Orbit Post Sitemap

#TSMC Q3 net profit exceeds expectations, profit hits a record high $LINK's market action is really awkward. The 9-day cooperation has landed, and Stellar's data service is officially up and running. Normally, this kind of news would be enough to push the price up, but instead, it dropped 6.8% in 24 hours and 7% over 7 days, sitting right at the 7-day low, with volume only 1.4 times the usual. Good news causing a sell-off boils down to two things: either the news was already priced in long ago, with smart money having exited after a 10% rise in 30 days and 46% in 90 days; or this asset doesn't react to news at all and only follows Bitcoin's moves. Today, Bitcoin fell below 81k, 180,000 people were liquidated, and the whole market is forced to reduce positions, so who has spare funds to buy on good news? I lean towards the first explanation. Real good news was already discussed in the market three weeks ago; those popping champagne when the announcement comes out are often the last to enter. Let's see if the 12.02 level can hold. For a rebound, we need to see volume pick up and prices recover. This current low-volume, slow decline can't hold. $LINK Predict Fun's single-day capture fees reached $313,000, surpassing $300,000 again in a single day after the World Cup final day, ranking fourth in BNB Chain protocol fees. Many prediction market platforms gradually lost activity or even ceased operations after the World Cup, but Predict Fun experienced a significant secondary growth. The core driving force behind this is the short-term price fluctuation market represented by BTC 5min, contributing over 60% of the fees. In the most recent full trading day, Predict Fun's Crypto 5min market set a new record for single-day trading volume, surpassing Polymarket comprehensively in the three core metrics of trading volume, unique trading users, and number of trades. Additionally, Predict Fun is also continuously growing in the 15min market, TradFi, and Esports sectors, with the 15min market particularly gradually competing for Polymarket's market share The market sentiment has definitely shifted. Big Brother Maji used to basically go all-in on longs, but this time his total contract position is about 59.76 million U, split directly into three parts—short BTC, short HYPE, and long ETH. He's playing a left-right ambidextrous game. Breaking it down more clearly: On the BTC side, a 40x short opened at 84070, with an unrealized profit of about 220,000 U, and funding fees actually earned 56,600 U, so the account looks favorable; but with 40x leverage, even a small price bounce can hit him hard, so don’t just focus on the unrealized profit. On the ETH side, a 25x long opened at 2541, with an unrealized loss of 340,000 U. The market is weak but he’s stubbornly holding the long position—either betting on a rebound or using it for portfolio hedging. That strong spike this morning, I seriously suspect it was aimed at his position. HYPE is the harshest: a 5x short opened at 65.8, with an unrealized loss of 4.21 million U, funding fees collected 1.03 million U, which helps ease the cost, but price volatility doesn’t show mercy. So this is not a full bearish stance; it’s clearly a directional differentiation between coins. Going forward, watch three things: whether BTC can stabilize, whether ETH can rise, and whether this HYPE short can hold up. You can observe the smart money’s positions, but don’t copy blindly—what you should ponder is why they are positioning this way, not just follow because they buy.The storage sector surged fiercely some time ago, and now it is falling just as sharply. The growth in AI demand is a fact, but stock prices have already priced in expectations prematurely, along with the impact of risk-free assets. SanDisk's earnings report on October 29 is approaching, and the market has already started to reprice. I choose to continue holding short positions, bearish in the short term, but that does not mean I am pessimistic about the storage industry in the long run. No matter how good the performance is, it depends on how much valuation the market has given in advance. #Can SanDisk's storage business high growth continue? Where is the Core public chain headed? Screenshot from OKX on October 15 at 