are the ever-burning high-powered computing stoves and the two super spicy peppers continuously added: chips and data centers.
The ratio is the killer. Revenue is the main dish, computing power is the firewood in the stove, and chips are the chili peppers. The spicier the chili, the more the main dish’s original flavor is masked. What chefs fear most isn’t customers not coming, but the stoves multiplying and dishes getting pricier, until the whole table’s bill can’t even cover the gas.
Several Michelin-starred peers are fighting over the same batch of truffles, only to find the purchase price is already higher than the menu price—that’s the current state of this pot.
There’s another trick: flavored profits are plated with trimmed edges. The real ingredient losses going into the pot—R&D, depreciation, prepaid long-term contracts—aren’t written on the tasting menu. Good plating is skill, but customers judge by the spoonful they taste.
That so-called 2 trillion full set menu pricing, in October, invited tasters into private rooms, serving a flavored version, not the original. The pinch of salt that presses hardest on the tongue in the original is the ever-increasing computing power cost. Whether this price can hold depends not on how pretty the menu looks, but on this dish being served for three consecutive quarters, with turnover steady and broth not turning bitter.
The walk-in customers at the next table, following the orders, add more as they smell the aroma; the linked assets’ heat looks strong, but the cooking time is always controlled by someone else’s stove. Before picking up the chopsticks, first see clearly what’s in your bowl: is it the main dish, the seasoning, or that spoonful of super spicy chili? Eating chili as a meal will burn your stomach through.
This pot knows best: steadily serving a 10% gross margin and sustaining turnover is true skill; piling on firewood to create heat is smoke, not fragrance. #anthropicq3revenue16.8bAortic dissection is silent; the only sign is that the curve on the monitor sharply drops nearly eight points during trading.
On October 15, the heart of the South Korean market was bleeding on stage. The composite index plunged nearly 8% intraday, with the wound almost covering the entire anterior wall; Hynix and Samsung, the two main blood supply trunks, simultaneously went into spasm, the defibrillator pads (sidecar mechanism) were applied to the chest for the second time, the current was delivered, but the rhythm could not be restored.
The Nikkei 225 fell 4%, indicating ischemia had crossed the pericardium and spread to adjacent organs, with collateral circulation fully decompensated. The Nasdaq dropped 1.88%, meaning the distal embolus finally reached the left ventricle. Meanwhile, U.S. chip stocks declined for the sixth consecutive trading day—note, this is no longer an acute event. Acute ischemia causes pain and alarms; six days of continuous low perfusion is painless, causing myocardial fibrosis quietly, and by the time you hear it, it’s gallop rhythm.
Hot inflation data means systemic vascular resistance is up, afterload surges, and every ejection must fight higher pressure. Rising AI financing costs mean extracorporeal circulation perfusion flow is quietly reduced, but surgery continues and oxygen debt accumulates. Valuation is ventricular wall tension—the higher it is, the more oxygen consumed; storage price wars are aortic valve regurgitation, blood pushed out flows back, wasting part of each contraction’s work.
ASML’s equipment orders remain strong, which on imaging equals a coronary angiography report stating “main trunk patent.” But angiography looks at the pipes, ultrasound looks at the pump. Pipes are clear but the pump doesn’t move—the lesion is not in the vessels but in the myocardium itself. What the market is doing now is moving the probe from the vessel wall to the ventricular wall.
TSMC’s performance is the intraoperative frozen section cut by this scalpel. Only two things need to be distinguished: is it tachycardia—a rate-driven valuation reset, functional and reversible, adjustable with medication; or is the myocardium already necrotic—chip profit expectations systemically weakening, organic, requiring open-heart bypass surgery. Before the biopsy results, any large-scale operation is a blind puncture, and blind punctures most easily damage the conduction bundle.
The mapping target linked to Nasdaq and the chip chain now acts as a transplanted heart. Once the donor is ischemic, the recipient immediately rejects it; no one waits for pathology reports before starting anti-rejection drugs.
