Orbit Post Sitemap
South Korean Stock Market Falls to Crisis Valuation, Semiconductors May Face High Odds Opportunity
The KOSPI fell about 22% in July, with a maximum drawdown of 39% over the past 12 months, marking the worst single-month performance since the global financial crisis.
In fact, the South Korean stock market has already fallen into an extremely extreme risk-reward ratio.
Currently, the KOSPI's forward 12-month price-to-earnings ratio is only about 4.7 to 4.85 times, the lowest since 2001, even lower than the bottom during the 2008 financial crisis. Meanwhile, the KOSPI's ROE remains around 25%, and the EPS forecast for the next 12 months was recently revised upward by 0.9% in the past week, showing a clear divergence between stock prices and earnings trends.
This round of decline has been accompanied by a thorough position cleanup. Foreign capital continues to flow out, foreign holdings in South Korean semiconductors are already low, margin balances have dropped from $25 billion to $22 billion, and leveraged ETF sizes have also decreased from $53 billion to $26 billion, nearly halving.
Historically, after the KOSPI experiences a drawdown of over 30%, the median gains in the following 3 months are 16%, 6 months 43%, and 12 months 59%.
The biggest risk in the South Korean market remains the downward revision of tech earnings. However, against the backdrop of continued rising memory prices, significant position clearing, and earnings expectations still being revised upward, this may be one of the most prominent rebound directions with the best odds in the current global market.$ETH This poor trend suddenly pushed the price to 1900 at five or six in the morning. I was still asleep, but after waking up, it was slowly falling again
I have a long position, reduced once at 1870, and again at 1850, so I still have some profit now
But 1900 did not stay, indicating that selling pressure above was heavier than expected. That rally felt more like a quick sweep rather than a real trend start
The price has now fallen back below the 1-hour moving average, so the remaining positions are not expected to be bought again when it returns to 1900
If it can hold around 1840, I'll keep some to watch for repair
If it recovers to 1865, there will be another chance to test 1875. If the rebound can't even break 1855, I will continue to reduce my position
1838 fell again, the remaining positions were directly disposed of, not taking 100x for a second raid
This time, the direction wasn't completely wrong, but during execution, one piece was missed: the person wasn't at the market price, and the exit plan wasn't included in advance
From now on, if you dare to stay overnight at high multiples, you must first take profits
This time, I learned my lesson. What's important next isn't predicting the next high, but that whenever a position leaves your sight, exit conditions must be written in advance in your order.
#30年期美债, the top or a new beginning?
#美日确认联合购汇 After Microsoft released its earnings report, its stock price rose nearly 9% in after-hours trading. Azure's growth accelerated from 40% to 43%. A wave of joy erupted.
But those who look closely at the financial report feel a chill down their spine.
Microsoft has extended the depreciation period for data centers directly from 15 years to 25 years.
Then, in a way, the 2026 capital expenditure guidance was "lowered" from $190 billion to $175 billion.
CFO Amy Hood came out to reassure them: Don't worry, it's not about cutting AI investments, it's a change in accounting standards.
The actual investment was not a single cent less.
The book figures suddenly dropped by 15 billion.
Here comes the question—
There's a strict rule in the chip industry: one generation every 18 months.
The GPU you buy today will be electronic waste in three years. Five years later, no one even maintains the driver.
Microsoft said my data center can last 25 years.
What does 25 years mean?
In 2001, the original Xbox was released.
In 2026, tell me, those servers back then can still run AI?
Economist David Woo flipped the table: This is not real economic improvement, but accounting treatment to reduce depreciation. The logic of AI investment continues to deteriorate, and the AI bubble has officially burst.
"Big short" Michael Burry is even more ruthless—he estimates that from 2026 to 2028, tech giants will inflate their book profits by $176 billion by extending depreciation periods.
Think about this logic:
On the surface—profits have increased, capital expenditures have "decreased," perfect.
In reality—chips are phased out in 3 years, depreciated for 25 years. The depreciation deducted annually is all converted into book profit. When the day comes that batch of equipment is truly scrapped, a one-time impairment could blow up the financial report.
This is not accounting treatment; it is planting a bomb 25 years later.
Some might say: Microsoft Azure revenue has surpassed 100 billion, Copilot has over 30 million paying users, AI monetization is accelerating, so why call it a bubble?
Let me tell you—
Revenue is growing, but that doesn't mean the depreciation logic is sound.
Moody's has raised its 2026 capital expenditure forecasts for the six major hyperscale cloud vendors to $785 billion. Morgan Stanley is even more aggressive, with 795 billion.
The entire industry is burning money wildly to build data centers.
But how do you break even after burning it?
The longer depreciation is amortized, the longer it takes to verify the break-even period.
Microsoft has "made" profits 25 years from now into today's financial statements in advance.
Wait 25 years? The AI industry has been an era for 18 months.
The most heartbreaking part is—
When stock prices rise, retail investors rush in to buy in.
They think Microsoft's AI investment is "slowing down" and profits are "growing."
In reality, nothing has changed—the investment hasn't decreased, but the profits are calculated.
What they are buying is not the future of AI, but the premium of accounting magic.
$MSFT $GOOGL $META #折旧年限延至25年, Microsoft's capital expenditure guidance was lowered قانون وضوح سوق الأصول الرقمية لعام 2025/2026
بشكل مختصر هو هو محاولة أمريكية كبيرة لوضع قواعد واضحة لسوق الكريبتو بعد سنوات من الفوضى والخلاف بين الجهات التنظيمية، لأن المشكلة الأساسية في أمريكا كانت أن هيئة الأوراق المالية الأمريكية SEC كانت تعتبر أغلب العملات المشفرة أوراقًا مالية مثل الأسهم، بينما هيئة تداول السلع CFTC كانت ترى أن بعض العملات مثل البيتكوين والإيثيريوم أقرب للسلع مثل الذهب أو النفط، وهذا الخلاف خلق حالة يسمونها “التنظيم عبر الدعاوى القضائية” حيث الشركات لا تعرف هل مشروعها قانوني أم لا إلا عندما تُرفع ضدها قضية، لذلك جاء هذا القانون ليحدد من يراقب ماذا، ومتى تصبح العملة سلعة رقمية بدل ورقة مالية .
الفكرة الجوهرية في القانون هي أن المشروع إذا كان لا يزال مركزيًا ويوجد فريق يتحكم به بشكل واضح ويعد المستثمرين بالأرباح، فهنا يبقى تحت رقابة SEC باعتباره شبيهًا بالأسهم والاستثمارات التقليدية، أما إذا أصبح المشروع لامركزيًا بشكل كافٍ ولم يعد فريق معين يتحكم به، فقد ينتقل إلى تصنيف “Digital Commodity” أي سلعة رقمية تخضع لـ CFTC، وهذا التحول مهم جدًا لأن رقابة CFTC عادة أخف من SEC وأسهل على الشركات والمشاريع.
القانون أيضًا يضع قواعد للمنصات والشركات العاملة بالكريبتو، مثل منصات التداول والوسطاء والحفظ، ويجبرهم على فصل أموال العملاء عن أموال الشركة، والإفصاح عن تضارب المصالح، والاحتفاظ بسجلات واضحة، إضافة إلى متطلبات تتعلق بمكافحة غسل الأموال وتمويل الإرهاب ضمن قوانين Bank Secrecy Act الأمريكية.
#clarity
#CLARITYActVoteWatch 🚨 One of the most expensive mistakes in crypto isn't buying a bad project—it's ignoring its tokenomics.
A token priced at $0.10 isn't automatically cheaper than one priced at $1,000.
Why? Because price alone tells you very little. Supply tells the real story.
Too many traders focus on the chart while overlooking what happens when millions—or even billions—of new tokens enter circulation.
Even the strongest narrative can struggle if continuous token unlocks create persistent selling pressure.
That's why understanding tokenomics is essential.
Before investing, ask yourself:
✅ How much of the total supply is already circulating?
✅ What is the fully diluted valuation (FDV)?
✅ When is the next major token unlock?
✅ Who receives those unlocked tokens, and are they likely to sell?
Projects with high FDVs, low circulating supply, and frequent unlocks can experience ongoing dilution—even if the underlying technology is excellent.
👀 Projects with notable token unlock schedules:
$ARB • $OP • $STRK • $ZK • $BLAST • $MANTA • $ALT • $DYM • $TIA • $SUI • $APT • $SEI • $PYTH • $JUP • $W • $EIGEN • $REZ • $ETHFI
🌐 DeFi & RWA $ONDO • $MKR • $AAVE • $UNI • $PENDLE • $ENA • $SNX • $CRV • $COMP • $LDO • $RPL
🤖 AI & DePIN $TAO • $FET • $NEAR • $RNDR • $AKT • $AIOZ • $GRT • $THETA • $FIL • $AR
🐸 Memecoins $PEPE • $WIF • $BONK • $FLOKI • $POPCAT • $BOME • $DOGE • $SHIB • $MOG • $BRETT
The best investors don't just ask:
"Is this a good project?"
They also ask:
"Does the tokenomics support long-term performance?"
Great technology doesn't always make a great investment if the token's supply dynamics are working against holders.
Don't just study the chart.
Study the supply.
That's where many of crypto's biggest opportunities—and biggest risks—are found.
Not financial advice. Always do your own research.
#Crypto #Tokenomics #Bitcoin #Ethereum #Altcoins #Trading #RiskManagement #BTC #ETH #DailyOrbit This market script is even more melodramatic than a TV series.
Just after rumors broke that the US and Iran were about to sit down for negotiations, Brent plummeted 7.3% in a single day, and the nearly 25% increase from July was wiped out in just a few days;
Immediately after, Iran's official government poured cold water on the matter: they never even engaged in dialogue with the US!
Everyone is discussing: Can this easing last longer than the July wave?
My view: This easing is purely a market imagination boost. If the range isn't strong, landmines could explode at any time. Before the written agreement was released, my approach remained unchanged: focusing on rebounds and high prices.
By the way, let's talk about how retail investors deal with these endless rumors of truth and falsehood, so you don't get ripped off.
My little trick for identifying messages
1. Empty Talk ≠ Implement Policies
If there's no official stamp, just take it with a grain of salt—don't take it seriously.
This negotiation rumor is a typical example: the market hyped up first, then the rumor was debunked on the spot. When you encounter such breaking news, don't rush to invest heavily and enter the market to play the game.
2. Don't be fooled by the edited news from self-media
The media loves to take things out of context and stir things up, packaging half a sentence as a major positive news.
Before placing an order, try to dig up the original information; don't believe whatever others say, or you'll definitely be the buyer.
3. Betting with your feet on the funds means the market won't act out of control
True good news means steady market movement, not a sharp rise and then a pullback.
A sharp drop after a rally shows that even big funds themselves don't believe this story, just short-term selling and retail investors fleeing.
How to play in practice
1. Treat news as entertainment; when making trades, look at hardcore indicators
Oil prices, US Treasury yields, and inflation are the real big brothers; rumors can only stir things up for a day or two.
Without a formal written agreement, I won't change the general direction of the rebound at high levels.
2. Don't get carried away by rumors—control your positions
This roller coaster market has maxed out the leverage risk.
Even if you want to try a rebound, just play with small positions—don't rush in and debunk rumors—a single reversal can seriously damage your principal.
3. Don't rush to expect high-risk boxed meals; wait until the melons are fully ripe before eating
The market always follows its usual pattern: buy rumors, sell facts.
When news first breaks, it's easiest to be exploited. It's better to quietly watch the drama and wait until the dust settles before deciding whether to enter the market.
💡 Honestly
Retail investors are already at the bottom of the news chain; competing with institutions for intelligence is just asking for trouble.
Instead of staying up late every day to check news and guess what's true, it's better to stick to your trading rules.
Publicity can hype the day's market trends, but pricing ultimately depends on real data. Don't believe rumors lightly, don't gamble on rumors in heavy positions.
Unless both parties sign a solid, written agreement, I won't change my overall strategy. Before this, the rebound was still dominated by high-level trading.