17:02: CORE/USDT perpetual at 0.01798, down 2.49% in 24 hours, lowest at 0.01712, trading volume 1.5581 million USDT. EMA5 0.01812, EMA10 0.01839, EMA20 0.01878—all three lines are above the price, with the short-term moving average below the long-term moving average, a classic bearish setup; MACD's DIFF is -0.00047, histogram -0.00018, still below the zero line; open interest shrank to 1.29 million USD, already below the sub-chart scale of 1.59 to 1.92 million. The only line still worth watching below is the 0.01712 just hit today. I went long at 0.02088, which is exactly that upper resistance line on the chart. Nominally down 13.9%, which translates to about -69.4% of principal on 5x leverage, excluding funding fees; the profit/loss on screen is -195.93. To get back to 0.02088, it needs to rise 16.1%. The position is more critical than the drop. btc126 wrote today that the distance from the historical low is 8.80%, that low was on July 28 at 0.0167. Three things I haven't figured out: On December 22, 2025, the Core Foundation released the 2026 roadmap, stating that repurchases of CORE would be funded by income from AMP, BTC LST, SatPay, ETF/ETP, but no funds have flowed back yet. Actual annual inflation is 35.73%, while the protocol's target is 3% to 5%. During the September validator over-issuance event, the hard fork was not rolled back, five exchanges suspended deposits and withdrawals, the official promised a technical review which has not been made public to date, and on October 1 block production was handed over to an independent operator. Three questions for everyone: When will verifiable on-chain data for repurchases be available? If it breaks below 0.01712 to test 0.0167, do you call that a risk or an opportunity? How much trust is left in an unexplained over-issuance? I personally went long at 0.02088; I won't discuss the next steps here. All data have sources and timestamps; this does not constitute investment advice. Contracts involve leverage and liquidation risks.#TSMC Q3 net profit exceeds expectations, profits hit a record high TSMC Q3 revenue hits a new high, what other highlights are there in the October 15 earnings report? Do you think TSMC can break records again with this earnings report? Latest data shows TSMC's Q3 revenue soared to 1.49 trillion TWD, approximately $46.7 billion USD, a year-on-year increase of 51% $TSM Market forecasts net profit to reach 746.3 billion TWD, which if realized would mark six consecutive quarters of record-high profits. In this AI boom, it remains the steady water seller on the fishing platform. Compared to Q3 results, two variables attract more attention: ▶️ Gross margin and guidance Last quarter's gross margin was 67.7%, with official estimates for Q3 between 65% and 67%. Against the backdrop of 3nm capacity being snapped up by Nvidia and Apple, watch if management will raise full-year revenue growth guidance and increase capital expenditures. ▶️ The Texas fab drama Elon Musk proposed the Terafab chip park plan, with TSMC possibly only leasing. The Arizona fab is still unresolved, so moving to Texas is not urgent; it seems more like a long-term layout post-2030 or a geopolitical bargaining chip. Compared to chip manufacturing, the real bottleneck is advanced packaging CoWoS; packaging and testing capacity can't keep up, so AI chips still can't be delivered on time. Looking ahead: ▶️ Short-term caution for potential profit-taking Even if data is strong, if gross margin guidance slows, the stock price may experience short-term volatility and pullback. ▶️ Long-term moat is very deep Tech giants' self-developed chips still rely on TSMC's technology and yield in the short term; in the coming years, it remains the biggest beneficiary of the AI dividend.This market is really brutal and heart-wrenching! $SKHYNIX has been plunging these past two days, dropping nearly 15% in 24 hours, now barely holding around $939. After last night's CPI data release, it just tanked like it took a laxative. The worst hit are the big players! From last night until now, three major whales got absolutely hammered, closing nearly $3.88 million worth of long positions combined, with actual losses hitting $887,000. Some cut losses actively with tears, others tried to hold on but got forcibly