Price crashes are always just symptoms. What truly chills the spine is—the monitor alarms always come from the myocardium that still has electrical activity. #kospidropsnearly8%Money votes with its feet
Three signals from the funding side, none showing any favor. First, the funding rate: on 10/13 it was still +0.0062% long position fee, turned negative on 10/14, and today has slid to -0.0041%—shorts have started to consistently collect rent, the sentiment steering wheel has turned. Second, on the open interest side: today saw a net outflow of $86.91 million, total market contract open interest shrank back to $7.6 billion; over 7 days, the cumulative net inflow is only about $3 million, all the positions added during the rise were paid back during the fall, meaning a whole week of effort was wasted. Third, the US stock $BTC ETF net outflow in October has exceeded $450 million, just last week alone $679 million, about eight thousand coins, were withdrawn; the $65.4 million outflow on 10/12 came entirely from IBIT under BlackRock—if even BlackRock is pulling out, bulls still trying to bottom-fish here should first ask who is carrying their sedan chair. Oh, the seller is still shouting on TV that the bear market is over and $100,000 is back. Calls are free, withdrawals are real money, the market has already answered with the candlesticks. I just came across a post that made me feel a bit amused yet somewhat bittersweet. Have you ever had that feeling: your mainstream coins are all in the red, while altcoins seem forgotten in the corner? A friend shared his positions—BTC and ETH both showing floating profits, but DOGE and PEPE have been red all the way down. He joked that he treated the junk coins like lottery tickets, luckily with a light position. This statement is actually more worth pondering than the losses themselves. The market has a very smooth logical loop for junk coins. When BTC rises, it means funds feed the mainstream first, altcoins get drained, so their drop is expected. When BTC is sideways, funds are cautious, no one dares to move, altcoins have no support, so their drop is also normal. When BTC falls, panic selling hits junk coins first, altcoins fall even harder, which is still normal. Each statement sounds right, but combined they mean: regardless of rising, falling, or sideways, altcoins can't find their own reason to rise. Behind this is not just simple sector rotation, but an issue of emotional pricing power. When the market only wants to give a premium to BTC and ETH, risk appetite doesn't truly spread out. It's not that funds don't see altcoins, they see them but are unwilling to bear that uncertainty. Tokens like DOGE and PEPE rely on attention, sentiment, and marginal buying. Once these three contract simultaneously, the rebound becomes very light and short-lived. There is also a more bullish path: if BTC stabilizes and slowly raises its lows, ETH starts to follow, then the market will give altcoins some room to test and err again. At that time, the first to move are usually not the worst junk coins, but those with communities, liquidity, and narratives. Junk coin rebounds usually lag and are highly dependent onThe market collectively took a hit, but ZK quietly surged 15 points. How much gold content does V God’s "milk" really have?
Today, the entire crypto market is still paying off CPI debts, with total market cap dropping about 3.4%. Bitcoin is gasping around 80,000, yet the top gainer is ZK (ZKsync): it surged from 0.0148 to 0.0185 in 24 hours, a counter-trend +15%, quite a divergence.
The trigger is the new catalyst. V God publicly congratulated the zkSync Airbender team on X today, saying zkEVM’s "speed battle has been completely won," and the next focus should be entirely on security. This Airbender Prime achieved 99% proof within 10 seconds for 7,200 mainnet blocks on a single Blackwell GPU, a solid narrative.
My take: this is a typical "technical milestone + leader endorsement" emotional pulse. A few days ago, the founder proposed a buyback, setting the stage. It’s reasonable for funds to ride this wave. But be clear, its 24-hour volume is just over $2 million, market cap under $200 million, RSI at 75, price breaking above the Bollinger upper band. The market cap is small, easy to pump and dump, and chasing with this volume has very low tolerance for error.
Watch two signals: whether it can hold above the 0.018 upper band with volume instead of pulling back after a spike; and whether the ZK ecosystem has follow-up verification progress after "shifting to security." Relying on the hype of a single tweet usually doesn’t last long. DYOR, not investment advice.