Have you been exploited by this fake news these past couple of days? $CL $BZ
#美伊重回谈判桌, oil prices pulled back 【法老看盘】
都在问法老，韩股昨天涨14%，今天跌5%，存储股多空信号打架，到底听谁的？
法老直接说，韩股这走势跟抽风一样，暴涨暴跌的本质没变。存储股现在一边是AI基建的长期故事，一边是韩国流动性危机的短期现实，多空在这个位置反复拉扯。
先看发生了什么。
韩国KOSPI指数在昨天创纪录暴涨14%后，今天开盘跳空低开近4%，随后继续下探，全天跌幅扩大至5.2%，基本吐回了昨天一半的涨幅。SK海力士昨天涨了15%，今天跌了8%；三星电子昨天涨了8%，今天跌了5%。
为啥涨得快跌得更快？
韩国股市高度集中于三星和SK海力士两家，加起来一度占KOSPI总市值超60%。昨天的大涨更多是空头集中回补和情绪修复，不是基本面改善。今天市场回过神来——韩国央行加息周期没结束、AI泡沫担忧没消退、油价还在高位——直接反手做空。
对大饼意味着啥？
存储股跟大饼现在是跷跷板关系。存储涨，资金从大饼流向芯片股；存储跌，部分资金回流加密。韩股这边反复抽风，大饼短期跟着情绪走，62000-63000区间震荡。$BTC $ETH $SOL
记住，好单子是等出来的，不是追出来的。
关注法老，财富不迷路！#韩股重挫5%，存储多空信号对峙 🟢 $ZHIPU Liquidated Short Alert
🟢 $ZHIPU Long Setup
Entry Zone: $120.80 – $122.20
🔸 TP1: $123.50
🔸 TP2: $126.00
🔸 TP3: $129.00
SL (Stop Loss): $119.50
ZHIPU recorded a short liquidation of $1.735K at $121.49978, indicating that bearish positions were forcefully closed as buying pressure accelerated. The liquidation reflects strengthening bullish momentum and increased market volatility. If buyers continue defending the current price zone, ZHIPU could extend its move toward the listed upside targets.
Trade with proper risk management.
Do you think buyers can keep this momentum going? 👇
#30YrYieldTopOrStart #CreatorRewards #DailyOrbit Monday Pullback? Here's Why the Market Often Starts the Week in the Red
A sea of red across the market isn't unusual at the start of a new trading week. Today action shows major cryptocurrencies including $BTC, $ETH, $SOL, $XRP, $LINK, $UNI, $AAVE, and $ONDO all trading lower, reflecting a broad-based risk-off mood rather than weakness in a single asset
Monday sessions often bring increased volatility as traders react to weekend developments, reposition portfolios, and wait for fresh macroeconomic catalysts. That can lead to temporary selling pressure before the market establishes a clearer direction
While short-term price swings grab attention, it's important to watch whether key support levels hold and if trading volume confirms the move A broad market dip doesn't always signal the start of a larger downtrend it can simply be part of the market's normal weekly reset
Do you see Monday weakness as a buying opportunity, or do you prefer waiting for the market to reveal its direction before making a move?
#DailyOrbit #OKXOrbitTopics #EarningsWeekAhead 《美股涨疯了，币圈还会远吗？》
近期美股科技股持续发力一路走高，$MSFT 微软、$SKHYNIX 存储板块轮番拉升，赚钱效应直接拉满！
不少人开始期待，风险情绪回暖之后，币圈能跟上这波多头行情。
可现实却十分残酷，#30年期美债，顶部还是新起点？ $BTC 大饼、以太坊依旧维持窄幅震荡，迟迟没法突破区间。
这一轮美股上涨依靠 AI 业绩、订单数据支撑，资金优先选择有基本面的标的。
场外增量资金扎堆美股，迟迟不愿流入加密赛道，监管不确定性持续压制行情。
以往美股走强带动币圈普涨的联动效应，正在不断弱化。
外围利好只能短暂提振情绪，没有持续买盘，反弹很难走远。
美股牛市热火朝天，币圈却持续磨盘，我们真的能坐等行情轮动到来吗？#折旧年限延至25年，微软资本开支指引下调
最近微软财报有个细节特别有意思：他们把数据中心的折旧年限从15年拉长到了25年，结果2026年的资本开支指引直接从1900亿降到了1750亿，账面少了150亿美元。很多人第一反应是“微软是不是要缩减AI投入了？”但仔细看财报电话会实录，管理层说得明明白白：“除会计准则变更外，投入预期没有任何变化。”说白了，这150亿不是真省下来的钱，只是换了个记账方式——原本算进资本开支的融资租赁，现在变成经营租赁不计入了，但数据中心该建还得建，芯片该买还得买。
更关键的是，这150亿的“消失”背后藏着微软的底气。就在同一场电话会上，他们宣布Azure云服务营收同比增长43%，需求持续超过供给上限，还新增了31个数据中心。如果真没钱投AI，哪来的底气扩产？所以别被账面数字忽悠了，微软的AI投入一分没少，只是把“花钱”的账本换了个写法。
很多人觉得折旧年限拉长，意味着回本周期变长，AI投资回报的验证也会拖得更久。但这个逻辑放在微软身上，其实没那么简单。
首先，折旧年限调整的依据是“多年运营数据和长期使用规划”，不是拍脑袋决定的。微软的数据中心用了十几年还没坏，延长折旧年限反而更符合实际资产寿命。其次，AI投资的回报不能只看折旧周期。微软现在搞的是“订阅+用量”双重收费模式，Copilot按人头收费，GitHub Copilot按开发者收费，连客服业务都按用量积分收费。这意味着收入不再只靠“卖人头”，而是和企业用AI的实际消耗挂钩。就算折旧摊到25年，只要AI调用量持续增长，收入就能覆盖成本，回本周期反而可能缩短。
不过风险也确实存在。如果AI芯片迭代速度比数据中心物理寿命快得多，25年折旧意味着资产还没摊完就过时了。但微软早就留了后手：他们把模型、上下文、调度工具完全解耦，企业可以随便换模型，不用被单一技术绑定。这种架构设计，本质上是在降低技术迭代带来的资产贬值风险。所以评估微软的AI回报，别只盯着折旧年限，更要看它的商业模式能不能扛住技术迭代。
穆迪把2026年六大超大规模云厂商的资本支出上调到7850亿美元，这个数字确实吓人，但它改变的不是“投入节奏”，而是“投入逻辑”。
以前大家看AI基建，只看“花了多少钱”；现在得看“钱花得值不值”。微软就是典型例子：他们资本开支账面降了150亿，但Azure营收增速反而从40%加速到43%，商业剩余履约义务（已签约未确认收入）达到6780亿美元，同比增长84%。这说明钱没少花，但花得更精准了——每一分投入都能转化成订单和收入。
反观Meta，同期资本开支301亿美元，营收增速28%但营业利润增速只有8%，营业利润率从43%暴跌到31%。同样是投AI，微软的钱变成了“可量化的收入”，Meta的钱更多是“内部成本”。穆迪上调的数字，其实是在提醒市场：AI基建的军备竞赛没停，但已经从“比谁花钱多”进入“比谁花钱值”的阶段。
所以别被7850亿这个数字吓到，它不是“投入加速”的信号，而是“投入分化”的开始。未来能活下来的云厂商，不是花钱最多的，而是能把每一分投入都变成订单和现金流的那一个。
微软这150亿的“消失”，本质是会计处理的变更，不是AI投入的收缩；折旧年限拉长，不代表回本周期变长，反而可能因为商业模式升级而缩短；穆迪上调的7850亿，不是投入节奏的改变，而是投入逻辑的分化。
评估AI时代的科技公司，别只看“花了多少钱”，更要看“钱花得值不值”“商业模式能不能扛住技术迭代”“投入能不能转化成订单和现金流”。微软的财报，其实给所有AI玩家上了一课：真正的竞争力，不是账面数字多好看，而是能不能把每一分投入都变成可持续的收入。
微软这份财报不仅影响了传统科技股，对币圈也释放了三个关键信号：
1. AI代币的“叙事泡沫”正在被戳破：过去币圈很多AI项目靠“未来算力需求”讲故事，但微软财报证明，市场现在只看“实际算力变现”。那些没有真实收入、纯靠融资烧钱的AI代币，可能会面临和Meta类似的估值回调。投资者会更关注项目是否有像Copilot那样“按用量收费”的商业模式，而不是空谈“AI革命”。
2. 算力租赁赛道的“重资产”风险凸显：微软把融资租赁转为经营租赁，本质是把“表内负债”变成“表外承诺”。币圈很多算力租赁项目也采用类似模式，用长期租赁承诺掩盖现金流压力。穆迪上调7850亿资本支出，提醒投资者：算力租赁不是“轻资产生意”，如果项目方没有像微软那样“需求超过供给”的真实订单支撑，表外负债可能随时变成流动性危机。
3. DePIN项目的“效率叙事”迎来窗口期：微软财报显示，AI基建的投入逻辑已从“规模”转向“效率”。DePIN（去中心化物理基础设施网络）项目的核心价值，正是通过分布式算力提升资源利用效率。如果传统云厂商的“重资产”模式开始被市场质疑，DePIN的“轻资产、高效率”叙事可能获得重新定价的机会。Robinhood刚刚用一份财报，揭示了美国零售投资者偏好正在发生结构性转变。 据BlockBeats消息，Robinhood第二季度报告显示——预测市场（事件合约）收入达1.56亿美元，不仅创下历史新高，还首次超过了加密货币交易收入（1亿美元）和股票交易收入（1.29亿美元）。事件合约同比增幅超过十倍，成为Robinhood最大单一收入来源。 这标志着Robinhood正在从一家“散户炒股平台”转型为“散户赌事件平台”。 三大收入来源对比： 预测市场收入：1.56亿美元（历史新高，首次登顶） 股票交易收入：1.29亿美元 加密货币交易收入：1亿美元（同比下降38%） 加密货币交易收入同比下降38%——在比特币和以太坊过去数月价格波动有限的背景下，散户对“赌方向”的需求正在从加密资产本身向更广泛的事件合约转移。 预测市场的崛起意味着什么？ 用户行为在变化：零售投资者越来越倾向于交易“事件结果”（如选举、政策决策、经济数据）而非单纯的资产价格。这本质上是一种“结构化押注”的娱乐化趋势——更像体育博彩，而非传统投资。 监管风险同步上升：美国监管机构对预测市场的关注度正在升温——CFTC已对#韩股重挫5%，存储多空信号对峙
暴跌45%的闪迪能抄了？
别急，先看清楚这三个信号
闪迪从6月高点下来已经跌了快一半，市值蒸发了大几百亿美金。很多人开始心痒痒，觉得跌到位了。8月可能是个更好的入场时点。
但这波暴跌不只是情绪宣泄，有几个硬伤得先看清楚：
第一，跌的是估值，不是基本面。
闪迪、美光、海力士三家业绩都是历史最炸裂的，美光毛利率干到84%。但市场现在不看你赚了多少，是怕你赚不长久——84%的毛利率本身就成了“周期见顶”的信号。
第二，筹码结构还没完全出清。
韩国那边7月干了2.3万亿韩元的杠杆平仓，监管还提高了融资门槛。虽然上周五KOSPI暴力反弹了快18%，但更多是空头回补，不是真买盘回来了。科技股的成交占比和机构持仓都没怎么降，筹码拥挤度还在高位。
第三，长协是把双刃剑。
三星、海力士签了五年期长期协议，锁了60%-70%产能，预付款都收了。但长协只能在涨价时锁定利润，一旦需求转弱，客户照样会要求重新谈判——2021年芯片荒时Microchip就这么干过，几百亿的合同说废就废了。
目前存储的 基本面没变，但筹码和情绪还没到位。想抄底的，等DRAM ETF的换手率降下来、散户恐慌盘出清再说。
想做多的轻仓试水可以，重仓梭哈再等等。想做空的可以逢高布局。$SNXX 3. How to interpret the BTC market (the part you care about most)
Scenario 1: No vote in the August window / Failure outright (higher probability)
The market had already priced in part of the expectation, and once it falls short, a wave of positive expectations may fade and long positions may close out and pullback from the market.
But this does not mean the bear market is restarting: ETF funds, halving cycles, and US dollar liquidity remain the main themes; The bill is just a "catalyst," not the sole determining factor of trends.
Scenario 2: Senate narrowly passes (low probability)
In the short term, it will trigger a round of compliance narrative rally, institutional capital expectations will heat up, and BTC will rebound. But the key warning:
- Legislation passed only in the Senate ≠ immediately;
- The market often "surges and then falls back on the day the news lands";
- The bill itself is a regulatory framework, not a direct permission for unlimited token issuance. Medium- and long-term benefits will gradually materialize, not a one-time unlimited price increase.