liquidated in batches. Now all three accounts are wiped out and gone—it's a bloodbath! But here's the interesting part. This plunge pushed the funding rate into negative territory for a while, indicating panic selling was rampant. Yet this morning, bottom-fishing funds rushed back in aggressively, flipping the funding rate from negative to positive in just two hours—intense battle going on. Looking at the liquidation chart, there's still over $5 million worth of long liquidation exposure lying around $888, just about 5.5% below the current price. If it can't hold and dips further, expect another chain reaction of liquidations. The long-side risk is far from over.$ETH $BTC 🔥 10.15 Evening Report On the same day, $BTC $ETH $SOL all weakened together. But the reasons behind the three bearish candles are different. Different reasons for the decline: $BTC is driven by sentiment selling, no large holders fleeing on-chain. There is still support near the 400-day moving average. $ETH is a capital withdrawal: ETF net outflows continue, wallet balances are moving out. Some are actively reducing positions, not just following the drop. $SOL is pressured by contracts: Perpetual shorts are increasing, prices pushed down by leverage. Spot outflows are limited, no large-scale chip turnover. Just looking at the order of decline can easily mistake sentiment for trend. Distinguishing the source helps identify where the noise is. #US CPI unexpectedly rose to 3.7%, tonight PPI faces another test CPI一热，全场挨锤。$ZEC 、$SOL 、$BTC 都在跌，但跌法完全不同。 $ZEC 摔得最重。24h跌12%，现价1096，持仓量缩13.6%，1900万清算几乎全是多单。资金费年化近10%，杠杆太拥挤，一反转就踩踏。1000是最后防线。 $SOL 被资金面拖着，基本面却最扎实。现价102.3，24h跌5%，ETF连卖7天。Q3应用收入3.65亿，同比增42.6%；10月21日Alpenglow上主网，确认时间从12.8秒压到150毫秒。破100看90。 $BTC 最能扛。现价80927，比ZEC少跌10个点，鲸鱼持仓303万枚，创近一年新高。但美元现货在撤，ETF上周流出6.8亿。跌破80000看78000。 三者本质：ZEC在还杠杆债，SOL在等资金回，BTC是资金没走净。今晚还有一场，继续跌还是止住？边走边看。#BTC现货ETF再度净流出，机构买盘能否回暖？#美国CPI超预期升至3.7%，今晚PPI再迎考验 #台积电Q3净利润超预期，盈利创历史新高 #奈飞增长或放缓，广告业务成新看点 #TSMCQ3RecordProfit TSMC just delivered another record quarter. The margin story is more nuanced. TSMC reported Q3 revenue of NT$1.494T, up 50.9% YoY, and net profit of NT$763.06B, up 68.7%. Net profit set a quarterly record and beat the NT$746.34B LSEG SmartEstimate cited by CNBC. The technology mix shows where demand is moving: · 2nm reached 8% of wafer revenue, up from 3% in Q2 · 3nm contributed 33%, followed by 5nm at 28% and 7nm at 10% #TSMCQ3RecordProfit $TSM $xTSM#台积电Q3净利润超预期，盈利创历史新高 台积电Q3财报出来了，数字炸裂，但股价跌了。或许预期太高，人们已经透支了股价。 营收同比增长50.9%到1.49万亿新台币，净利润同比增长68.7%到7631亿，创季度新高，超过市场预期。2nm制程占晶圆营收的8%，先进制程整体占比79%。AI芯片需求持续为业绩提供支撑，这个基本盘很硬。 但股价跌了2.67%。问题出在毛利率指引上。公司给的65%到67%，低于Q2实际的67.7%。市场担心的是，AI芯片再旺，如果毛利率往下走，说明先进制程的成本压力在吃掉利润。另外资本开支会不会进一步上调，也是接下来要盯的。#Strive increased its BTC holdings by 2,822 coins in two days, with total holdings exceeding 35,000 BTC $BTC In eight weeks, it grew by 62%, while the largest holder increased by less than 1%. ▪️ Added 12,527 BTC from 8/14 to 10/9, a holding growth of about 62% ▪️ During the same period, Strategy's holdings rose from 840,447 to 848,250 BTC, an increase of less than 1% ▪️ Increased by another 2,822 BTC in two days, with holdings exceeding 35,000 BTC ▪️ Metaplanet sold 10,000 BTC in Q3 but bought back 11,000 BTC, a net increase of 1,000 BTC ▪️ Twenty One has not increased holdings since July 2025, and MARA has remained basically flat since May The divergence is not about who holds the most coins, but about "holding rankings" in terms of stock volume and "growth rankings" in terms of incremental volume. 840,447 to 848,250 reflects changes in stock volume, while 12,527 versus 20,246 reflects changes in incremental volume—two completely opposite rankings within the same industry. The direction is to watch — small-cap companies continuing to absorb incremental volume is the real trend; if the large-cap company accelerates again, the pattern returns to stock volume.🔥盘面解读｜韩股重挫，存储板块承压，今晚美股开盘预判 配图核心信息：韩国KOSPI盘中大跌超6%，触发Sidecar程序化卖盘暂停机制；SK海力士大跌超8%、三星电子跌近7%，连带日经225跌超3%，日韩芯片股集体走弱。$SNDK $MU ✅事件逻辑 韩股本轮杀跌核心就是半导体存储板块，市场担忧长鑫存储技术突破带来长期供给冲击，叠加通胀、利率预期扰动，资金集中抛售存储芯片标的。三星、SK海力士作为全球DRAM两大龙头大跌，会直接给美股存储股（美光MU、闪迪SNDK）带来情绪传导压力。 📊今晚美股开盘推演 1. 盘前情绪：存储板块大概率低开承压。日韩存储巨头大跌，属于板块负面情绪外溢，美光、闪迪开盘容易受拖累。但要区分：这是远期竞争预期利空，并非企业当期业绩暴雷。 2. 分化机会 美光MU：业务覆盖DRAM，受韩股下跌联动冲击更强，波动更大。 闪迪SNDK：主营NAND闪存，和DRAM赛道无直接竞争。