#ZK #ZKsync #Layer2$ETH CPI data released, overall inflation exceeded expectations, core inflation rose simultaneously, and the Fed's excuse of "only relying on energy price increases" was completely torn apart. The market responded with a sharp drop; what should we expect next?
After the data came out, Bitcoin $BTC directly fell below 82,000. ETF funds were already flowing out, and now with tightening macro expectations, buyers dare not enter. If 80,000 doesn't hold today, the next support is 78,000.
Ethereum $ETH is the worst hit; it was already the weakest among the three. ETF continues to bleed, staking yields can't keep up with U.S. Treasuries, and institutions are not stepping in at the bottom. When Bitcoin weakens, it falls the hardest, and today it continues to decline, likely hard to stabilize in the short term.
Gold $XAUT also plunged, spot gold dropped nearly 2%. Core inflation exceeding expectations pushed up real interest rates, the dollar and U.S. Treasury yields both rose, and non-yielding gold can only take a hit in the short term. But gold quickly rebounded today because ETF allocation funds are still flowing in, and central banks have not stopped.
The biggest difference between it and Bitcoin is that Bitcoin's buyers are on-chain whales, while gold's buyers are central banks and sovereign funds. The latter's capital attributes and patience are on a completely different level. Hold your Bitcoin and Ethereum spot positions firmly, unload all short-term leverage. Gold is under short-term pressure but fundamentally stronger than Bitcoin. $BTC $ETH $ZEC #USCPI exceeds expectations, rising to 3.7%, tonight PPI faces another test #TSMC Q3 revenue hits record high, profit performance to be revealed 💸Capital Confession Session: Who's Really Buying, Who's Just Pretending to Be Excited?
$BTC: Contract net inflow in 24h is 194 million, spot +44.7 million, short-term both green. But don't rush to pop the champagne—weekly -2.8 billion, monthly -5.3 billion, 12h still -102 million. This 194 million is just a recent rebound in the last few hours, more like an oversold bounce test, not a reversal signal. 🩹
$ETH: Scene of schizophrenia. Contract +57.55 million, spot -110 million. Leveraged traders are betting on a rebound, holders keep running. Both sides fighting, big money lacks consensus. If spot keeps selling, contracts will eventually be dragged down. ⚔️
$SOL: Contract -16.2 thousand, basically flat; spot +33.39 million. Someone is quietly accumulating, no one dares to leverage. Weekly spot +20.06 million, the only mainstream to turn positive. Dollar-cost averaging logic > contract speculation. 🌱
Summary: BTC tests inflow, ETH longs and shorts clash, SOL spot quietly buys. Three ailments, one common point—weekly and monthly charts are bleeding, big money remains cautious. Although BTC is strong, funds are just testing the waters, not going all in. 🚨
Conclusion: You can watch the rebound, but don't rush to call a reversal. Money is testing the temperature, not jumping into the pool. 🧊#US September CPI YoY rises to 3.7%, core inflation exceeds expectations #TSMC Q3 revenue hits record high, profit performance to be revealed #TSMC Q3 Revenue Hits Record High, Profit Performance Yet to Be Revealed TSMC's Q3 revenue reached NT$1.49 trillion, a year-on-year surge of 50%, setting a new historical high. The full earnings report and briefing will be at 14:00 today, with profit and gross margin being the main focus.
But don't just focus on TSMC; this AI chip craze is actually tightly linked with our crypto space.
Recently, when semiconductors sneeze, Bitcoin catches a cold. Micron and Hynix fell, BTC slid directly from 83,000 to 81,600, and ETH dropped nearly 4%. Simply put, the money speculating on AI and crypto is from the same group—when risk appetite tightens, tech stocks and crypto assets get hit together.