Scenario 3: The August window is postponed or postponed to next year
Market sentiment is neutral to bearish, with short-term fluctuations digesting and returning to the original range (the 64,000/74,000 range logic you previously focused on continues to dominate). Anyway, there are plenty of them. Don't be fooled by how things affect your mood—just let it go. It's not that he hasn't lost out before. If prices rise, it is all that heaven has given us. We must believe firmly and always look forward.🚨 The moment everyone calls a coin "guaranteed," it's usually no longer early.
That's where most traders get trapped.
They don't buy the wrong project.
They buy the right project... too late.
The pattern rarely changes:
• Smart money accumulates quietly.
• Price starts climbing.
• The crowd notices.
• FOMO takes over.
• Early buyers start taking profits.
Take $LAB.
The breakout was impressive, but one big green candle doesn't automatically mean altseason is here.
Real trends need more than hype.
They need:
✅ Strong demand
✅ Deep liquidity
✅ Growing on-chain activity
✅ Fresh capital
Right now, liquidity remains focused on:
$BTC • $ETH • $HYPE • $BNB • $AAVE • $LINK • $PENDLE • $JTO
If participation expands, watch:
$SUI • $ARB • $OP • $ONDO • $SEI • $ENA • $XMR • $KAITO
The traders who last aren't the ones chasing every breakout.
They're the ones who stay patient and follow where the money is actually flowing.
Patience beats FOMO more often than people think.
#DailyOrbit #30年期美债, the top or a new beginning?
Currently, the 30-year Treasury yield is in the range of 5.23%–5.28%, and from late July to early August, it briefly reached around 5.28%, marking the highest level since 2007. In the short term, the rise has been steadily rising from around 4.9% at the end of June, but the upward trend is not moderate.
This isn't about a slight rise and then a pullback. It is hitting a nearly 20-year high, and it comes at a time when the Federal Reserve has just maintained a policy rate of 3.5%–3.75%, with three board members publicly inclined to raise rates. The market is voting on price: long-term interest rates no longer follow the federal funds rate exactly.
Why is the price rising so strong?
Three forces overlapped.
The first is the stickiness of inflation. Core PCE remains above 3%, still some distance from the 2% target. The market is increasingly skeptical of the narrative that "inflation is dead," and term premiums have been repriced.
The second is fiscal supply. With large deficits and a slow pace of bond issuance by the Ministry of Finance, long-term supply continues to pressure the market. When the private sector also issues large amounts of bonds for AI infrastructure, the competition for funds will be more directly reflected in long-term yields.
Third, the Federal Reserve's communication style has changed. The new chairman places greater emphasis on letting the market watch data and inflation for themselves, rather than overly relying on policy path signals. The result is that the long-term end sets its own prices and is no longer easily suppressed by "dovish expectations."
Historical comparisons are also noteworthy: in 2007, when the 30-year yield was above 5%, the federal funds rate was much higher than it is now. Now, policy rates are lower, but the long end is demanding higher compensation, indicating that the market is charging extra for ongoing fiscal risks and inflation uncertainty.
The top, or a new beginning?
From a technical perspective, resistance has formed multiple times in the 5.0%–5.2% range. This year has already tested and effectively broken out for the third time, with each pullback lower point rising. This structure resembles an upward breakout of an ascending triangle rather than a simple top pattern. If it can hold above 5.2% and continue to expand, the next target area will target 5.5% or even higher.
Fundamentals have not given clear signs of a top. As long as inflation does not quickly fall back to the target and fiscal policy does not undergo substantial contraction, it will be difficult for term premiums to fall sharply. In other words, the current level is more like the lower edge of the "new normal range" rather than a phased top.
Of course, the top can always suddenly appear? If economic growth slows significantly, employment data deteriorates sharply, or inflation unexpectedly falls rapidly, the long end can quickly take profits. But based on current data, this scenario is not the main theme.
What does this mean for other markets?
The continuous rise in long-term yields means that the discount rates for all assets are becoming more expensive. Stock valuations, especially long-duration growth stocks, will feel pressure; Real estate financing costs will also be directly transmitted. For the crypto market, risk assets are sensitive to real interest rates, and long-term upside usually signals a cooling of risk appetite, unless a stronger liquidity hedging narrative emerges.
Conversely, if it is later confirmed that this is only a temporary overrush and a long-term pullback, it will be a temporary positive for risk assets. Key points to watch remain are inflation data and the Treasury's bond issuance pace.
The more likely scenario is that the 30-year Treasury has not reached the "top," but has entered a higher trading range. Above 5% may become part of the new normal, rather than a transient extreme. The market is relearning how to price assets in a higher long-term interest rate environment.You might not know yet
Samsung's foundry business may be reaching an important turning point.
In recent years, Samsung's wafer foundry has been plagued by one problem: production capacity has been ramped up, but orders are insufficient and machines cannot be fully operated.
This is exactly what the semiconductor industry fears most.
Wafer fabs require huge investments, with fixed costs for equipment, plant space, and labor. If the production line utilization rate is low, no matter how advanced the technology is, it will be difficult to make money.
But now things are changing.
According to industry sources, Samsung's foundry capacity utilization is expected to approach full capacity in the second half of this year, with utilization rates currently rebounding to around 70%-80%.
Around 2024, the utilization rate of some advanced process production lines at Samsung once fell below 50%.
The core driving force behind this is two words: AI.
With the explosive demand for AI computing power, the demand for high-bandwidth memory (HBM) continues to grow. The advanced packaging and substrate chip manufacturing behind HBM are also driving wafer foundry demand.
At the same time, cloud computing giants, AI chip companies, and high-performance computing clients are all increasing orders for advanced process technologies.
Especially the 2nm process, which is becoming the key for Samsung's comeback in the coming years.
Currently, companies like Qualcomm, AMD, Google, and even Tesla are all considered to be focusing on or laying out Samsung's advanced nodes.
Of course, for Samsung to truly escape losses, there is another key indicator:
2nm yield.
Simply put, producing 100 chips can meet the standard ratio.
Currently, Samsung's 2nm yield is still improving. If it can steadily break through around 70% in the future, it means more major customers may genuinely entrust Samsung with mass production.
Therefore, Samsung's real winning or losing hand in foundry is not just "whether the factory has orders," but rather:
Can advanced processes be produced stably?
If demand for AI chips continues to grow in the coming years and Samsung's 2nm and 1.4nm processes gradually mature, then Samsung's foundry may be shifting from a "cash-burning business" to a new profit growth point.
The market has been closely watching the competition between Intel and TSMC, but don't overlook Samsung.
The next round of competition in the semiconductor industry is likely not just about chip design, but also:
Competition of advanced manufacturing capabilities + AI supply chain integration capabilities.
Samsung is making a comeback at the table.Amazon's $50 billion is more likely to be seen as a "cloud computing contract for the next eight years" rather than a traditional equity investment.
Many people focus on $50 billion, but I'm more interested in another set of figures: OpenAI has committed to spending $100 billion on AWS over the next eight years.
If this promise is fulfilled, what Amazon is truly buying is not OpenAI shares, but rather the most stable and highest-quality computing power demand in the future AI era. Cloud computing is inherently high in gross margin and cash flow; as long as OpenAI continues to train its models, that money will keep flowing back to AWS.
Of course, this deal is not without risks.
Amazon holds preferred shares, not common stock; the true value of the equity will only be realized when OpenAI goes public or experiences liquidity events in the future. If OpenAI maintains a closed structure and does not go public for a long time, this portion of paper profits could be locked up for many years. So I believe the biggest variable in this investment isn't how much OpenAI is worth, but how much the $100 billion cloud order can actually be fulfilled.
If OpenAI continues to lead in the coming years and the demand for large model training and inference keeps growing, Amazon has already secured its largest AI customer in advance; But if model competition turns into a price war, training costs fall, or OpenAI starts spreading more computing power across other cloud providers, the $100 billion may not be fully realized.
So my judgment is clear: this is not a bubble, nor is it just a bet, but a business deal that binds customers.
Many people see $50 billion in investment, but I see Amazon locking in cash flow for the next eight years in advance. In the AI era, what is truly valuable has never been the shares themselves, but who can continuously sell computing power, infrastructure, and earn cash.
That's why when I look at the AI sector, I focus less on which models are stronger, and instead focus more on computing power, cloud services, and infrastructure. Because in the end, the one who makes money is often not the best storyteller, but the one who keeps collecting rent.
$AMZN
#亚马逊向OpenAI投500亿美元: Bet or bubble Microsoft changed data center lifespans from 15 years to 25 years—this is not accounting, it's magic
On July 29, Microsoft released its earnings report. Revenue reached 90 billion yuan, exceeding expectations. Azure grew 43%, exceeding expectations. The stock price surged nearly 9% in after-hours trading.
A wave of joy erupted.
But if you look closely at the financial report, there's a detail that keeps me awake.
Microsoft announced that it will extend the estimated service life of data centers and office buildings from 15 years to 25 years.
CFO Amy Hood said this reflects "our operational history and the anticipated use of our assets."
In plain language: we believe data centers can last 25 years, so there's no need to rush to depreciate.
And then? The 2026 capital expenditure guidance has been lowered from $190 billion to $175 billion.
The book value was 15 billion.
But the actual investment has not changed at all.
Wait, isn't this just a change in the number?
Yes. Purely accounting operations.
Extending the depreciation period → reducing annual depreciation expenses→ improving profit figures→ capital expenditure guidance "declining"→ the market feels "AI investment has cooled down?" Good news! ”
One operation, a win-win-win.
But here's the question—can a data center really last 25 years?
An AI chip can be replaced every 18 months.
GPUs are updated every two years, servers are phased out every three to five years.
You tell me a data center can last 25 years?
The chips installed inside have been replaced by more than ten generations, and the house is still depreciating?
Economist David Woo bluntly said: the AI bubble has officially burst. He is already building AI-related short positions.
Michael Burry is even more ruthless—he says these tech giants have inflated profits by 20% through accounting tactics. If the actual 2.5-year depreciation cycle is used, these companies' profitability would disappear immediately.
What's the most ironic?
In the fourth fiscal quarter, Microsoft's capital expenditure was $41 billion, up 69% year-over-year. In one year, 88 new data centers were opened.
Moody's has raised its capital expenditure forecast for the six major hyperscale cloud vendors for 2026 to $785 billion.
AI investment has not stopped but is accelerating.
But as the depreciation period is extended, the costs of these investments are "hidden" in the financial statements.
Short-term profits look good, but what about the long run?
Tell me 25 years from now that these data centers are still in use?
This reminds me of the 2000 internet bubble.
At that time, companies also used various accounting maneuvers to turn losses into profits.
And now? Hide depreciation, increase profits, and stabilize the stock price.
When the tide goes out, who's swimming naked?
Microsoft's real capital expenditure in one quarter is 41 billion, with an annualized increase of over 160 billion.
The GPUs, servers, and data centers invested in this money will never last 25 years in real life.
At that point, the issue will either be scrapped early and a one-time loss provision will be made, or the depreciation period will be "adjusted"—pushing the issue forward.
This isn't asset management; it's called passing the ball.
$MSFT $META $GOOGL #折旧年限延至25年, Microsoft's capital expenditure guidance was lowered Here are a few things to know before the market opens this week
Core signal: S&P Gamma has shifted upward from its lows, with the market betting on a continued rebound
U.S. stocks quickly rebounded 170 points from the plunge following the Federal Reserve's FOMC meeting. Gamma concentration has rapidly shifted upward from previous lows, with the largest positive Gamma currently concentrated at the S&P 7500 level, followed by 7550 and 7600, indicating that derivatives traders are quietly positioning themselves in the 7500 to 7600 range for next week's open.
The VIX has been declining continuously over the past three days, being pushed back below the 20 level, with the market returning to a typical volatility-compressed environment.
Geopolitics: Trump claimed to have withdrawn strikes against Iran, but Iran directly denied this
Trump posted on social media that the planned strike against Iran had been canceled, and stated that both sides had reached a framework for the agreement, including the full restoration of passage through the Strait of Hormuz and the elimination of Iran's nuclear threat. However, Iranian military and diplomatic officials refuted this claim, calling it psychological warfare and a list of illusions, emphasizing that no agreement has been accepted, the Strait of Hormuz remains under security control, and that the U.S. withdrawal of strikes is due to the exhaustion of Middle Eastern air defense systems. The U.S. State Department has issued a security warning to American citizens in the Middle East, indicating that geopolitical uncertainty remains high.