若资金理性区分赛道，后续有机会走出相对抗跌行情。 3. 大盘变量 不能只看韩股，美债收益率、今晚PPI数据、纳指整体情绪，影响力同样很大。如果美债收益率回落，会对冲一部分芯片板块的恐慌。TSMC's Q3 financial report is out, with explosive numbers, but the stock price fell. Revenue grew 50.9% year-over-year to 1.49 trillion New Taiwan Dollars, net profit increased 68.7% year-over-year to 763.1 billion, setting a new quarterly high and exceeding market expectations. The 2nm process accounts for 8% of wafer revenue, and advanced processes overall account for 79%. AI chip demand continues to support performance, this fundamental base is very solid. But the stock price dropped 2.67%. The issue lies in the gross margin guidance. The company gave a range of 65% to 67%, lower than Q2's actual 67.7%. The market worries that even if AI chip demand remains strong, if gross margin declines, it indicates that cost pressures in advanced processes are eating into profits. Additionally, whether capital expenditures will be further increased is something to watch next. For BTC, this is not direct, but the underlying logic is connected. TSMC is a core supplier of AI computing infrastructure; its revenue and profit validate that AI demand is real, and the narrative of the computing power economy remains intact. BTC's long-term logic as the underlying anchor still holds. However, the weak gross margin guidance indicates that the cost of burning cash is rising, affecting the entire semiconductor chain.$SKHYNIX SNDK $SKHYNIX Hynix starts to consolidate while SanDisk plunges sharply!!! Institutions begin to support Hynix, so Chai didn't dare to short it, but this time Chai plans to short SanDisk hard. Recently, Chinese manufacturers have substantially overtaken; Yangtze Memory has completed overtaking in the NAND market, and ChangXin Memory's breakthrough in the DRAM field has caused panic across the entire sector, coupled with many institutions lowering target prices, leading to a brief loss of market confidence in SanDisk. Additionally, continuous insider selling at SanDisk and CPI data exceeding market expectations, the dollar appreciating, and tech stocks collectively under pressure. In the short term, Hynix has triggered a protective cushion mechanism; institutions start to support the stock, and retail investors and whales begin to look forward to extra shareholder returns after the Q3 earnings release. SNDK current price is suitable for entering short positions: 1348, first target 1322, stop loss 1400. The above is personal sharing only and does not constitute investment advice. #TSMC Q3 net profit exceeds expectations, profits hit record highs #US CPI exceeds expectations, rising to 3.7%, PPI test tonight $NEAR 永续50倍空单，5.075开，现4.563，浮盈+504.43%。 思路很朴素：顶部横盘缩量到极致，波动率压到地板，说明筹码准备派发。一根放量阴线直接把价格从 5.075 砸下来，典型的破位信号，追空不追多。50倍杠杆，止损 5.2。走势一路向下，不给任何舒服的下车位，盈利直接起飞。 这位置我打算先减一半仓位落袋，剩下的一半止损提到 4.8 让利润跑。4.2 能放量跌破就继续拿，跌不破就全平走人。$BTC $ETH #台积电Q3净利润超预期，盈利创历史新高 Netflix guided Q3 revenue growth of 12% and an operating margin of 33.2%. The year-ago margin was 28.2%, but that quarter included a $619M Brazil tax charge, so the comparison flatters it. A beat or miss hinges on ads. #NetflixQ3AdsAndMargins 最脆弱的一环从来不是方向，是仓位太挤之后大家都不肯先走。 你猜现在最难受的是空头还是追高的人？ 刷到Maji大哥的永续持仓更新，我盯着那组数字看了很久。整体还是BTC空加ETH多再加HYPE空的组合，但账面盈亏已经跟着行情撕成两半。 - BTC空单四十倍全仓，持仓243.96枚，名义约1994万U，开仓84070，未实现盈利约56.87万U，资金费还收了6.41万U，爆仓价104380附近。 - ETH多单二十五倍全仓，持仓7222枚，名义约1743万U，开仓2541，浮亏约92.27万U，爆仓价1623.5。 - HYPE空单五倍全仓，持仓26.83万枚，名义约2168万U，开仓66.24，浮亏约390.94万U，但资金费收入约104.2万U，爆仓价99.76。 这组仓位其实在讲一件事。钱现在更愿意奖励看空BTC的人，却还在惩罚看空HYPE的人。前者顺着宏观避险和风险偏好收缩的方向，后者撞上了叙事和买盘的惯性。ETH夹在中间，既没跟上BTC的疲软，也没吃到HYPE那种情绪溢价，所以它的多单成了拖累。 我更在意的是情绪层面。HYPE这笔浮亏已经接近390万U，换成一般人早就砍了，但大哥#台积电Q3净利润超预期，盈利创历史新高 ⚠️Risks to watch 1. Slight sequential decline in gross margin, mainly due to higher costs during the ramp-up phase of the 2nm new process, which drags down gross margin in the short term ​ 2. Heavy reliance on major AI clients; if AI capital expenditures cool down in the future, orders will face pressure ​ 3. Overseas factory construction (USA, Japan) has high costs and long build cycles, which will continue to increase capital expenditure pressure ✅Summary in one sentence AI computing power demand continues to explode, driving TSMC's advanced process orders to be full and pushing profits to a record high for six consecutive quarters; 2nm mass production is proceeding smoothly, maintaining technological barriers, but high capital expenditure and AI demand fluctuations are the two major variables going forward.