TSMC's guidance tonight is essentially a "thermometer" for AI market sentiment. If the gross margin hits the 67% ceiling, it means demand for computing power is still burning hot, which is positive for crypto sentiment; conversely, if the guidance is conservative, funds may continue to withdraw from high-risk assets.
Keep an eye on 14:00; TSMC's earnings briefing could be the prelude to tonight's Bitcoin direction choice. $BTC $xTSEM #TSMC Q3 Revenue Hits Record High, Profit Performance Yet to Be Revealed TSMC delivers an impressive "half report card": Q3 revenue approximately NT$1.494 trillion, up 50.9% year-over-year and 17.6% quarter-over-quarter, setting a single-quarter record; September monthly revenue reached NT$511.857 billion, soaring 54.6% year-over-year. The numbers are explosive, but this is on the revenue side; profit details will be disclosed at the October 15 earnings call.
The record revenue is understandable, fundamentally driven by three combined forces: Nvidia, AMD, and other AI chips fully occupying advanced process capacity, with CoWoS packaging still a bottleneck; Apple's new A/M series supporting the high-end consumer segment; 2nm moving from trial production to mass production, starting to contribute to revenue. In short, customers are not bargaining over price but competing for capacity.
However, profit is the real test. The company’s own Q3 gross margin guidance is only 65%–67%, lower than Q2’s 67.7%. The reason is straightforward: early depreciation and yield costs during 2nm ramp-up dilute margins by about 3–4 percentage points; overseas plants in the US and Japan have higher costs, diluting margins by 2–4 percentage points long-term. So the market is really watching three numbers: whether the gross margin can hold at 65%, whether Q4 guidance will increase quarter-over-quarter, and the capital expenditure outlook for 2027.Today's breadth data is painful: Among 283 coins in USDT perpetual contracts, only 19 are rising, with a median of -7.32%. $BTC fell 2.74%, seemingly "resistant to falling," but actually it's just falling slower, not not falling.
My three actions today:
1. No adding positions. In a market with a median of -7%, "cheaper because it fell" are the most expensive four words.
2. Cut leveraged positions first. Leverage amplifies the median decline; -7% multiplied by leverage is not a correction, it's an accident.
3. Watch ETH, not BTC. ETH fell 5.43%, SOL fell 3.98%—if the decline continues, altcoins and the main market have an order; the main market is just lagging.
The lesson is not that the market must fall, but that in a down market, "holding" and "being held" are two different things. I don't hate losing money; I hate not having control over my positions.
On days like today with a median of -7%, whose position will you cut first? Hold BTC, or look at the ETH in your hand first? $BTC $SLX
Not looking good, not looking good at all!
The institution incubating this project is about to enter liquidation, so what's the point of playing this game?
Solstice (the project issuing the SLX token) was incubated by **Deus X Capital**.
Deus X Capital founded Solstice Labs in 2024 and provided it with funding and resource support; it is one of the projects incubated under its umbrella.
Currently, this institution has announced it will shut down and enter liquidation (planned to be completed by January 2027), but the SLX project itself was still incubated by it.
---
What this institution does:
Deus X Capital is a crypto and fintech investment company supported by a family office, classified as an institutional investor. It was established in 2023, manages about $1 billion in assets, and its main businesses include investment, project incubation, and related operations. It is considered a relatively legitimate professional institution in the crypto field.There are some reasons to hold, so let's ask again
If the reason to buy $PEPE is to wait for sentiment to return, I think we should admit that our judgment is about market attention.
You can't say it's popular when it doesn't rise, then suddenly change your tune when it falls, saying you should hold it as a long-term project.
I'm more concerned about whether trading can remain active after the hype dies down.
If more and more people are discussing it but the price still can't push up, don't rush to explain it as having more room ahead. After all, there is real money between attention and buying.
For $SUI, I watch out for an illusion: just because the project is progressing well, you think you don't need to watch the supply during the holding period.
It has a total cap of 10 billion tokens, but tokens are still gradually released as planned, so the circulating supply will change.