US Treasury yields soared, unexpectedly reigniting expectations of rate hikes
The yield on the U.S. 30-year Treasury surged to 5.27%, its highest level since June 2007. What is even more concerning is that after the Federal Reserve held it steady, yields have not fallen but actually risen. Currently, market traders have dashed expectations for a rate cut in September, now factoring in a 72% probability that the Fed will raise rates by 25 basis points, with the 30-year mortgage rate approaching 7% to 8%, putting pressure on real estate and consumer credit. Wall Street's confidence in the Fed's policy path is clearly wavering, which is the core macro variable to continue monitoring in the coming weeks.
For the first time in 15 years, the US and Japan jointly intervened in the foreign exchange market
After the Bank of Japan sold $59 billion to buy yen, the New York Fed and the Federal Reserve made a rare joint effort to sell EUR/USD to buy yen. This is the first joint intervention by the US and Japan in the yen exchange rate since 2011, and another coordinated central bank action following the previous Japan-South Korea joint intervention. Such frequent interventions disrupt the rhythm of arbitrage trading, and the pressure may further affect yield fluctuations in the Treasury market, which deserves continued attention.
S&P Technical Perspective: SPY 749-750 is a critical watershed
The above levels between 749 and 750 are strong psychological and technical resistance levels. Currently, during the night session at 751, let's see if tomorrow's opening can hold above the upper level. Previous attempts failed multiple times (including July 22 and last Friday's intraday pullback after touchdown). If it can truly break through, the market will quickly challenge the historic high of 760.40. Below, 740 to 741 is a short-term double bottom support. If it falls below it, the market could drop to 736 or even the previous recent low of 727.
Gold: Positive gamma recovery, major long capital inflows in the options market
GLD is attempting to form a larger bottom, Gamma concentration is turning positive again, and the options market has seen very clear net inflows (calls to increase holdings, puts to sell). The forward target is directly targeting the strike price of 400 or even 500, with single transactions totaling over $30 million. If GLD successfully breaks through and holds above the 370 to 375 resistance zone, it will continue to push toward 400, making it a valuable indicator to watch independently of the stock market. Now, the falling dollar and rising bond market are positive factors for gold.
This week, macro data is packed with earnings reports
On the macro data front, this week includes JOLTs job openings, ISM services PMI, and Friday's nonfarm payroll report, all of which will verify whether current rate hike expectations are reasonable.
Regarding financial reports
Palantir released an after-hours report on Monday.
Tuesday is the focus of the week, with AMD and SpaceX (the first earnings reports after their IPO), Arista, Astera Labs, and Kratos all releasing reports after the market closes.
I will focus on AMD's financial report. Currently, the stock price is fluctuating between 460-580, and has been trading sideways for two months. Gamma is mainly concentrated at 500 (maximum), 450, and 600. However, on Friday, net option capital flow was clearly negative, with put and call selling dominating. Negative gamma at 450 and 400 indicates a cautious market sentiment and is a key turning point. If the earnings report falls short of expectations, it could trigger a continued downward trend.
Wednesday was the earnings reports from SanDisk and Western Digital for the storage/semiconductor industry chain.
Before Thursday's market open, it was Datadog; after the close, it was AAOI.BTC🔥 rebound horn: ETF net absorption of 233 million, 65K bottom with an impact of 80,000
The bear market's corpse has grown cold, and the ox's hooves are digging the ground 🚀
———
📍 Price Sketch
Bitcoin hovered in the $65,200–$65,500 range, rebounding nearly 3% from last week's low. Retail investors remain anxious, while major players are quietly accumulating shares.
📍 Institutional blood bags
On July 30, spot ETFs saw a single-day net inflow of $233 million, with BlackRock IBIT taking $183 million; Immediately after, on the 31st, there was a brief outflow of -265 million, seen as a final release of selling pressure. Continuous flow shows that 'smart money' has completed a position replenishment above 60K.
📍 Macro Strategy
The Federal Reserve kept the rate unchanged at 3.50%–3.75%, but sent a signal of 'hawkish hold on to action.' Market expectations for a rate hike in September first surged to 77%, then cooled rapidly—policy uncertainty increased volatility and reinforced the narrative of "inflation-resistant digital gold."
📍 Leverage bloodbath
Within 24 hours on July 30, 286 million USD in derivatives positions evaporated, with both long and short positions declining, and OI plunging to a two-month low. After high leverage is flattened, spot buying is more likely to leverage the upward trend.
📍 Playing cards on the chain
Whale addresses > 37 new 1k BTC in three weeks (Glassnode data), with exchanges seeing a net outflow of 15,800 BTC—a typical cold wallet migration signal. Meanwhile, Strategy continued to hold 844,000 tokens worth 75,500 yuan, preferring to sell shares and buy back preferred shares rather than sell coins 💪.
———
🎯 Strategic perspective
Short-term resistance at 68K: Breaking below previous high without filling the gap, a breakout would trigger FOMO.
Main upward target is 80K: ETF & options "wall" focused, short Delta hedging will spontaneously ignite.
Extreme pull at 57.5 K: June low; breaking below would break the bull market structure.
I bet it will reach 80,000 within five weeks, with aggressive traders adding positions in batches on pullbacks of 62K-63K; Conservative investors can wait for the weekly chart to reclaim 68K before following.
The above does not constitute investment advice.
#BTC #Crypto #First US-Japan Joint Intervention in the Yen Since 2011
What does the US intervention in the yen signify?
1. Since 1998, this is the first time the US and Japan have jointly bought yen; the last coordinated intervention was in 2011 (to suppress the yen). 2. Exchange rate background: the yen has fallen to a nearly 40-year low, with the USD/JPY rate reaching a high of 163.73, then rebounding after the intervention. 3. Special operational point: the US sold euros to buy yen, an unconventional use of dollars.
II. The core economic motivations for US willingness to intervene (most important) 1. Prevent Japan from passively selling US Treasuries. Japan’s solo intervention in the forex market requires dollars; without the FIMA tool, it can only sell US Treasuries. Japan is the largest overseas holder of US Treasuries, and large-scale selling would push up US Treasury yields, impacting the US financing market. 2. Block cross-border risk contagion. Continuous yen depreciation triggers Japanese bond sell-offs, and rising yields would transmit to the global bond market, increasing long-term interest rate pressures in both the US and Japan. 3. Trade demands. The US has long believed the yen is undervalued, indirectly boosting Japan’s export competitiveness, and there is a demand to correct this imbalance. 4. Secure a policy window for the Bank of Japan. Stabilize the exchange rate in the short term while waiting for the Bank of Japan’s subsequent rate hikes (the fundamental support for the yen).
III. Geopolitical significance 1. Symbolizes an upgrade of the US-Japan alliance, fulfilling security and economic commitments to the ally; 2. Sends a strong external signal, serving as a strategic warning to China.#折旧年限延至25年, Microsoft's capital expenditure guidance lowered $BTC When the world's richest man says "money will no longer matter," and OpenAI's head says "inequality will still exist"—how should we interpret the signals behind this?
The tech community has been in an uproar these past couple of days. During an interview with The Economist at Tesla's Texas factory, Musk dropped a bombshell: by 2036, money will no longer matter. His logic is very hardcore—AI combined with humanoid robots will produce goods and services far beyond human consumption capacity. When resources become extremely abundant, money loses its purpose as a tool for "distributing scarce resources." He even predicted that the biggest economic problem in the future would be deflation rather than inflation.
However, Sam Altman was not convinced. When asked about this forecast, he said bluntly: "Clearly I think money will still matter in 10 years." Altman acknowledges that AI and robotics can bring tremendous prosperity and make many goods cheaper, but he points out a fatal issue—the concentration of wealth and power. He put it bluntly: "Money won't magically solve wealth inequality."
This debate precisely reveals the core value of cryptocurrencies.
Musk is depicting a "post-scarcity" world—when resources are infinitely abundant, fiat currency as a medium of exchange may indeed depreciate. But Altman's warning is closer to reality: the wealth created by technology does not necessarily flow to everyone automatically. The productivity explosion brought by AI is likely to benefit capital holders first, rather than ordinary citizens.
And this is precisely the very meaning behind the existence of Bitcoin and cryptocurrencies.
In a future where AI and robotics can produce infinitely, who will define value? Who will ensure fair distribution? Fiat currency can be printed infinitely, power can be highly centralized, but the Bitcoin limit of 21 million coins will not change. In a world where deflation may become the norm, scarcity itself will become the hardest asset.
Musk said money will no longer matter—he might be referring to fiat currency. Altman says money still matters—he might be referring to a store of value. When these two narratives converge in 2036, holding truly scarce assets may be the last line of defense for ordinary people against the "AI billionaire era."
What do you think? Ten years from now, will your asset allocation still be the same as it is today? Let's talk in the comments.590亿美元买来的假安全感：美日联合购汇砸穿156，套息交易的核弹几时引爆？
这次干预撑不了多久。
7 月 31 日到 8 月 1 日之间，日本财务省和美国财政部联手动用了高达 590 亿美元的外汇储备弹药，在公开市场上暴力购入日元，把美元兑日元从 163-164 的 40 年新高区域一路砸到了 155-156。这是 15 年来两国首次协同外汇干预，市场当天直接被打懵了。
散户和那些做空日元的投机客看到 8 个大数的瀑布暴跌，第一反应是恐慌。社群里全在庆祝"日元暴跌终于被遏制了"、"美日联手出拳比日本单方面干预猛十倍"、"协同干预是终极王炸，空日元的人全完蛋了"。大家觉得，只要美国爸爸下场帮忙，日元贬值的趋势就可以被逆转，套息交易者终将被扫出出局。
可如果你觉得 590 亿美元的蛮力就能改变全球资本的流向，那说明你完全忽略了驱动日元暴跌的底层动力引擎一直在全功率运转。
因为这笔干预改变的只是盘面的报价，而没有改变美日之间那道致命的利差鸿沟。
美联储的联邦基金利率依然锁死在 5.25%-5.50% 的历史高位，7 月议息会议按兵不动，内部甚至有多位理事投下反对票要求继续加息。而日本央行虽然在去年底终于结束了负利率，但其政策利率依然在 0.5% 附近趴着不动。
这之间存在着接近 5 个百分点的恐怖利差。
全球的对冲基金和宏观交易者，为什么要疯狂借入几乎零成本的日元去买美债和美股？因为这条利差提供了近乎无风险的套息回报。只要这道利差鸿沟不缩窄，日元作为全球套息交易的最大融资货币的宿命就不会改变，任何干预都只是在大坝上捅了个临时沙袋。
590 亿美元砸下去，大坝没有堵死，反而向全球投机客暴露了一个致命的信号：日本的外汇储备弹药库是有极限的。
这才是真正让我后脊梁骨发凉的地方。
一旦市场认清干预弹药即将耗尽、日元重新滑向 160 甚至更高的弱势区间，全球所有用日元做融资杠杆的套息交易者，将被迫在同一时间进行集体平仓——他们必须疯狂买入日元、同时抛售手里的美股、美债和一切风险资产来偿还日元借贷。
这就是所谓的"套息交易踩踏"（Carry Trade Unwind）。
而在这场跨市场的流动性大核爆中，24 小时全年不关门的加密市场，将第一个被砸穿。因为在所有资产类别里，加密的订单簿最薄、深度最浅，是全球大资金在紧急去杠杆时最方便、最快速的流动性"提款机"。
我自己在 2024 年 8 月亲身经历过一次日元套息交易的小型踩踏。当时日本央行意外加息 15 个基点，比特币在一个周末内从 6 万美元直接砸到了 4.9 万附近。我那天凌晨被强平警报吵醒，爬起来看手机，满屏全是红色，我慌得连裤子都没穿就跑到电脑前平仓。那一次我亏掉了接近两成的本金。前天看到美日联合干预 590 亿的新闻时，我没有感到一丝安慰，反而浑身发冷。因为我太清楚了——干预越猛烈，说明裂缝越大，等弹药烧完后的反扑，会比上一次更加血腥。
我当即把我所有加密持仓的杠杆降到了最低，核心仓位全部切换成现货，准备迎接这枚核弹的下一次试爆。
接下来的几周，死死盯住两个指标：美元兑日元能否守住 155 的干预底线，以及日本财务省外汇储备的月度消耗速度。一旦汇率重新滑向 160、且外储加速流失，套息踩踏的倒计时就正式启动。在那之前，管住你的杠杆，守好你的现货。
#美日确认联合购汇 1. Maximum macroeconomic suppression: US Treasury yields + Fed expectations
BTC and ETH are interest-free risk assets. Recently, concerns about inflation have resurfaced, leading the market to delay expectations for Fed rate cuts, and the 10-year U.S. Treasury yield remains elevated. Risk-free returns are rising, with funds leaning toward bonds, continuously suppressing crypto asset valuations.