#台积电Q3净利润超预期，盈利创历史新高 台积电Q3营收创新高，10月15日财报还有哪些看点？ 你觉得台积电这次财报还能再刷纪录吗？ 最新数据看，台积电三季度营收飙到1.49万亿新台币，折合约467亿美元，同比大增51% $TSM 市场预测净利润达7463亿新台币，真落地就是连续六季盈利创新高。在AI这波行情里，它依然是稳坐钓鱼台的卖水人 相比三季度成绩，大家更关注两个变数 ▶️毛利率与指引 上季毛利率67.7%，官方预估三季度在65%到67%。在3纳米被英伟达和苹果抢爆背景下，看管理层会不会上调全年营收增速和追加资本开支 ▶️得州建厂这出戏 马斯克抛出Terafab芯片园区计划，台积电或只是承租。亚利桑那州厂还没搞定，跑去德州并不紧迫，像是2030年后的远期布局或地缘博弈筹码 相比芯片制造，真正卡脖子的是先进封装CoWoS，封测产能跟不上，AI芯片照样交不出货 后期预计 ▶️短线谨防利好出尽 即便数据拉满，若毛利率指引放缓，股价短期可能震荡回踩 ▶️长期护城河极深 巨头自研芯片短期离不开其技术与良率，未来几年它仍是AI红利最大受益者 你觉得它会去得州接马斯克的盘，还是稳扎稳打搞好CPI一出，很多人第一反应：物价又烫了，黄金该闪亮登场了吧？✨ 现实反手一巴掌：金价跳水，1.65%瞬间蒸发，4096失守，4088都摸到了。😵 BTC却像没事人，81000—83000区间散步。🚶‍♂️ 逻辑不复杂：通胀顽固→美联储不敢松→高利率续命→美债收益率往上窜→黄金持有成本变贵→资金撤退。📉 黄金的剧本早改了：从前是“通胀来了我涨”，现在是“美联储鹰了我就跪”。🦅 说到底，黄金、BTC、美股，看起来各玩各的，其实都在押注同一道题：流动性松不松。💧 别跟K线较劲，先看央妈脸色。市场没变疯，只是换导演了。🎬#台积电Q3净利润超预期，盈利创历史新高 台积电Q3营收再创新高，但毛利率环比回落是本次财报最大看点。AI需求景气确定，但大规模扩产带来折旧压力，正在开始侵蚀毛利率。市场现在早已消化Q3的高营收，全部押注Q4营收指引与资本开支计划。 资本开支上调，远期利好产能、巩固先进制程垄断地位；短期代价就是利润率承压。马斯克Terafab洽谈代工属于远期题材，短期无业绩贡献。接下来财报会议，重点盯三件：Q4营收指引、资本开支上调幅度、管理层对未来毛利率中枢的表态。阿美总部遭袭，新闻让人紧张，但先别把办公楼受损写成原油已经减产。 美联社10月14日援引地区官员称，达兰的阿美总部遭打击，高层办公室所在楼层受损，未报告人员伤亡。这个消息说明核心企业的办公设施也受到威胁，可它本身没有告诉我们，油田产量或出口装船减少了多少。 我觉得这时候最该分开的，是已经发生的损失，和市场担心下一次袭击的风险。前者要靠生产、炼厂和港口运行信息核对；后者可能在尚未停产时，就让买家更担心交付。这两层都影响价格，但不能用同一组叙述替代。 如果后续确认生产照常，供应缺口就不该凭想象扩大。可下一次呢？连续袭击会让人继续追问，关键设施还能不能安全运转。风险仍在，具体损失却需要另外核实。 我希望伤亡和破坏都能尽快停止，也不愿把每一次袭击都当成刺激行情的素材。接下来会更留意阿美的运行说明，以及实际出口有没有改变。做能源判断，承认“目前不知道减产多少”，比为了一个鲜明观点随手填补证据更负责。总部受到打击已经够严重，不需要再给它添加尚未证实的供应损失。 #沙特阿美总部遇袭，原油供应受关注 #台积电Q3净利润超预期，盈利创历史新高 Q3净利润确实炸裂，但股价反而跌了2%多，典型的“利好落地即利空”。市场跌的不是现在，而是Q4毛利率指引（65%-67%）低于Q2。AI需求是确定的，但2nm爬坡和海外建厂的成本开始显现，加上资本开支可能进一步上调，资金担心利润率见顶。 短线是估值消化，长期看AI结构性逻辑没坏。我选择等回调企稳再考虑，不追高。 大家怎么看？毛利率下滑是短期扰动还是长期趋势？ #TSMC Q3 net profit exceeds expectations, profit hits record high ⚠️ Risks to watch 1. Gross margin slightly declined quarter-on-quarter, mainly due to high costs during the ramp-up phase of the 2nm new process, which drags down gross margin in the short term ​ 2. Highly dependent on major AI clients; if AI capital expenditures cool down in the future, orders will face pressure ​ 3. Overseas factory construction (USA, Japan) is costly and has a long build cycle, which will continue to increase capital expenditure pressure ✅ Summary in one sentence AI computing power demand continues to explode, driving TSMC's advanced process orders to be full and pushing profits to a record high for six consecutive quarters; 2nm mass production is smooth, maintaining technological barriers, but high capital expenditure and AI demand fluctuations are the two major variables going forward.$BLUR 链上全是大额出货，这波空单我拿定了 兄弟们，绝大部分人都在看多，那咱们就反着来，看空！链上数据已经把底牌亮出来了。 据Arkham监测，今早一笔高达9857万枚BLUR（约400万美元）从匿名地址直接转入Coinbase Prime。这是大资金明牌出货。做市商Wintermute也没闲着，累计向OKX、Binance、Coinbase转入了约260万美元的BLUR。做市商往交易所转币，从来不是为了护盘，是准备卖。 最狠的一刀来自韩国交易所Upbit，已经把BLUR加入交易警告名单，并暂停了充提业务。这套流程老玩家都懂，接下来大概率就是下架。一旦下架，流动性瞬间枯竭，想跑都跑不掉。 链上数据比K线诚实。我0.0127开的空单继续拿着，防守放0.0132上方，下方先看0.012，跌破就是0.011。剩下的交给时间。 $BTC $ETH #美国CPI超预期升至3.7%，今晚PPI再迎考验 TSMC just delivered another record quarter. The margin story is more nuanced. TSMC reported Q3 revenue of NT$1.494T, up 50.9% YoY, and net profit of NT$763.06B, up 68.7%. Net profit set a quarterly record and beat the NT$746.34B LSEG SmartEstimate cited by CNBC. The technology mix shows where demand is moving: · 2nm reached 8% of wafer revenue, up from 3% in Q2 · 3nm contributed 33%, followed by 5nm at 28% and 7nm at 10% #TSMCQ3RecordProfit #台积电Q3净利润超预期，盈利创历史新高 台积电这份Q3财报看着亮眼，利润创了季度新高，但毛利率不如上个季度。利好提前被市场消化，财报一出来资金就跑路，股价直接收跌，所以我觉得后续得看AI芯片订单能不能稳住。聊一下今天“预测一哥”的事情，我跟当事人并不熟悉，所以下面论事不论人。 其实很简单就是有些事情“不上称四两重，上了秤千斤都压不住”。预测市场在很多国家的法律都是灰色地带，不光是国内。而且大部分国家都把预测市场视为赌场或者菠菜的变种，基本都要有相应的牌照。 有时候自己私下玩玩基本不会有人管你，但是真要是搞社区，拉人去玩就会出问题。都说预测市场有技巧，但是99%人做不到稳赚不赔，赚钱了大家一团和气，赔钱了呢？输多上头以后可能真的会“点”了你这个拉他去玩的人。 刑法303条就是关于赌博的： 《刑法》第303条第1款：赌博罪 是以营利为目的，组织多人赌博或以赌博为业。 《刑法》第303条第2款：开设赌场罪 是经营赌场、控制赌博平台、做代理接受投注，或参与相关利润分成。 《刑法》第303条第3款：组织参与国（境）外赌博罪 是招揽中国公民去境外赌场赌博，达到法定严重情节。 所以自己玩玩没啥，拉人去玩还是要三思。因为我的简介中也写着预测市场，也有不少项目方找我做返佣分成，我是从来不做，有些钱不能挣，毕竟安全第一。1. AI算力仍是主引擎
高性能计算平台营收占比68%，环比增长22%，是增长核心。AI需求已经从GPU/加速器延伸到AI代理CPU、网络芯片，客户仍在抢产能。 2. 先进制程占比继续提升
7nm及以下先进制程合计占79%。其中3nm占33%、5nm占28%、2nm占8%。2nm占比从上季度的3%提升到8%，说明客户向更先进节点迁移正在加速。 3. 订单能见度延伸到2027年