This doesn't mean that release necessarily causes a price drop, but with the same demand, the tradable quantity might differ.
Being optimistic about the ecosystem is one thing; deciding at what valuation to buy is another matter to consider separately.
For $AVAX, you need to be careful not to mistake staking rewards as price protection.
Its staking rewards come from new issuance, while network fees have a separate burn mechanism.
The number of coins increases, but the account value can still decline.
So I won't ignore whether the original buying reasons have changed just because it can generate yield. Rewards can be counted as income but can't answer why you still want to hold.CPI data released, overall inflation exceeds expectations, core inflation rises simultaneously, the Fed's excuse of "only relying on energy price increases" is completely shattered. The market responded with a sharp drop; what should we expect next?
After the data came out, Bitcoin $BTC directly fell below 82,000. ETF funds were already flowing out, and now with tighter macro expectations, buyers dare not enter. If 80,000 doesn't hold today, the next support is 78,000.
Ethereum $ETH is the worst hit; it was already the weakest among the three. ETF continues to bleed, staking yields can't keep up with U.S. Treasuries, and institutions are not stepping in at the bottom. When Bitcoin weakens, it falls the hardest, and today it continues to decline, likely hard to stabilize in the short term.
Gold $XAUt also plunged, spot gold dropped nearly 2%. Core inflation exceeding expectations pushed up real interest rates, the dollar and U.S. Treasury yields both rose, so non-yielding gold can only take a hit in the short term. But gold quickly rebounded today because ETF allocation inflows continue, and central banks have not stopped buying.
The biggest difference between gold and Bitcoin is that Bitcoin's buyers are on-chain whales, while gold's buyers are central banks and sovereign funds. The latter's capital attributes and patience are on a completely different level. Hold your Bitcoin and Ethereum spot positions firmly, unload all short-term leverage. Gold is under short-term pressure but fundamentally stronger than Bitcoin. $BTC $ETH $ZEC #USCPI exceeds expectations, rising to 3.7%, PPI test coming tonight #TSMC Q3 revenue hits record high, profit performance to be revealed The US dollar surged. Bitcoin was dumped. The bet on rate hikes doubled.
September CPI is 3.7%, August was only 3.4%. Core CPI is 2.6%, August was only 2.4%.
Both numbers are higher than expected.
The biggest increase was in gasoline, 9.1%. One person is bearing half of the increase.
The first thing to break isn’t the coin.
It’s the idea that "Bitcoin is only for risk speculation."
When I saw the gasoline number, my first reaction was to rush to fill up.
Then I found it was even more expensive than last week.
People are like that.
Alright, back to the main topic.
The Federal Reserve already raised rates by 25 basis points on September 16, now at 3.75%–4.00%.
Will they do it again? The next meeting is October 27–28.
Before the CPI, the market was betting about 20% chance of a hike. When the data came out, it became 47%.
In one trading session, it doubled.
The two-year US Treasury yield is 4.872%, up 7 basis points that day. The dollar also rose.
Look clearly:
Bitcoin pays no interest.
The two-year US Treasury pays you 4.87%.
That’s all there is to it.
Now comes the hardest part.
I’m a bit biased, but this round of decline is not the coin’s fault at all; it’s because everyone suddenly remembered interest rates.
So I wrote two consecutive articles about interest rates yesterday, because this is the key focus now.
A few days ago, those shouting "next stop 100,000" have all gone silent these past two days.
And people who usually don’t follow financial news opened accounts for the first time this week, checked deposit rates first, and then never came back. #TSMC Q3 Revenue Hits Record High, Profit Performance Yet to Be Revealed TSMC's Q3 revenue reached a historic high, mainly benefiting from the AI chip foundry boom. However, the market is more focused on profit quality: new process ramp-up and overseas factory construction have increased costs, raising doubts about whether the gross margin can hold at 65%-67% and if profits can match the revenue. Earlier AI expectations have been fully priced in, so funds chose to cash out before the earnings report, leading to a stock price pullback. #Netflix growth may slow, advertising business becomes new focus #TSMCOct15Earnings
TSMC's revenue record before October 15 had already been absorbed by the market. If we wait until the financial report day to see it as a positive surprise, the investment rhythm might have been delayed.