Volatility in Middle Eastern crude oil has driven up inflation expectations, further reinforcing the idea that "high interest rates will persist longer."
2. Characteristics of Capital
BTC spot ETF capital inflows slowed significantly, with frequent large single-day net outflows; Institutional funds shifted from one-sided long to buying low and selling high.
The market lacks incremental funds, and the market is dominated by stock speculation and leveraged capital stamping back and forth, with sharp rises and falls becoming commonplace.
3. Market Style: Funds tend to hedge against leading coins, with BTC's market cap share continuing to rise; Altcoin liquidity continues to shrink.
2. Bitcoin (BTC).
Core positioning
Narrative: Digital value storage (digital gold), total supply of 21 million constant, crypto assets are the preferred choice for institutional allocation.
✅ Bullish logic
1. The long-term supply-demand logic of halving still exists, and medium- and long-term chip holdings continue to exist;
2. The U.S. spot ETF channel is now open, opening a long-term entry channel for compliant funds;
3. Recurring global fiscal debt and geopolitical conflicts have occasionally stimulated safe-haven buying.
❌ Bearish pressure
1. The high interest rate environment persists, greatly weakening the attractiveness of holding;
2. Price increases heavily depend on new capital and lack of organic cash flow;
3. Regulatory policies are uncertain, with the pace of U.S. crypto legislation progressing slowly;
4. Multiple failed attempts at key resistance above, gradually eroding bullish confidence.
Short-term market characteristics: range-bound fluctuation, resistance under pressure above, and medium- to long-term holdings taking hold below.
Features: Lower volatility than Ethereum, a market indicator, and an anchor in sentiment across the crypto community.
3. Ethereum and ETH
Core positioning
Public chain infrastructure, supporting DeFi, NFT, Layer 2, and RWA tokenization; After the merge, it will have a token burn mechanism and generate on-chain fee income.
✅ Bullish logic
1. The only dual narrative combining value storage + public blockchain; The Layer2 ecosystem continues to expand, with on-chain applications growing over the long term;
2. The market continues to debate the approval expectations for the US spot ETH ETF, which will bring incremental capital once implemented;
3. Network fees and burn mechanisms create native value capture.
❌ Bear pressure (key reason for recent weakness compared to BTC)
1. The ecosystem realization cycle is long, making it difficult to quickly convert short-term into price increases;
2. ETF approval progress is much more uncertain than Bitcoin, with policy expectations repeatedly pulling at the market;
3. Stronger speculative attributes, with a high proportion of retail funds; During market correction phases, the decline is usually greater than BTC's;
4. On-chain activity fluctuates in stages, lacking sustained explosive positive developments.美伊重回谈判桌，油价剧烈回吐：地缘风险溢价快速消退，80美元关口成关键 美东时间周日，特朗普在空军一号上宣布，应沙特等中东盟友及伊朗方面的请求，取消原定对伊朗的军事打击，转向双边谈判。 特朗普称霍尔木兹海峡“已有协议”，并确认谈判于周一（8月3日）下午正式启动。 伊朗方面同步确认缓和进展，表示调解方正协助恢复伊美谅解备忘录，且与阿曼就霍尔木兹海峡的谈判已进入最后阶段。 关键反转点：7月下旬曾出现短暂的缓和信号，但一周内被打破；本次缓和以正式双边谈判为支撑，持续性仍有待验证。 特朗普的“极限施压”时间线 2026年2月底 宣布对伊朗展开军事行动，强调持续施压 战争升级，供应中断担忧 油价快速上涨，风险溢价增加 3月 多次表示“目标即将完成”，但同时威胁扩大打击 不确定性高企 油价高位震荡，波动加剧 8月2日（周日） 取消打击，宣布谈判，称海峡“已有协议” 地缘缓和，供给预期修复 单日暴跌7.3%，溢价快速消退 未来关键看点 1. 谈判持续性：7月下旬的缓和曾在一周内被推翻，本次双边谈判能否真正推进，是市场反复定价的核心变量。 2. OPEC增产节奏：油价回落后，OPEC是否按原计划增产，#SPCX首份财报将公布, the $100 billion ban is about to be lifted
SpaceX earnings report eve: A decisive battle between bulls and bears that will determine the fate of the stock price
After tonight's market close, SpaceX (SPCX) will release its first earnings report since going public. This is not just a performance disclosure, but more like a key node in a multi-party game.
The stock price has dropped more than 50% from its post-IPO high and has fallen below $110. Bears are gearing up, with open interest exceeding $24 billion—a rare display in the US market.
Let's break it down and see what is the main focus tonight:
First, is Starlink, this "cash cow," still strong enough? This is the only profitable segment in the company, with over ten million users in Q1, contributing nearly 70% of revenue.
Tonight, the main focus is on user growth and average monthly revenue per user (ARPU). The ARPU in Q1 was $66. If this drops too much, it means the low-price strategy is eroding profits, and simply stacking users may not be enough to support valuation.
Second, has the rate of AI (xAI) burning money taper down?
In Q1, AI capital expenditure reached $7.7 billion, accounting for 76% of total expenditure. The market expects another 12 billion yuan tonight. Most of the money is spent on data centers, serving as the "foreman" for computing power at Anthropic and Google. The key is to see if there are long-term contract orders that prove the money can be burned to recoup costs
。
Third, the day after the earnings report, the "nuclear button" was lifted. On August 6, the first batch of 911 million shares restricted by insiders was unlocked. This volume is equivalent to more than doubling the existing circulating stock, and the short-term selling pressure is real. However, the market has also fallen for so long in advance, and some of the negative factors have been digested.
Operational Advice:
Tonight's earnings numbers may not be that important; what matters is management's guidance and the solid support after the lock-up. Morgan Stanley believes the AI business is undervalued and has set a $300 target price. But it's like a pancake—it takes time to verify. $SPCX $BTC /USDT is facing short-term selling pressure, trading at 62,655.4 (-1.45%) with $156.92M volume. Despite today's pullback, Bitcoin remains the market leader. Smart traders are watching key support levels for the next breakout opportunity. Volatility creates opportunity for disciplined traders.
#30YrYieldTopOrStart #USJapanYenIntervention #EarningsWeekAhead Last Friday, Trump even shouted on social media, "The gun is loaded," threatening to launch an "unprecedented military strike since World War II" against Iran. The U.S. State Department even issued a security warning to American citizens in the Middle East, advising them to "consider leaving the area." And then? In less than 48 hours, the script was completely twisted. Trump announced on Air Force One: Strike canceled, negotiations on Monday. Saudi Arabia, the UAE, and Qatar have collectively mediated peace, and Iran has also "requested cancellation." As soon as the news broke—Brent crude plunged 7.3% intraday, hitting a low of $81.55. WTI crude fell below $80. July's nearly 25% monthly gain was given back by nearly a third in a single day. What about Bitcoin? It has gone up. Breaking through $63,000, Ethereum rose over 2%, and SOL increased over 3%. US stock futures surged, with gold breaking through $4,080. Social media erupted in cheers: "Peace is here!" Risk assets take off! " Don't rush. This "peace" for Bitcoin is not so simple. Breaking it down, there are two completely opposite transmission paths: Path A — Positive news. Oil prices plunged→ inflation expectations fell → opened room for Fed rate cuts→ liquidity improved, → risk asset valuations rose. At the July Fed meeting, there were already three opposing votes. Why? Because of oil prices. Oil prices have surged to $100, CPI has rebounded directly—how will the Fed cut rates? Now that oil prices have plummeted, the logic for rate cuts has been reestablished. Path B — Negative news. Geopolitical conflicts cancel → risk aversion cools → gold and BitcoinBTC community temperature update: speed 0.77x, current bulls and bears close
Putting BTC's short-window numbers together with the full-day average makes the picture much more complete than just looking at the popular rankings.
On August 3rd, at 13:00 (China time), OKX Onchain OS recorded 41 mentions of BTC in one hour, including 37 times on X and 4 times in the news; The total 24-hour volume was 1,278.
After conversion, the latest hour is 0.77 times the hourly average for Long Window, which is about 23% lower than the 24-hour average. This ratio only answers whether discussions have heated up, not whether buying has increased. If you write it directly as a breakout signal, you take an extra step and make an inference that the data does not support.
The structure of tone is another line. Within one hour, 29% are slightly bullish, 32% bearish, and about 39% neutral, which is considered 'close to bull-bear'; Within the 24-hour period, the trend is 24% bullish and 34% bearish. The gap between the short and long windows is the part worth tracking going forward.
On the source side, BTC is currently mainly driven by X. When a message is widely shared, mentions quickly increase, but independent information may not necessarily increase year-on-year. The trending list cannot tell us whether each piece of text comes from different participants, nor does it weigh by account influence or fund size.
Long window sources can be used as background: BTC has 1,132 times in 24 hours, 146 news events. If the proportion of sources in one hour suddenly deviates sharply, it could mean new news first broke out on a certain channel, or news updates just haven't caught up yet. Both explanations are reasonable, so we still need to wait for the original announcement or the next round of source distribution confirmation.
I would treat Bullish and Bearish as thermometers under the same ruler, not as exact voting. There is a lot of neutral content, usually just everyone watching and not yet forming a unified direction; An increase in bearish bias may also mean more risk discussions, but it doesn't mean every poster has truly established a short position.
The next step to observe is whether spot trading volume expands, whether perpetual contract funding rates and open interest are moving in the same direction, and whether liquidations are concentrated. These three sets of data answer real trading participation and leverage structure, and cannot be replaced by community mentions. If there are macro or industry events, the official original text should be directly verified.
How did I know I was mistaken this time? If the next round of BTC mentions returns to near the average and the gap between bullish and bearish will narrow, this change will likely be just short-term noise. Conversely, two consecutive rounds of increased speed, expanded news sources, and simultaneous increases in spot and derivatives transactions are more like the main market theme is taking shape.
You also need to keep the intraday difference. The community activity levels differ naturally between early Asian trading, US trading hours, and near major announcements; A single 0.77x is not suitable for annualization, nor should it be used for hard comparisons with raw counts from other platforms. Continuous snapshots are more useful than a single beautiful number.
So I first recorded BTC as "discussions have slowed down, and the tone of the short window is close to bulls and bears." The official rankings stop here, with no proof that funds are betting in the same direction. If the next round also improves both the diversity of sources and market transactions, it won't be too late to raise confidence in judgment.On March 26, 2021, Bill Hwang, the world's fastest-losing man, was liquidated, holding a large amount of Chinese concept stocks with leverage, but that was when Chinese concept stocks only fell 27% from their peak.
On July 31, 2026, Wall Street's 24-year-old stock legend Leopold was forced to liquidate his holdings, heavily investing 3-4x leverage in AI. The price change of thematic funds was similar to SOXX, dropping 29% from its peak.
There was also the once-popular Wooden Sister, who once shone brightly, but after the tide receded, many of them faded away. Every few years, a certain sector surges, and at the same time, a celebrity stock god appears, but the good times don't last long. Looking back, it's actually no different from Dutch tulip foam.
When the heat turns red and purple, pay attention to protecting profits. There was before, there is now, and there will be more in the future.
Today, I stopped my loss on a tech stock at the open. Last Friday, I was oversold, but it didn't hit the daily limit on Friday, so I stopped at the opening today.Brothers, extra! Extra Edition! Presidential Media is also a big account of chives!
Trump Media's move is even more reckless than retail investors.
At an average price of $118,522, 11,542 BTC was bought, spending $1.37 billion.
After seven months of continuous selling, it has already sold 7,281 coins, with an average selling price of only $74,855.
Gang reportedly sold 2,628 coins, about $165 million.
The total book loss was $555 million.
This isn't allocation to Bitcoin; it's paying tuition to the market, and even a presidential-level large $ETH allowance #亚马逊向OpenAI投500亿美元: Bet or bubble
Recently, there was big news: Amazon gave OpenAI $50 billion in exchange for a $100 billion commitment to spend $100 billion on Amazon Web over the next eight years. Many people don't understand: investing 50 billion just to earn 100 billion? Isn't this just left hand switching to right hand?
In fact, this money was not an investment at all, but a "computing power consumption voucher" issued by Amazon to OpenAI.