台积电预计Q4营收523亿—543亿美元，中值约533亿美元，明显高于市场预期的503.5亿美元；全年美元营收增速指引从“略高于40%”上调至约44%。 也要看到压力点 毛利率环比略降。 从67.7%降到66.4%，主要原因是2nm初期量产摊薄利润，海外扩产也会带来成本压力。 资本开支极高。 2026年资本支出维持600亿—640亿美元，创历史新高；市场预计2027年可能进一步升至750亿—900亿美元。扩产强度前所未有，短期会压制利润率。 #台积电Q3净利润超预期，盈利创历史新高 #TSMCQ3RecordProfit TSMC's revenue has already exceeded expectations, so what is the Thursday earnings call really about? TSMC's Q3 revenue was actually announced early, about $46.7 billion, higher than the company's own forecast of $44.6 to $45.8 billion. The Thursday earnings call isn't about revenue, but about the costs behind that revenue. The first hurdle is gross margin. The Q3 guidance is 65% to 67%, with a midpoint of 66%, which is 1.7 percentage points lower than Q2's 67.7%, due to#奈飞增长或放缓，广告业务成新看点 1$NVDA $AAPL From a comparative perspective, NVIDIA’s current share buybacks primarily serve to offset the dilution caused by employee stock-based compensation, making it difficult to consistently thicken EPS through buybacks alone, unlike Apple. If NVIDIA were to emulate Apple’s mature capital allocation strategy by consistently using its substantial, stable free cash flow for secondary market buybacks and subsequent share cancellations—thereby achieving genuine net share count shrinkage—its EPS would receive sustained additional accretion, assuming its AI computing foundation remains solid. This would significantly boost market confidence in its valuation and serve as a key incremental driver for valuation expansion, potentially paving the way for NVIDIA to challenge a $10 trillion market cap. For young investors, Apple-style stocks should undoubtedly be the core holding in your primary portfolio position.$BTC 🔥 10.15 晚报 同一天，$BTC $ETH $SOL 一起走弱。 但三根阴线背后的东西不一样。 跌的原因不同： $BTC 是情绪带着卖，链上没见大户集体出逃。 400日均线附近还有承接。 $ETH 是资金在撤： ETF 连续净流出，钱包余额往外走。 有人主动减仓，不只是跟跌。 $SOL 是合约在压： 永续空头加码，价格被杠杆推下去。 现货流出有限，筹码没大规模换手。 只看跌幅排序，容易把情绪当成趋势。 分清来源，才知道哪里是噪音。#CoreWeave股债双跌，CEO回应融资担忧 $CRWV 借 100 还 100 的纸，现在只值 78 块。 ▪️ 10/14 股价跌 15% 至 65.78 美元，为 2 月以来最大单日跌幅 ▪️ 2032 年到期、票面 9.625% 的无担保债券，收益率升至 14.8% ▪️ 该债券二级价格跌破面值的 78 美分 ▪️ 信用违约互换升至纪录高位 ▪️ CEO 在内部备忘录中称市场「不理性」，称合同积压足以覆盖债务 ▪️ 同期甲骨文跌 9%、Nebius 跌 11%，纳指跌 1.4% 分歧不在 AI 算力需求够不够，在这条链的定价先发生在债市。 股票跌 15% 量的是估值，债券跌到 78 美分量的是信用——同一家公司，两个市场给出两种判断。重资产、长回收期、设备折旧快，三者叠加意味着它必须持续再融资。债券的期限比客户合同长。 方向给观望 —— 债券价格回到 90 美分以上才算信用修复；继续下探则融资链条会先收紧。#TeslaQ3ProfitAndMargins # TSMC Q3 Net Profit Exceeds Expectations, Earnings Hit Record High TSMC's revenue has already surpassed targets, so what else is there to consider at Thursday's earnings call? TSMC's Q3 revenue was actually announced earlier, about $46.7 billion, higher than the company's own forecast of $44.6 to $45.8 billion. The Thursday earnings call won't focus on revenue, but on the costs behind that revenue. The first hurdle is gross margin. The Q3 guidance is 65% to 67%, with a midpoint of 66%, which is 1.7 percentage points lower than Q2's 67.7%, due to the 2nm ramp diluting margins by about 3 to 4 percentage points. Market consensus is around 66.5%, so staying at the upper end of the range would be good news. The second hurdle is capital expenditure. In July, the forecast was raised from $52 to $56 billion up to $60 to $64 billion, so $64 billion is not a number waiting to be revised upward but the current ceiling. Management previously said capital expenditure would be "significantly higher" over the next three years; this time, it depends on how much they reveal. Regarding stock price, the ADR closed at $485.80 on 10/6, with this year's high at $487.47; the latest available closing price is $472.20 on 10/8. 演示很精彩，然后呢？ $FET 我最想看到的，是AI代理完成一次任务以后，客户还会不会再用。 官方展示的代理支付支持FET，也包含USDC等方式，所以代理交易增长，不能直接按同样比例算成FET需求。 这不是否定产品，而是需要把使用路径弄明白。 客户为什么付款、用什么付款、需要提前留多少余额，这些细节决定业务增长怎样影响代币。一次成功演示能吸引关注，反复发生的付费交易才更值得跟踪。 $AR 的生意则有个不同之处：用户预付费用，网络要承担长期存储。 上传费用中，一部分进入存储基金，用来支持未来的存储支出。 所以收到一笔费用，不能全部当成眼下可以分配的利润。 我会同时看新增存储需求和长期维护成本。订单增加值得认可，但承诺保存多久，也意味着要负责多久。 $OP 我想看的是生态投入有没有后续效果。 治理安排涉及资金配置，不能只看某个项目拿到了多少支持。 假如激励结束，用户也跟着离开，那前面的活跃就需要重新评价。 如果应用能留下用户、逐步自己获得收入，投入才更有说服力。发出去多少容易统计，留下什么更值得研究。