On October 8, the company had already announced September revenue data: a 54.6% year-over-year increase, and a slight 0.6% month-over-month decrease. This variance carries deep significance — the strong long-term growth trend does not necessarily mean every month must set a new record compared to the previous month. Minor monthly fluctuations are not enough to shake the big narrative of AI demand The market opened all green this morning 😵. The higher-than-expected CPI data caused a widespread pullback in the Crypto market. It's clear that it's being lumped into the same risk valuation framework as AI stocks. Sectors like RWA have dropped more than 6%. Ajian believes that what the market cares about right now isn't $BTC or $ETH or any particular token, but rather interest rates, financing costs, and whether high-valuation assets can hold up.
I've also emphasized before that in the current high-interest-rate economic environment, crypto asset valuations are walking a tightrope and are especially sensitive to market sentiment. Short-term funds love to target these types of assets first. If the token you hold long-term doesn't have enough income support, it's risky in this environment. #US CPI unexpectedly rose to 3.7%, and tonight PPI faces another test The Dow fell 0.09%, almost unchanged; the Nasdaq dropped 1.88%. Two worlds in the same trading day.
Putting these two numbers together makes it clear: money hasn't left the US stock market; it has shifted from tech stocks to healthcare and consumer staples. The information technology sector fell 3% in a single day, while healthcare actually rose 2.4%. Eli Lilly and Merck rose by three to four points, a typical defensive rotation.
The root cause is still oil. In the past ten days or so, attacks on ships in the Strait of Hormuz have clearly increased. On October 11, daily crude oil passing through the strait was only about 2.9 million barrels, compared to over 20 million barrels a few days ago. Brent crude briefly rose above $106, gasoline prices increased by more than 30% year-over-year, and the September CPI rose 3.7% year-over-year, exceeding expectations. The 10-year US Treasury yield surged above 5.2%, the 2-year yield approached 4.9%, and New York Fed's Williams also said another rate hike this year makes sense.
Assets with high valuations fear this combination the most. Nvidia fell 3.2%, Meta dropped 3.3%, the Philadelphia Semiconductor Index fell more than 10% in five days, and CoreWeave dropped 15% in one day.
I tend to believe that as long as oil prices remain above $100, any rebound in tech stocks can only be seen as a bounce. Today, we first look at TSMC's earnings report. If the AI demand story is strong enough, semiconductors may get a breather; if not, this round of adjustment may not be over. After that, the FOMC meeting on October 27-28 awaits.
The xdm tech position can be held for now, adding on dips. In this kind of market, less trading is better. The midterm elections are approaching, and major turning points may come for the US-Iran and Russia-Ukraine conflicts. #Saudi Aramco headquarters attacked, crude oil supply under focus This CPI move has dashed the previous optimistic expectations of "peak rate hikes." #USCPI unexpectedly rose to 3.7%, and tonight PPI faces another test
In the short term, the macro trend is pulling downward. Rising rate hike expectations, higher US Treasury yields, and a stronger dollar mean risk assets are generally struggling. BTC will not strengthen independently.
But in the medium term, there is long-term capital buying around 80,000. The ETF's single-day counter-trend inflow of 167 million is no joke. It shows that there are buyers at this level.
At this stage, the biggest taboo is trying to guess the bottom. Whether 80,000 can hold depends on upcoming PPI data and the FOMC meeting. Before the direction is clear, position sizing is more important than judgment. # Netflix Growth May Slow, Advertising Business Becomes New Focus
Three key variables truly determine the ceiling of the advertising business
1. Whether advertising ARPU can catch up with subscription ARPU
Currently, the per-user revenue of ad-supported plans is still lower than that of ad-free plans. The gap may only narrow if advertiser budgets, programmatic buying, and live CPM increase.