OpenAI lacks computing power the most when building AI models, just like humans can't live without electricity. Amazon holds the world's largest cloud service (AWS) and self-developed chips, and OpenAI cannot do without them.
The $50 billion Amazon gave OpenAI most of the money cannot be spent casually; it must be used to buy Amazon's cloud services and chips. It's like "suppliers invest money in customers, and customers then use the money to buy the supplier's products." For Amazon, this 50 billion yuan is not an expense, but rather locking in 100 billion in revenue in advance, and even selling its own chips.
This isn't betting on whether OpenAI will succeed, but betting that "AI needs computing power" will never change. Whether OpenAI wins or loses, as long as AI continues to develop, Amazon's cloud services can still make money.
Many people worry: Amazon is holding preferred shares and will have to wait until OpenAI goes public before converting to common shares. If OpenAI never goes public, won't the 50 billion be wasted?
Actually, there's no need to worry too much. Preferred stock is more stable than common stock; even if OpenAI doesn't go public, Amazon can recoup its costs through cloud orders. Moreover, OpenAI's valuation is already very high. Even if it doesn't go public, it could be acquired or repurchased shares, so Amazon's money won't be hard to get back.
The real risk isn't OpenAI not going public, but whether OpenAI can actually spend $100 billion to buy Amazon's services. If the AI industry cools down, OpenAI won't need so much computing power, and Amazon's investment will lose money.
This deal is not a bubble, but a clever move by Amazon. It is not betting on OpenAI itself, but rather on the trend that "AI needs computing power."
Ordinary people looking at this deal don't need to understand preferred shares or IPOs; they only need to focus on one indicator: whether Amazon Web Services' revenue is steadily growing. If revenue keeps rising, it means OpenAI is truly spending money on services, and the investment is a success; If income doesn't increase, it means there may be problems with the order, and the investment carries risks.
1. Don't chase after the "AI concept"—look for "real profit": Many AI projects only tell stories, not make money. When investing, don't just listen to their "future prospects"—look at whether they currently have real customers and can sell their services.
2. Pay attention to "water sellers," don't chase "gold miners": In the AI industry, companies selling computing power and cloud services (such as Amazon and NVIDIA) are more stable than those making AI models. No matter which AI company wins, it must use their computing power. Ordinary people investing in AI should first look at these "water sellers" and avoid betting on any single AI company.
3. Don't trust book numbers—look at the "real money": tech company valuations and capital expenditures are easily beautified by accounting. To judge whether a company is reliable, don't just look at its valuation—check whether it has real income and healthy cash flow. Amazon's $50 billion investment is essentially about locking in orders with cash flow—that's its greatest ingenuity.
Simply put, Amazon's 50 billion is not foolish money, but smart money that binds customers. Ordinary people looking at this deal don't need to understand technical jargon; they just need to focus on whether AWS revenue is growing. The investment logic in the AI industry has shifted from "believing in the future" to "verifying the present." Only companies that can truly make money can survive. 晚饭桌上聊AI，很多人第一反应还是大模型、显卡、机器人。可今天这条消息，主角不是一个会聊天的软件，也不是一台会走路的机器，而是一家做光互连的芯片公司：量引科技近期完成天使轮数千万元融资，由珠海科技产业集团领投，珠海正方集团、险峰跟投，钱将用于扩充团队、迭代流片和补充设备。 它影响的不是某一个普通消费者今晚买不买手机，而是AI算力链条里一群更靠后的角色：GPU厂商、封装厂、数据中心、光芯片团队，以及所有被算力成本牵着走的应用公司。说白了，前台越热闹，后台的数据搬运就越吃紧。 读者现在最该核对三件事：第一，量引科技成立于2024年，聚焦硅光PIC、CPO和OIO；第二，公司称1.6T微环调制器芯片已经完成流片，目前仍在测试；第三，与国内GPU厂商合作处于PoC阶段，并在与封装厂推进3D堆叠联合研发。流片不是量产，PoC不是订单，这两个边界很重要。 【真正拥堵的地方，不在屏幕上】 我们平时用AI，感受到的是一句话生成一段文案，一张图几秒钟冒出来。体验像点外卖，按钮一按，东西送到。 但机房里不是这样。 那里更像一个巨大的后厨。GPU在烧火，存储在备菜，服务器在传菜。菜做得再快，如果传菜通道堵了Strategy $MSTR 二季度亏了86亿美元。
第一眼是真吓人。
但这 86 亿里有 83 亿是比特币的账面浮亏。就是它手里那堆币二季度跌了，按规矩没卖也得当亏损记上。真正做生意收进兜里的，才1.22亿。
有意思的是，它那个没人搭理的软件老本行，订阅收入这季度偷偷涨了54%，被比特币的巨亏盖得死死的没人看见。更离谱的是，一边亏成这样，它一边还在加仓。二季度又买了8.4万枚币，总持仓奔着84万枚去了。
所以这公司到底该值多少钱，华尔街自己都吵翻了。
执行主席Saylor一如既往地头铁。有人问他怎么看做空的，他说他根本不想Chanos那个人，对方压根没搞懂比特币在干嘛。Chanos 是华尔街的老空头。他早几年就玩买比特币、做空 MSTR，说白了同一份东西MSTR敢卖你两块五，他就赚这个溢价迟早收敛的钱。
现在还真被他说中了。MSTR股价对它手里那堆币的溢价，从去年11月就基本趴在一倍以下，以前那层让股东白赚的水分快没了。当然也有看好的。Cantor 那边的分析师给了增持，目标价 212 刀，理由是它现金变厚了、还能继续融到钱，大家之前最怕的爆雷缓了一大截。
现在买MSTR，基本约等于买了个带杠杆的比特币，涨跌都比现货猛一截，还得多扛它优先股和债务那一层。想赌比特币弹性、又不怕多担点风险，它是个趁手工具。只想干干净净拿点比特币敞口，那还不如直接买币省心。#30年期美债, the top or a new beginning?
The 30-year U.S. Treasury yield broke through 5.2%, hitting a nearly 19-year high, and above 5% has become the new normal. The market is in turmoil, with one side calling for the top and the other calling for a new starting point.
The logic behind calling for a top is simple: if the economy can't hold on, interest rates will naturally fall.
What does a 5.2% long-term interest rate mean? The 30-year mortgage rate is approaching 8%, and long-term financing costs for companies are soaring. High interest rates themselves are the strongest tightening policy; the Fed doesn't need to raise rates again—the market has already done it for them. BOC Hong Kong's view is very straightforward—the rise in bond yields itself is equivalent to the effect of raising interest rates. Rising corporate financing costs and increasing mortgage pressure for households will ultimately backfire on the economy, forcing the Federal Reserve to pivot.
The logic behind calling for a new starting point is also very strong: global money is running low, and the interest rate center is permanently raised.
Goldman Sachs released a report this week with a headline that was explosive—"This is the largest capital demand cycle in human history." AI infrastructure, reindustrialization, defense restructuring, and sovereign debt—four demand curves are all surging upward. The era of "excess savings" over the past few decades is coming to an end, and capital has shifted from abundance to scarcity. Mark Wilson, Head of Goldman Sachs' European Hedge Fund Business, put it in words: "We are in the most capital-hungry investment cycle ever."
What's even more critical is that the Federal Reserve is actually making things worse. After Walsh took office, he deliberately reduced forward-looking guidance, leaving the market completely unclear about the policy path. In July, the FOMC kept rates unchanged, but three opposing votes supported a rate hike, eroding market confidence in the Fed's ability to fight inflation. Even the St. Louis Fed President came forward to call out, urging the Fed to quickly rebuild its anti-inflation credibility. If the market begins to doubt the Fed's ability to control inflation, long-term interest rates could spiral out of control even further.
My judgment: In the short term, it is near the top area, but it is not a turning point where you can go all in to buy the bottom; rather, it is more likely the starting point of a new central zone for high-level consolidation.
The underlying logic supporting the high level has not reversed. Fiscal deficits continue to widen, long-term debt supply continues to rise, willingness of major overseas buyers is waning, the Middle East situation remains volatile, and energy inflation remains looming. These factors make it difficult for long-term rates to return to the low levels seen in previous years.
But calling it a "new starting point" is too absolute. The 5.2% yield is already backlashing on the economy. Continuing to short long-term bonds at this level is already not very attractive. A real trend turning point requires a dual confirmation of the continued decline in core inflation combined with a clear economic slowdown.
There's also a huge geopolitical variable in between—the effective blockade of the Strait of Hormuz, with Brent crude surging to $96.6. For every 10% increase in oil prices, inflation expectations jump along with them, and long-term bonds get dumped again. As long as the US-Iran conflict does not subside, the selling pressure on long-term debt will not truly end.
Refuse to be black or white. In the short term, there will be repeated high-level fluctuations, so it is not advisable to heavily position or bet on the direction. At this level, we are waiting for both inflation and economic confirmation, not to guess the top or bottom.The Korean stock market fell another 5%, but I think it's too early to say the "storage bull market is over."
The most interesting part is:
Last Friday, the KOSPI surged 17.9%, and today it dropped another 5%.
Samsung and SK Hynix also plunged sharply.
This no longer feels like normal trading; it feels more like funds trampling each other.
Now, two completely opposite signals have appeared in the storage market:
Observers see:
AI is still aggressively buying HBM, Samsung is making record profits, and SK Hynix is also recording revenue.
Bearish people see:
With valuations too high and capacity expansion too aggressive, Chinese storage companies are catching up again.
So my judgment now is simple:
The storage bull market may not be dead yet,
But those speculating on storage may be the first to be washed out by high leverage.
This is also why a 18% increase in one day can still fall 5% the next.
What really matters is not how much it dropped today.
The question is whether, after this round of trading, Samsung and SK Hynix can continue to set new records.
If earnings continue to rise, today's sharp drop feels more like a shakeout.
If orders and profits start to decline, that's the real turning point in the cycle.
For the crypto world, this also offers some reference:
If the main theme of AI chips—the strongest risk asset—continues to cool down, high-beta assets like $ETH and $SOL will find it difficult to remain completely affected.
I haven't been bearish on AI yet.
I just feel that the market no longer allows you to "tell only stories, ignore prices."
$ETH $SOL #韩国股市 #AI芯片
#韩股重挫5%, storing long-short signals in a standoff $HYPE HYPE Fixed at 52, Waiting for Forced Movement
HYPE continues to steady and move in the band at $52, currently price at $52.34, down 0.85% in the last 24 hours. After a succession of days of weak declines, it has pulled back more than 30% from its June high of $78.
The news was mixed with positives and negatives. On the negative side: this week there are ~$22.74 million HYPE token unlocks (although the actual claim is below the limit), creating potential supply-side pressure; and growing community conversation about the possibility of TradeXYZ "going alone" — a platform that controls more than 90% of Hyperliquid's HIP-3 volume.
On the plus side: most of the protocol's revenue is used for buying and burning HYPE, with millions of dollars coming in every day. On the technical side, $52 is an important support level — the 32% pullback from ATH — and with the RSI approaching oversold territory (38-39), a technical bounce is possible.
At $52, if you stay, look at $55–56; If it falls, the next level will be $50 and then $48. The short-term direction depends on the catalysts — without it, a continual grind will occur.
#30年期美债,顶部还是新起点? #美日确认联合购汇 $BTC #财报观察员:本周四场开奖,Circle压轴 $ETH Low circulation, high FDV, and dense unlocks—these three words together form the scenario where retail investors are most easily "gently harvested." Do you know why many coins seem cheap but become more anxious the more you buy them? The list of unlocking stress that the community has been repeatedly mentioning lately—I glanced at it: $ARB $OP $STRK $ZK $BLAST $MANTA $ALT $DYM $TIA $SUI $APT $SEI $PYTH $JUP $W $EIGEN $REZ $ETHFI. To be honest, this list has grown long enough to be like a "guide to persuading people to quit." But what really concerned me wasn't the list itself, but the market's reaction to it—everyone clearly knew the unlock was coming, yet they couldn't help but bet on a rebound before the unlock and then saw the price get crushed on the day of the real volume surge in volume. This is the most subtle part. On the surface, it seems everyone is "trading to unlock expectations." But what is actually being traded? The question is, will those who receive the tokens sell immediately? If the unlocking targets are teams and early-stage institutions, selling pressure is almost a certain event; If the unlocking target is an ecosystem fund or a market maker, the price may actually be protected. The same event leads to two completely different endings, which is why looking only at the "unlock date" is meaningless; you must clearly see whose pockets the chips are going to. Looking deeper, the sentiment in this round of market has actually been split in two. Half of the people are focusing on new coins with low circulation and high FDV, knowing that this structure cannot sustain sustained growth🔥 Liquidation Heatmap + In-depth Analysis of Long-Short Position Ratio (2026.8.3)
1. Overview of 24-hour full-scale liquidations across the entire network
Dimension data
Total liquidation across the entire network: approximately 282 million ~ 606 million USD (by scale: CoinGlass is about 282 million, with some platforms including weekends totaling 606 million)
Long-short ratio: Short liquidations dominate (62%~83%), typical of a short squeeze market
Number of liquidations across the entire network: about 54,000~95,000 liquidated
The largest single long BTC order on Hyperliquid was about $2.76 million
📌 Interpretation: Today is a pattern where bears are being wiped out. Geopolitical easing (US-Iran negotiations) → Oil prices plunge → Risk assets rebound, triggering a chain of short blowouts and forming positive feedback to push the market upward.