#TSMC Q3 Net Profit Exceeds Expectations, Earnings Hit Record High TSMC's Q3 earnings report is very strong, but what I think is truly worth noting is not just the "exceeding expectations," but that despite such a good report, the stock price still fell. Q3 revenue was 1.49 trillion TWD, up 50.9% year-over-year; net profit was 763.1 billion, up 68.7% year-over-year, continuing to set a new record high. AI demand remains the biggest driver. This indicates one thing: AI is no longer just a story; there are real orders and real spending. Companies like NVIDIA sell GPUs, but ultimately they rely on TSMC for manufacturing. As long as advanced process technology and computing power demand continue to expand, TSMC is the most direct "shovel seller" in the AI industry chain. But the problem is also clear. The market's expectations for AI are too high now; it's no longer enough to just have "good performance," but you are expected to continue exceeding expectations every quarter. TSMC has raised its full-year revenue growth forecast to about 44%, and capital expenditures remain high, indicating the company still has confidence in AI demand. So I think what the AI sector really needs to focus on now is not "whether there is demand," but: Whether such large capital expenditures can ultimately continue to convert into profits. The fundamentals are not bad; what really needs caution is— Whether the market has already priced in the good news for the next two to three years in advance. @OKX星球 Bitcoin $BTC has already fallen back to around $81,000, and Ethereum $ETH dropped to a low of $2,315. Market risk appetite is declining, especially with capital becoming cautious toward high-volatility assets. I believe the main reason for this drop is that the US CPI exceeded expectations, causing market concerns that the Federal Reserve's rate cut pace might be affected; secondly, after prices broke key levels, leveraged long positions were liquidated en masse, further amplifying the decline. From a technical perspective, Bitcoin $BTC is first watching whether $80,000 can hold, with resistance above at $82,000–$83,000; Ethereum is watching the $2,300 support, and any rebound must first reclaim $2,400–$2,450; if Bitcoin breaks below $80,000, then Ethereum’s $2,300 level might also break. The most important thing now is not to guess the bottom, but to see if there is decent support after the decline. Without volume-driven stabilization and without reclaiming key resistance levels, one should not easily mistake a rebound for a reversal. Going forward, focus on US interest rate expectations, market trading volume, and leveraged liquidation situations to see if the market can stabilize first. #美国CPI超预期升至3.7%，今晚PPI再迎考验 #BTC现货ETF再度净流出，机构买盘能否回暖？ #特斯拉交付超预期后盈利修复成焦点 特斯拉Q3交付量虽然高于市场预期，但同比依旧下滑2.1%，暗藏隐忧。交付结构是最大压力点：高毛利的美国本土车型销量回落，低毛利中国产车型占比抬升，会直接拖累整车毛利率。 市场现在博弈财报盈利修复，但销量仅靠摊薄成本很难扭转利润压力。除汽车业务外，Robotaxi、Optimus进展是额外变量，若AI业务没有实质性突破，仅靠交付数据很难带动股价走强，财报存在不及预期的风险。$ETH 跌到有人开始算周线目标位了 跌了一晚，$ETH 的持仓者开始翻周线图。 有人把目标位放到 2130，再往下看 2000。 这个价位是什么：2130 不是支撑位，是上一批多单的成本区。 价格到那儿，等于把他们的止损单排队扫一遍。 谁在这儿挂单：做市商不看方向，只看两边挂单厚不厚。 一边厚就往下推，推穿了再收回来。 接下来会发生什么：$BTC 跟着下 8 万，$ETH 的深度只会更薄。 薄了以后，同样的卖量能砸出更大的跌幅。 倒推一下，2130 到 2000 这一段中间没有密集成交。 真跌进去，速度会比上一波快。 #渣打开放BTC与ETH交易，银行加密业务扩展 #$ETH $BTC 《鹰派夜，多头被抬走》💥 20:30，数据落地：核心2.6%，超预期，偏鹰。🦅 盘面先沉默，再阴跌。$BTC 从82797滑落，$ETH 在2500边缘喘气。 21:58，第一刀。单分钟，多单爆仓940万，155笔。😱 22:05，第二刀。BTC多单176万、ETH多单125万，一起归零。价格停在82125、2438。 22:15复盘：1小时全网爆6243万，九成是多的血；24小时2.4亿，74596人下车。🚑 开奖前多空互砍，开奖后单边猎多。有人赌“利空出尽”，结果利空没尽，止损和爆仓互相踩踏。 费率也分裂：BTC仍正，多头不死心；ETH转负，空头接管。ETH日内-2.99%，ETF连流10天。📉 BERA +7.3%到0.2365，费率-0.0109%，空头加仓；A +4.59%到0.1045。妖币的涨，像借来的火。🔥 我的结论：杀多未完。BTC 82000是命门，破位看81000；ETH 2400失守，下探2300。 #美国CPI超预期升至3.7%，今晚PPI再迎考验 #台积电Q3净利润超预期，盈利创历史新高 #奈飞增长或放缓，广告业务成新看点 #台积电Q3净利润超预期，盈利创历史新高 台积电Q3财报出来了，数字炸裂，但股价跌了。 营收同比增长50.9%到1.49万亿新台币，净利润同比增长68.7%到7631亿，创季度新高，超过市场预期。2nm制程占晶圆营收的8%，先进制程整体占比79%。AI芯片需求持续为业绩提供支撑，这个基本盘很硬。 但股价跌了2.67%。问题出在毛利率指引上。公司给的65%到67%，低于Q2实际的67.7%。市场担心的是，AI芯片再旺，如果毛利率往下走，说明先进制程的成本压力在吃掉利润。另外资本开支会不会进一步上调，也是接下来要盯的。 对BTC来说，这事不直接，但底层逻辑连着。台积电是AI算力基建的核心供应商，它的营收和利润验证了AI需求是真实的，算力经济的叙事没断，BTC作为底层锚定物的长期逻辑还在。但毛利率指引偏弱，说明烧钱的代价在上升，整个半导体链条的估值需要重新消化。短期BTC还在扛CPI超预期的压力，台积电财报的影响是次要的。 操作上别因为台积电利润超预期就去追科技股。等PPI落地，看美债收益率和美元怎么走，再决定接不接。