2. Whether live and highly engaging content can be continuously supplied
Live content like NFL, WWE, MLB has higher ad monetization efficiency and can boost user engagement, but it is costly and concentrated in scheduling, requiring proof of sustainability.
3. Whether AI advertising systems can improve delivery efficiency
Netflix is advancing AI-driven ad planning, creative adaptation, audience insights, and data clean rooms, which relates to whether advertisers are willing to increase budgets long-term.
Investment Judgment
Netflix's advertising business is worth tracking, but one should not focus solely on the slogan of "doubling ad revenue." More importantly, watch: ad plan ARPU, advertiser retention, live ad fill rate, programmatic share, and whether profit margins are diluted by content costs.
If these indicators improve, the ad business will shift from a "story" to "cash flow"; if not, it will only continue to support valuation rather than become a performance inflection point. With the release of this CPI data, many people have actually realized that the previously repeated market narrative of "inflation easing, rate cuts by year-end" is not so certain.
Everyone had assumed that the Fed's rate hike cycle was already over, so assets like $BTC and $ETH, which have no interest, could maintain their valuations in a high interest rate environment.
However, this time both CPI and core inflation exceeded expectations, effectively overturning that premise—the US Treasury yields surged immediately, and now even holding risk-free US Treasuries can yield over 5%, so institutions naturally don't want to stay in volatile, cash-flow-negative crypto; the continuous net outflows from ETFs are the most concrete evidence.
The chain of liquidations only accelerated the decline a bit; the real problem is that most of the current buyers are still old whales in the crypto space doing short-term trades, who won't hold for more than a few months before exiting.
In contrast, gold fell but quickly rebounded, supported by long-term allocations from central banks and sovereign wealth funds; they don't gamble on whether rates will rise next month. This kind of capital resilience is something the crypto space simply can't match right now.
Investment carries risks, decisions should be made cautiously, and the above analysis is only a market viewpoint sharing, not any investment advice.
#US CPI unexpectedly rises to 3.7%, tonight PPI faces another test TSMC’s record revenue figures were already public before October 15, so presenting them as an unexpected positive surprise on earnings day may be missing the bigger picture.
On October 8, the company reported that September revenue increased 54.6% year-over-year but slipped 0.6% compared with August. This highlights an important distinction: strong annual growth doesn’t necessarily mean revenue will rise every month.
#TSMCOct15Earnings # TSMC Q3 Revenue Hits Record High, Profit Performance Yet to Be Revealed
Why so strong?
● Advanced process fully loaded: 3nm accounts for about 30% of wafer revenue, 5nm about 33%, and 7nm and below total 77%; 2nm has also started scaling (about 11%), mainly benefiting from Apple's A20 Pro/M6 chips and AMD Venice server CPUs.
● Continuous investment in AI infrastructure: Next-generation AI processors like NVIDIA Vera Rubin platform and AMD MI450X are queued up competing for capacity, and CoWoS advanced packaging is also in short supply.
● Capital expenditure significantly increased: Full-year capital expenditure for 2026 has been raised from $52-56 billion to $60-64 billion, with 70-80% invested in advanced processes. #US CPI unexpectedly rises to 3.7%, PPI faces another test tonight Brothers, my personal analysis is that oil prices will surge to $100 before the midterm elections. First, let's talk about the US-Iran situation. Recently, Iran blocked the Strait of Hormuz, then the Houthi forces in Yemen attacked Iran, and Iran continues to stir up the Middle East situation. Iran's goal is very clear: to control oil as a powerful weapon. Excluding seasonal factors, global demand for oil is roughly the same every year, but now that Iran controls the oil, global demand cannot be met, which will definitely affect everyone's lives, and this impact will reflect on countries. To solve this problem, there must be US-Iran talks, but the US has no sincere intention to negotiate. Currently, the US is still trying to find solutions, but the root cause remains in the Middle East. So oil prices definitely won't come down anytime soon and may even hit previous highs because Iran will surely try to finalize the talks before the midterm elections. Otherwise, once the midterms are over, the "King" will definitely attack. So oil prices may not continue to rise, but they certainly won't fall.