------
2. Key ranges in the BTC liquidation heatmap
According to Hyblock / CoinGlass liquidation heat data:
🔴 Above is a dense short position zone (a breakout could trigger a short squeeze)
• $64,000–$65,000: Short liquidation tight band; bearish stamping after a breakout can provide upward momentum
• $65,800–$66,000: Deeper cluster of short stops
🟢 Below is a dense long position zone (breaking below triggers a bullish stampede)
• $62,000–61,000: Dense long liquidation band; once broken, it will accelerate long liquidation
• $60,000: Psychological threshold + chip support
⚪ Current price zone (lower liquidation pressure)
• Around $63,000: Currently in a relative "vacuum zone," with low liquidation pressure, belonging to a tug-of-war zone between bulls and bears
📊 BTC 24H liquidation distribution (CoinGlass real-time)
• Long liquidation: approximately $34.38 million (56.15%)
• Short liquidation: approximately $26.84 million (43.85%)
• Exchange rankings: Binance 32.3% > Hyperliquid 25.62% > Bybit 14.07%
------
3. Key ranges in the ETH liquidation heatmap
🔴 Above is a dense short position zone
• $1,900–$1,930: ETH short positions are concentrated in the band; breaking through 1,900 is a key short squeeze
• $1,950–$1,980: Deeper resistance + liquidation zone
🟢 Below is a densely packed area with multiple orders
• $1,830–1,850: Dense long liquidation zone
• $1,780–$1,800: Neckline support + liquidation cluster
📊 ETH experienced a 24-hour liquidation
• Approximately $55.2 million of liquidations across the entire network, with short positions accounting for the majority
• Funding rate: 0.0041%~0.0065% (low near zero, no extreme greed)
------
4. Analysis of the Long-Short Position Ratio
🐋 Whales/Whale Holdings (Key Signals)
Indicators BTC and ETH bias
Binance Large Account Long-Short Ratio 1.63 (Long) — slightly bullish
Binance large account number of long/short accounts is slightly higher than 1.79 — slightly higher
OKX's long-short ratio is 1.72 — slightly higher
Hyperliquid whale positions long 48.57% / short 51.43% — slightly bearish
Whale ETH took a 5x long position at $3,147, with an unrealized loss of $4.02 million — ⚠️ dangerous
📊 Retail Investors vs. Institutional Divergence (Key Contradiction)
• Retail investors ($1K–$10K range): Consistently buying spot on dips + opening long, with a bullish sentiment
• Whales/Institutions ($1 million–$10 million tier): Net selling during rebounds, suppressing rebound heights
• ETH funding rate nears zero + open interest up 2.33% → Participation is rising but the direction remains unclear
🎯 Exchange Long-Short Ratio Overview (Real-time)
• Binance BTC/USDT Whales: Long Positions Dominate (1.63)
• OKX BTC user ratio: 1.72, slightly higher
• Overall sentiment: Fear and Greed index 27–28 (fear range), sentiment is somewhat pessimistic
------
5. Comprehensive Assessment and Operational Insights
⚠️ Core Contradiction
Retail investors desperately buying long positions vs. whales continuing to distribute — this is currently the biggest risk point. The rebound is driven by short squeezes, not by incremental funds actively going long.
🔑 Key price levels to watch list
Variety: Bullish Defense Line, Short Stop/Squeeze Point Signal
BTC 62,000–62,400 64,000–65,000 holds above 62K, turning bearish
ETH 1,830–1,850 1,900–1,930 Breaking above 1,900 confirms a short squeeze
💡 Practical advice
1. Upside potential: If BTC can break through 64K–65K with increased volume, short liquidation fuel is ample, so you can follow the trend to chase long, targeting 65,800
2. Downside risk: If it falls below 62K, the chain reaction of long liquidation will be ignited, possibly quickly testing 60K
3. Position warning: Hyperliquid whales have already posted losses on 5x long positions at ETH highs; if large players are forced to close their positions, it will trigger a waterfall
4. Position discipline: Currently, retail investors are crowded to long + whales are distributing, so it's not advisable to chase highs. It's safer to wait for pullbacks to support and low bullish positions
📌 In short
Today is a bearish squeeze rebound. The liquidation heatmap shows that the 64K–65K (BTC) and 1,900–1,930 (ETH) levels above are key short squeeze zones; However, the divergence from heavy retail investors and whale short positions indicates a weak rebound. The 62K (BTC) and 1,830 (ETH) levels are lifelines; if a break breaks, immediately turn to bearish.Crowding and Crowding List
This group doesn't rank by rate, but only seeks high-cost positions and their price feedback.
$MMT Current rate -0.1343%, closing -0.573% in the past 24 hours, at the 7th percentile of the most recent sample. Prices are down, and so are positions; the ebb tide of position is more certain than directional attribution. OI contraction indicates that risk exposure is being withdrawn; the fee rate only indicates which side has higher costs and cannot replace the detailed exit directions.
$BICO Current rate -0.0922%, closing -0.097% in the past 24 hours, at the 0% percentile of the most recent sample. Prices rise in sync with holdings; short-term trading is not just about replenishing old positions. Prices and OI rose in tandem, while rates remained negative, making this misalignment more sensitive to bears.
$SNDK Current rate -0.0309%, closed in the past 24 hours +0.018%, at the 3rd percentile of the most recent sample. Expanding positions while falling is accompanied by selling pressure and new positions, but OI alone cannot confirm the direction of short positions. Rates have shifted from past directions to the other side, and position sentiment is rearranging itself; The signal becomes more complete after expanding the position.#美日确认联合购汇
这次美日联手干预日元，美国为什么愿意亲自下场？过去几年大量资金借入低利率日元，去买美股、AI、BTC等美元资产，这就是大家常说的日元套息交易。美国真正担心的是，日元如果一直贬值，套息交易会越滚越大；等哪天日元突然大幅升值，资金集中平仓，冲击的就不只是汇市，而是全球金融市场。
所以这次美国一边支持日本稳定汇率，一边扩大FIMA工具，就是希望日本在买入日元的同时，不用大量抛售美债筹集美元，尽量把对美债市场的影响降到最低。
不过这只是在控制节奏，而不是改变趋势。只要美联储维持高利率、日本加息依然缓慢，美日利差还在，资金最终还是会流向美元。汇率干预可以稳定短期情绪，却很难改变长期方向。
这次更像是给市场争取一个有序退出的时间窗口，而不是重演2024年那种套息交易快速踩踏。接下来需要盯的还是美国非农、美联储政策和美日利差。I'm glad this article was featured in the official eighth issue of the weekly report. Let me briefly share the writing process. When I wrote this article, it happened that both giants were releasing their financial reports on the same day, and the contrast between the ups and downs was so obvious that I thought I'd put both sides together to see what exactly the difference was.
When writing, I ponder and ponder; what I fear most is just listing data. Later, I found a different approach to AI implementation. From this perspective, the article suddenly had a main thread and became much smoother to write.
The main point is to express that while both companies are talking about AI, Microsoft and Meta are taking completely different paths. One leverages existing business to boost efficiency, while the other is betting on the future but is currently under pressure. In fact, as long as you find a way to dig deeper and present your own viewpoint based on factual evidence,$SNDK $MU $SKHYNIX Big differentiation among the three storage giants! Micron Hynix was dumped 100 million yuan, while SanDisk secretly accumulated 36%. What signal is this?
Today, the storage world is experiencing a dramatic twist and bustle. Hyperliquid's data shows that Micron and SK Hynix lost a combined $103 million in holdings in the past day, while SanDisk bucked the trend by raising over $32 million in cash and its holdings surged by 36%. On the surface, the stock prices of the three companies seem to fluctuate little, but on-chain capital movements have already revealed the true attitude of the main players.
Why has this kind of differentiation occurred? External news is also coordinating—Apple CEO Tim Cook has publicly stated plans to increase DRAM suppliers, clearly aiming to break the monopoly of the three giants: Samsung, SK Hynix, and Micron. Meanwhile, the Korean stock market crashed today, with SK Hynix dropping more than 7%. All the negative news is piling on these two companies, so it's no wonder funds are flowing out.
In contrast, SanDisk's large address only added $2.51 million in multiple orders, with no one opening a short position. Although some are also adding to Micron, the bears are simultaneously increasing their positions, making it a strong hedging between long and short positions. SK Hynix is in worse shape, with whales all fleeing and even short sellers reducing their positions.
My view is that capital is shifting from the DRAM track to NAND flash. It's not that the sector as a whole is failing; it's that internal pricing is being reconsidered. Today, Bitcoin also stood above $63,000. Against the backdrop of the entire crypto market warming up and increased risk appetite, this kind of structural rebalancing deserves even more attention.
What happens next? Keep an eye on whether SanDisk's long positions can continue to expand, while also watching whether the short positions in Micron and Hynix are still increasing their positions. In terms of direction, let's skip DRAM for now; we can take a closer look at the NAND line.
Want to be the first to catch the main players' position adjustment movements? Give me a follow, and I'll keep an eye on the data for you. #30年期美债, the top or a new beginning? #美日确认联合购汇 #财报观察员: Four draws this week, with Circle as the grand finale Why does the crypto market tremble when the USD and JPY intervene together and the yen rises?
The US and Japan have confirmed joint currency purchases, with market expectations that the action mainly supports the yen and limits the continued rise of the USD against the yen. On the surface, a weaker dollar is beneficial for BTC.
However, many funds have borrowed low-interest yen to buy US stocks and crypto assets. If the yen suddenly appreciates, these funds may be forced to sell assets to repay loans.
Therefore, this news may not be bullish in the short term. Market outlook: initially bearish, focusing on whether BTC can hold $62,000.
Personal view: if it breaks below $62,000, go short accordingly, stop loss at $62,800, target between $60,500 and $60,000.
If the yen stabilizes and BTC climbs back above $64,000, then switch to bullish on $BTC
#美日确认联合购汇 #Coldcard安全事件升级, fourth wave attack warning
Industry institutions estimate that about 7 million $BTC will become targets of future quantum attacks due to public key exposure on-chain, with a total value of approximately $470 billion. However, there is no need to worry in the past; exposing public keys does not necessarily mean assets will be stolen, because a quantum computer capable of cracking Bitcoin's private key has yet to appear.
However, it is important to be cautious that technological advances are continuously shortening this window of safety. Google Research has issued a statement stating that the number of qubits needed to crack related cryptographic systems has dropped significantly; Ethereum researchers estimate that by 2032, the probability of quantum computers successfully cracking exposed public keys could reach about 10%. Currently, the lab can only crack extremely short test keys, which are still far from the 256-bit real key.
The Bitcoin community is also continuously discussing and researching related upgrade plans:
BIP-360 plans to introduce quantum-resistant addresses, encouraging users to proactively migrate assets.
BIP-361 is even more aggressive, advocating for phased phased phases out of legacy signatures and ultimately freezing wallets that have not migrated for a long time. Supporters of this plan argue that freezing dormant assets is better than having quantum hackers steal them and throw them into the market; Opponents, on the other hand, worry that this would be equivalent to making decisions for asset holders and could even cross Bitcoin's decentralization red line.
Of course, a group of startups have also started to seize the quantum-resistant track. American Fortress claims to add quantum protection to multiple blockchains without changing existing addresses and automatically freezes high-risk wallets before an attack occurs. However, its core papers have not yet been made public and the technology lacks independent audits, so it currently appears to be more hype and hype.