$TSM $xTSM Brothers, the CPI knife has finally fallen, and both BTC and ETH have taken another hit $BTC $81,150 | $ETH $2,370 The US September CPI rose 3.7% year-on-year, higher than the expected 3.6%, and core CPI also exceeded expectations. Bitcoin dropped directly from above 84,000 to around 81,000, Ethereum fell below 2,400, hitting a low of 2,357 This wave of decline is a double whammy of "inflation + leverage." In the past 24 hours, the entire network liquidated $977 million, with long positions liquidated at $895 million, accounting for a high 92%, and 184,000 people were taken out. Ethereum liquidations topped at $293 million, 92% of which were long positions, with the largest single liquidation on Binance BTCUSDT valued at $12.03 million But there is a key divergence in the capital flow. The Bitcoin spot ETF recorded a net inflow of $175 million on October 13, with Fidelity's FBTC alone accounting for $103 million, ending a streak of outflows. The Ethereum ETF also recorded a net inflow of $24.84 million on the same day, ending 10 consecutive trading days of bleeding. ETF funds actually flowed in during the decline Technically, 81,000 is a strong support at the 400-day moving average. Analyst JackYi pointed out that Bitcoin quickly rebounds every time it touches 81,000 on a pullback, with 84,000 as the resistance at the 20-day moving average. After Ethereum fell below 2,400, 2,300 is the next line of defense #美国CPI超预期升至3.7%，今晚PPI再迎考验 BTC现货ETF又传出流出消息，这次我想先看表格有没有填完。 这次查Farside，10月14日记录里，部分产品已经报出资金流向，但IBIT等栏目仍是空缺。表格能显示一个汇总数，不意味着当天所有产品都已上报。直接把暂时空白当成零，就可能把一个待更新的结果，写成正式结论。 前一交易日的完整记录是净流入1.757亿美元。这也说明，资金方向本来就在变化，单日标题很容易让情绪从一边倒向另一边。若上报时间不同，再把不完整的当天数据和完整的昨天比较，判断就更容易偏。 这也不只是数据洁癖。记录范围没对齐，后面的讨论可能只是在比较两组完成度不同的数字。 资金到底有没有回暖？至少先让比较的对象一致。 接下来我会等同一交易日各产品数据补齐，再看净流出集中在哪些基金、有没有扩大。空缺只是暂缺信息，既不能替多头证明有人买，也不能替空头证明没人接。行情紧张时，大家都想第一时间得到答案，我也理解。但花几秒确认统计口径，总比第二天拿着被修订的数字解释“机构态度突变”省心。 #BTC现货ETF再度净流出，机构买盘能否回暖？说实话，这波空单拿到现在，我最大的感受不是爽，是后怕。 当初1405空进去，$ZEC反弹到1384那天，我盯着屏幕，手放在平仓键上停了很久。屏幕上的浮亏数字每跳一下，心里就跟着颤一下。什么趋势、什么逻辑、什么技术面，那一刻全被情绪冲得一干二净。 差一点。真的差一点就点了。 让我收住手的，不是什么定力。是我刷了一圈社群，满屏都在喊“减半”“升级”“隐私叙事起飞”。那种整齐划一的兴奋劲儿，我见得太多了。每次所有人都朝一个方向喊的时候，往往就是最危险的时候。加密市场从来没有让大多数人舒服过，这几乎是铁律。 后来的事大家都看到了。$ZEC从1285一路砸到1088，24小时跌幅超过13%，成交额放量到2.72亿USDT，均线空头排列彻底成立。ETH从2800跌到2363，ETF连续10日净流出，贝莱德和灰度一起往外撤钱。CAP从0.099掉到0.076。 三个空单，全绿了。$ETH浮盈120%，$ZEC浮盈225%，CAP浮盈44%。 但我高兴不起来。 因为我知道，$ZEC在1285追多的、$ETH在2800接盘的、CAP在0.099冲进去的，现在账户大概率已经废了。那些人不是不聪明，不是不看盘，只是在所有人都在喊“涨”的时候，很难相信自己看到的其实是一个正在收紧的绞索。 别问我什么时候止盈。我只知道，眼下这些事情一件都没解决： ETF还在流出。10月以来，美国现货比特币和以太坊ETF合计净流出已经逼近10亿美元，以太坊ETF更是连续8天净流出。 NU7升级的决策，要等到10月20号才能定。测试网现在只是在跑，主网能不能按计划在11月5号上线还是未知数。这意味着$ZEC头顶上最大的叙事催化剂，目前只是悬在半空。 巨鲸在干什么？一边有机构清仓17.25万枚$ETH套现1.24亿美元转头去买UNI，另一边有巨鲸在Aster DEX上用20倍杠杆同时做空$BTC、$ETH、$SOL、$XRP和$DOGE，所有空头头寸全部处于盈利状态。聪明钱在做的事，和散户在喊的话，完全是两个方向。 山寨在塌方。$XRP衍生品市场单日多头爆仓超过6.1亿美元，创下历史纪录。更夸张的是，某持有3.52亿美元多单的巨鲸或机构，其79260枚ETH多单的清算价在2290到2301美元之间，距离当前价格只差60美元——随时可能被强平。 这些事，一件都没解决。在这些变量明朗之前，我不走。 你信不信随你。但如果你现在还在纠结要不要抄底，先别问自己能赚多少，先问自己一个问题： 你是在抄底，还是在给别人的空单送燃料？ $BTC $ETH $ZEC #美国CPI超预期升至3.7%，今晚PPI再迎考验 #台积电Q3净利润超预期，盈利创历史新高 #奈飞增长或放缓，广告业务成新看点 #美国银行遭重罚，华尔街Q3业绩分化 $BTC 净利少掉的 13.67 亿里，有 12 亿是一张罚单。 ▪️ Q3 营收 299.58 亿美元，同比 +3.2%，低于预期 305.8 亿 ▪️ 净利润 69.33 亿美元，同比 −17%，去年同期为 83 亿 ▪️ 其中包含 12 亿美元监管和解费用 ▪️ 每股收益 1.08 美元，低于预期的 1.10 美元 ▪️ 财报后开盘跌超 7%，收跌近 5% ▪️ 同期摩根士丹利股票交易收入同比 +41%，贝莱德单季净流入 2,050 亿 分歧不在它赚得少不少，在「少赚的 14%」里有多少是一次性的。 83 亿到 69.33 亿，差额 13.67 亿，其中 12 亿来自和解费用——这一项不会连着发生，但它躺在利润表里，与交易业务疲软混在同一行。剔掉罚款，剩下的才是经营变化。 方向给观望 —— 下一季不含罚单的净利回到同比正增长才算修复；否则降幅只是被罚单盖住。$BLUR 这波空单稳稳吃肉，找准位置挂空才是硬道理 我0.01283开的空，现价0.01254，浮盈6.07%，方向踩得死死的。 先看最新消息，Blur官方前几天宣布关停前端，引发市场恐慌抛售，巨鲸也在持续出货。链上数据显示，有大额BLUR持续转入交易所，抛压一波接一波。虽然盘中有反弹，但那都是诱多，追高的全被埋在山顶。 盘面看，上方0.01297压着557.77K大额卖单，0.01296还有519.30K，抛压层层叠叠。下方买盘稀稀拉拉，承接薄得像纸。多空比41%的多头对59%的空头，空头完全主导了节奏。 BLUR从高位一路砸下来，反弹一次比一次弱，每次冲高都是空的机会。我0.01283进的空单继续拿着，防守放0.0135上方，下方先看0.012，跌破就是0.011。找准位置，挂好止损，剩下的交给时间。 $BTC $ETH #美国CPI超预期升至3.7%，今晚PPI再迎考验 短线修复，不改空头底色 $BTC 、$ETH 大势仍偏空，但我不觉得会直接一根筋下杀。更像先给一段弱修复，诱多后再跌。 CPI 3.7%，超预期。降息交易被泼了盆冷水，加息风险重新摆上桌面——CME数据显示10月加息概率从22%飙到53.7%。美元指数拉到102.7，10年期美债收益率逼近5.33%，创24年新高。威廉姆斯、柯林斯、古尔斯比排队放鹰，AI投资过热正在往物价里灌通胀，美联储“更高更久”的逻辑被重新定价。美元、美债收益率、官员讲话，接下来任何一个都可能重写剧本。 若这种背景下突然急拉，我会把它当强牛回归信号，但概率太低。 $BTC今天盘中一度砸到80230，离8万就差一口气，随后拉回，近1小时全网爆仓4.84亿美金，多单占了4.64亿。ETF那边也没闲着，昨天$BTC现货ETF净流出2.61亿，$ETH流出1.29亿，贝莱德一家就撤了1亿多。但有意思的是，链上大户持仓已经爬到303万枚$BTC，月度增持速度是今年2月以来最快的。机构在跑，大户在接，这种背离本身说明不了谁对谁错，但至少说明有人在赌方向。 我的$BTC均价63000，是从上轮暴跌8万一路捡下来的。朋友笑我接刀，他们等5万；可我怕的是没仓位，踏空比浮亏更难受。前提是现货，合约插针太狠，麻吉大哥早上已被爆。 后面只看：$BTC能否失守8万，$ETH能否失守2360。$BTC 8万这个位置今天已经测过一次，下方7.8万和ETF持仓成本区间才是真正的底牌支撑。$ETH更弱，ETF连续10天净流出，累计跑了7.83亿美金，鲸鱼还在往交易所搬货，2382那个位置挂了一大面卖单墙。破位，空头续演；守住，还有拉扯。 现货慢慢熬，合约别赌命。 #美国CPI超预期升至3.7%，今晚PPI再迎考验 #台积电Q3净利润超预期，盈利创历史新高 #BTC现货ETF再度净流出，机构买盘能否回暖？