Now about Russia and Ukraine, the war has lasted almost five years. Now, even if they want to stop, they can't. One side must admit defeat. If they wanted to negotiate, Russia and Ukraine would have done so earlier and wouldn't have fought this long. Ukraine definitely wants to negotiate, but Russia refuses. So Ukraine attacks Russia's energy facilities daily to pressure Russia through other energy-concerned countries. But Russia has no other strengths except stubbornness. So no one knows when oil prices will come down. This is my personal opinion and does not constitute investment advice $BTC #TSMCOct15Earnings TSMC's revenue record was already known to the market before October 15. Treating the same fact as a sudden positive surprise on the day of the earnings report might be a step behind the rhythm.
The company announced on October 8 that September revenue grew 54.6% year-over-year but declined 0.6% month-over-month. This contrast is quite interesting: strong long-term growth does not mean every month is higher than the previous oneFrom October 14 to 15, the crypto market experienced a sharp plunge. Bitcoin fell below $81,000, hitting an intraday low of about $80,500; Ethereum dropped below $2,400, once reaching approximately $2,317 on October 15, with a nearly 5% decline over the past week. Data from Coinglass shows that the total market liquidation exceeded $1.1 billion, with Ethereum long positions liquidated at about $330 million, the highest among all cryptocurrencies. Notably, this was not an isolated "crypto-only event." During the same period, the US Nasdaq index closed down about 1.9%, the S&P 500 fell about 1.1%, gold dropped below $4,100, silver plunged sharply, and South Korean memory chip giant stocks crashed suddenly. This time, crypto assets behaved more like "high beta assets" in the global synchronized deleveraging of risk assets. 2. Direct trigger: CPI exceeded expectations, renewed hawkish rate hike expectations Let's look at the data. The US September CPI released on October 14 showed a year-on-year increase of 3.7%, higher than the market expectation of 3.6% and above August's 3.4%, with a month-on-month rise of 0.6%. Core CPI excluding food and energy rose 2.6% year-on-year (expected 2.5%) and 0.3% month-on-month (expected 0.2%). Gasoline prices contributed more than half of the monthly increase, driven by rising oil prices due to Middle East tensions (Brent crude oil once rose to about $105). Why this CPI data is particularly critical. The key lies in the changed policy background: the Federal Reserve just completed its first rate hike of 2023 in September, and the market was previously debating "whether#台积电Q3营收创新高，盈利表现待揭晓
What I find truly worth looking at in TSMC's earnings report is not how much revenue has increased, but how long the AI business can continue to thrive.
The Q3 revenue has been announced, reaching NT$1.49 trillion, a year-on-year growth of about 50%, hitting a new high. This figure indicates that demand for AI chips remains strong, but it also means the market had already set high expectations for today's earnings report.
Next, focus on three areas: first, the gross margin—whether it can remain stable under the pressure of expanding 2nm process capacity; second, CoWoS advanced packaging capacity—whether AI chip orders can continue to support growth; third, the demand outlook for Q4 and 2027.
My view is that the AI industry cannot be easily underestimated for now, but we also shouldn't blindly chase the stock just because of performance growth. TSMC's good performance only proves strong demand in the supply chain, not that all AI concept stocks are worth buying.
Especially now, the market is trading not just on current profits but also on growth expectations for the next few years. If the earnings beat expectations but management does not provide stronger signals for future growth, the stock price may still surge and then fall back.
Tonight, I am more focused on management's statements about future orders and capital expenditures. What truly determines how far the AI market can go is not how much money is made today, but whether it can continue to exceed expectations next year.
What do you think? Will this TSMC earnings report continue to fuel the AI market, or is it a case of good news being priced in and funds taking the opportunity to exit?