Quantum threats are a real test the Bitcoin network must face—not just upgrading cryptographic algorithms, but also finding consensus on security, property rights, privacy, and compliance.
It is believed that the Bitcoin community can handle this "quantum threat" crisis perfectly.#韩股重挫5%, storing long-short signals in a standoff
Korean stocks plunged 5%, with long-short signals locked in a standoff: it's not that the cycle is dead, but that leverage has died first
On the morning of August 3, South Korea's KOSPI fell more than 5% at one point, with Samsung Electronics and SK Hynix both dropping over 8% at the open, and Hynix hitting nearly 9% at its highest.
Just last Friday (July 31), KOSPI surged 17.91%, marking the largest single-day gain in history, while SK Hynix sealed a 30% limit-up—moving from the daily limit up to near the limit down within 48 hours. The two Korean stock giants have perfectly embodied the term "sentiment amplifier."
There are two surface trigger points:
• Kioxia's Q1 operating profit was 1.27 trillion yen, below the expected 1.37 trillion yen; US stocks in storage stocks fell first (Kioxia ADR -10%, Micron -5.9%, Hynix ADR -3.5%)
• Margin requirements for Korean individual stock leveraged ETFs increased from 10 million won to 30 million won starting July 31, causing leveraged funds to stamp out the market
But at the bottom is the first head-on split that stores long/short signals:
The bulls still hold firm
• HBM3E/4 gap of 50%–60%, with the HBM market expected to reach +58% in 2026 to $54.6 billion, accounting for nearly 40% of DRAM
• Samsung/SK Hynix/Micron have locked 70% of their advanced capacity into HBM+DDR5, squeezing consumer-grade DRAM/NAND to the point of no increment
• SK Hynix's Q2 revenue reached 79.32 trillion KRW and operating profit of 60.54 trillion KRW, both hitting record highs; Morgan Stanley also claimed "over 40% upside potential"
The bearish hand only came out this week
• DRAM contract price gains have risen quarterly from 93–98% in Q1 → 58–63% in Q2→ with Q3 projections of only 13–18%, a steep cliff
• Mobile Side Defects First: OPPO/vivo Reject Samsung's Q3 Offers, Top Domestic Manufacturers Follow Suit, Huaqiangbei DDR4 Drawdown About 35% in Month
• NVIDIA Rubin Ultra preview HBM4 has been reduced from 12-Hi 384GB to 8-Hi 192GB, and AMD MI455X removed LPDDR5X—it's not that AI doesn't want storage, it's that it can't afford to use so much
• Morgan Stanley Bluntly Says: Memory contract prices peaked in Q4
My views are straightforward:
This time, the Korean stock market's -5% drop isn't a falsification of the supercycle storage, but rather a temporary expiration of "flexible price hike trading" + a reshuffling of leveraged chips.
The medium-term scarcity of HBM is real, but the narrative of "a full doubling of DRAM prices" has reached the stage where customers vote with their feet. Under the three-way game of "original manufacturers holding onto pricing power (Samsung DS refuses MX long-term contracts"), buyers refusing orders, and system manufacturers downgrading allocations—Q3 began with "stable price and increased volume, but valuation downgrade," not "prices keep soaring and stock prices continue to multiply."
For those exposed to memory like XSKHY (SK Hynix ADR tokenization), XMU (Micron), and RAM, two things need to be clarified these days:
• Underlying stocks in Korea fell 8% on Monday, but XSKHY is a Nasdaq ADR. Weekend+ partially priced before market opening. Don't blindly short tokens just because KOSPI drops -5%; price differences can bite you
• For perpetual DRAM ETFs like RAM, volatility decay during sideways trading is even more devastating than a wrong direction. Don't use "long-term storage logic" as a reason for "leveraged trades can stay overnight."
Here's my personal arrangement for operation:
1. From August 3 to 8, focus only on "leverage" and avoid taking the "flying knife" trade. Before SK Hynix/Samsung Korean stocks shrink and stabilize, do not add positions on dips
2. Really want to secure a position: Wait for official Q3 DRAM contract price data (mid to late August). If prices rise 13–18% and HBM shipments continue to double, then look back at the XSKHY/XMU pullback support
3. Leverage trades only engage in gamma during earnings/data nights, not lying flat positions that say "I think the cycle isn't over yet."
In short:
Storage isn't finished; it's that over the past three months, the 'AI + shortage + leverage' triple layer premiums have been maxed out all at once. Now, let's talk about the leverage layer first, then the fundamentals. Bulls holding HBM can hold out until 2027, while short-term traders survive this week.
Do you believe the storage supercycle can last until 2027, or do you think prices will peak in Q4 and the Korean stock market will test the July low?📊 $SOL Contract Liquidation Express (August 3)
⚠️ Data repetition reminder: This image is exactly the same as the previous SOL liquidation data, and the following report contains the same content.
According to liquidation data, short-cycle bulls are being pinned down and rubbed wildly, but medium- and long-term bears have directly collapsed...
The liquidation amount in the past hour was about $210,700
Long positions were liquidated at about $210,600
Short liquidation at about $92.29
The liquidation amount in the past 4 hours was approximately $575,900
Long positions were liquidated by about $575,800
Short liquidation at about $92.29
The liquidation amount in the past 12 hours was approximately $2.758 million
Long positions were liquidated at about $2.0902 million
Short positions were liquidated by about $667,800
The liquidation amount in the past 24 hours was approximately $3.5442 million
Long positions were liquidated at about $2.3547 million
Short positions were liquidated by about $1.1894 million
From $SOL liquidation data, 1-hour and 4-hour long liquidations crushed the shorts, with long liquidations being 2,280 times and 6,240 times the shorts, respectively, showing a nuclear explosion-level intensity at the start of the long sell-off; The 12-hour bullish advantage narrowed sharply, dropping to 3.1 times, with short pressure significantly strengthening; the 24-hour bullish advantage further narrowed, dropping to less than 2 times. Dog Maker completed a fierce turnaround from selling long to short on SOL—short-term long positions were targeted and destroyed, medium- to long-term short positions were continuously harvested, and cumulative liquidations exceeded 3.54 million USD. Everyone should control their positions carefully to avoid being bought back.
🔥 Market Barometer | August 3rd
Today's three hot topics point to the same theme: the reset of global asset pricing anchors and the sharp swings in market confidence—the bond market is punishing the Federal Reserve, the currency market is joining forces to resist trends, and the stock market is making an extreme rebound under policy stimulus.
📈 30-Year U.S. Treasuries: Top or New Beginning?
This may not be the end yet.
On July 29, the Fed kept rates unchanged at 3.50%-3.75%, but the vote of 9 in favor and 3 against revealed internal divisions—three regional Fed chairs advocated for rate hikes, marking the first time since 2016. More crucially, Chairman Wash withdrew the forward-looking guidance, completely disrupting market expectations.
Three forces driving long-term bond yields soaring:
· Fed credibility damaged: Senior observers bluntly say "Walsh's message is not clear enough, bond market reacts ruthlessly"
· Side effects of joint US-Japan intervention: May require selling or collateralizing US Treasuries to obtain liquidity
· US-Iran conflict drives inflation expectations: oil prices remain high
The 30-year yield has surged to 5.27%, the highest since 2007. JPMorgan has raised its 30-year target to 5.4%; The options market is betting on a breakout above 5.4% before August 21. Brandywine bluntly stated, "Long-end investors do not believe his anti-inflation narrative." "
💴 US and Japan confirm joint foreign exchange purchase: first joint effort in 15 years
On August 3, the U.S. and Japanese Treasury Ministries simultaneously confirmed that they had jointly bought the yen on July 31. This is the first joint intervention since 2011, and the first yen purchase since the 1998 Asian financial crisis. Japan's Finance Minister clearly stated: "We will continue to intervene together without hesitation going forward." "
After the intervention, the yen rose to the 156-yen range per US dollar. U.S. Treasury Secretary Bescent said it "effectively curbed the disorderly fluctuations of the yen," while Trump said, "This is both a reflection of friendship and beneficial to the world economy." The last time such a partnership was during the Asian financial crisis—the forex market had already entered crisis response mode.
📉 KOSPI surged 14% intraday: an extreme rebound driven by policy
On July 31, KOSPI closed up 17.91%, marking the largest single-day gain in history. The South Korean government announced an injection of 20 trillion won (13.9 billion USD) into the sovereign wealth fund for AI investment; SK Group Chairman made a rare direct purchase of SK Hynix shares; Combined with joint intervention by the US and Japan, the Korean won has strengthened.
However, on August 3, KOSPI opened down 3.6%, with its intraday loss widening to 4.52%. Samsung Electronics and SK Hynix fell 7.8% and 7.5%, respectively. With an 18% surge in one day and a drop of over 4% the next, the volatility of the Korean stock market has shifted from "extreme" to "disorderly."
💎 Summary
Three events outline the core picture of global markets in early August 2026: the bond market is punishing the Fed's hesitation, the currency market is joining forces to fight the trend, and the stock market is making an extreme rebound under policy stimulus before quickly pulling back. The 30-year U.S. Treasury yield rose above 5.27%, the U.S. and Japan jointly intervened in the foreign exchange market, and KOSPI surged 18% in a single day—these are not the norm. When all three markets experience "abnormal" fluctuations simultaneously, the old order is collapsing, the new pricing system has yet to be established, and the chaos in between is the only certainty at present. #30年期美债, the top or a new beginning?
#30年期美债, the top or a new beginning?
#财报观察员: Four draws this week, with Circle as the grand finale 本周美股前瞻（2026.8.3-8.7）
个人观点，不构成投资建议
1.财报、宏观事件前瞻
• 周一：PLTR（盘后）
• 周二：SPCX、ALAB、ANET、AMD（盘后）
• 周三：SNDK、CRCL、GFS、WDC、IONQ（盘后）
• 周四：AAOI、NET（盘后）
• 周五：美国7月非农就业及失业率
2.关于存储板块的核心判断：
• 泡沫出清后市场将趋于理性，将用更严格标准审视各公司业务。
• 数据中心真正缺口、具技术壁垒与溢价能力的是HBM，而非普通DRAM/SSD/NAND/HardDisk等。
• 伴随中国存储公司的产能落地，美股的存储板块将告别单边普涨的历史，转向分化、震荡行情。
• 看好具备HBM研发生产能力的MU、SK Hynix、Samsung三家，但破新高的路程将不会一帆风顺，仍需一些耐心。
3.关于SPCX的判断：
• 解禁压力：财报发布后两个交易日（8月6日后），约有9.11亿股（规模逾千亿美元）迎来首次解禁，将使当前流通盘大幅扩容，形成潜在抛压。
• 潜在传闻：特斯拉正考虑出售中国业务，为TSLA和SPCX合并铺路，但合并面临监管问题，尚需时间，不应作为短期买卖的逻辑理由。
• 当前环境：在整体大盘偏谨慎的环境下，考虑SPCX的高估值，以观察为主。With 6 days left, $BABY will unlock 136 million selling pressure. For those who jumped on the wave in the past couple of days, get off the board quickly:
1. Next Monday, Babylon will unlock 136 million tokens, of which 85 million belong to early investors, 41 million to the team, and 10 million to advisors. At the current price of 0.011U, that's $1.5 million.
2. Currently, BABY's daily trading volume is $4.9 million, meaning next week's unlock volume will account for 30% of the day's trading volume. In theory, the market can still absorb the price within a 5% drop.
But the key point is, this unlock is a fixed monthly unlock, which is 36 consecutive months, with 30% of daily trading volume entering the market each month. That amounts to 1.63 billion coins annually, accounting for 40% of the current circulating supply.
3. Coincidentally, Babylon is engaged in syndicated staking and is collaborating with Aave V4 and others, which not only increases returns but also supplements liquidity. Will the unlocked tokens be re-staking? I don't think so.
Although the project team requires 1 $BTC for 20,000 $BABY, if the co-staking yield is high enough, some unlocked tokens will be locked by staking. But the problem is, everyone also calculates the score. Although co-staking yields about 9%, the token has an annual inflation rate of 5.5%, not to mention that the more tokens unlocked, the lower the price drops.
4. Under what circumstances will the decline stop? I feel the chances are too slim. Because it requires:
(1) The TBV mainnet is beginning to generate substantial real income;
(2) The official system must truly initiate the buyback and destruction mechanism;
(3) BTC's sharp rise boosts co-staking yields, locking in more tokens.
So, without a catalyst, every rebound is a window for selling. If you want to buy a BABY, at least wait until the buyback mechanism is implemented.