Orbit Post Sitemap
100 crypto projects will die by 2026, and 4 exchanges will shut down within a month
On July 17, BitMart released an impressive half-year report: assets under management grew by about 256%, a new predictive market product was launched, and it just obtained its Australian financial services license in June. The report also acknowledged that the backdrop was not good: Bitcoin fell 30% in half a year, Ethereum was halved, and spot ETFs saw record net outflows.
Nine days later, at 01:30 UTC on July 26, the same company announced an orderly shutdown. New user registrations stopped, deposits closed, futures accounts switched to reduced position mode, trading was fully halted on August 26, and completely closed on January 31, 2027. The platform token BMX fell nearly 60% that day.
Even more absurd is the statement from former global CEO Nenter Chow on X: he was notified of his dismissal on July 24 and has not been involved in any management or decision-making since then. The news of the shutdown, like everyone else, was seen in the announcement.
Three days ago, BitMEX had just announced the exchange shutdown at 04:00 UTC on September 23, ending an 11-year hiatus.
Looking further back, AscendEX was shut down on July 1, and EXMO was put into liquidation after being placed on the UK's sanctions list against Russia.
Within a month, four well-known centralized exchanges exited.
RootData's 2026 crypto industry dead projects list has reached the 100th and is still being updated.
The common feature of that batch of deaths in 2022 was violence:
Luna lost its value in three days, 3AC margin recovery defaulted, FTX misappropriated customer assets and was run on the bank, and Celsius froze withdrawals. The death happened instantly, the user's assets evaporated instantly, and the judicial process dragged on until today.
The common trait of this batch in 2026 is decency.
The wording of the announcement is almost identical: after a careful assessment of operating conditions, market environment, and future strategic direction, it has been decided to exit in an orderly manner.
In plain terms, this means business is no longer profitable. No hacking, no bank run, no law enforcement raids—just the accounts can't keep up.
Starving and exploding are two completely different market signals. An explosion means systemic risk is spreading; if one family falls, the whole world will be taken down; Starving to death means the individual business fails, and risk is isolated on their own balance sheet.
#参议院CLARITY法案下周或表决: Favorable Moments or Shortcoming? $PIEVERSE $ETH ETH's relative move today warrants a closer look. At roughly three times BTC's 24-hour gain, with the Iran strike pause pulling risk appetite back into markets, the outperformance looks positioning-driven rather than narrative-driven. Rotation into ETH ahead of broader alt momentum is a known pattern; whether this is that setup or just a one-session catch-up is still unclear.
The macro backdrop adds friction. Jobless claims dropping gives the Fed less reason to move quickly on cuts, keeping real rates elevated and limiting the liquidity tailwind crypto needs to sustain a rally. Google and Tesla earnings this week matter more than most traders expect; a growth miss there could reprice the whole risk-on move. I'd want more confirmation before treating this bounce as structural.
Just my read, not advice.Anthropic feeds the Korean giants, A-shares' 3 trillion yuan boosts China's Changxin—Is the AI storage cake big enough for three to share?
Today, the A-shares market went crazy.
Changxin Technology debuted on the STAR Market, opening up 471.59% to 49.5 yuan/share, with a total market value surpassing 3.31 trillion yuan.
Surpassing Industrial and Commercial Bank of China, it topped the A-shares market cap rankings.
The first hour of trading saw turnover exceed 100 billion yuan, becoming the first A-share stock to break 100 billion yuan in single-day turnover. One lot earned 20,000 yuan, with 9.42 million accounts rushing to subscribe.
A company founded only in 2016 has become the largest Chinese enterprise by market value in just ten years.
What does this mean? Changxin's Q1 revenue was 50.8 billion yuan, up 719% year-on-year; net profit attributable to the parent company was 24.762 billion yuan, up 1688%. The half-year earnings wiped out all losses accumulated since its founding.
AI storage is truly highly profitable.
But the other side of the story was written seven days ago.
At the San Francisco AI Summit, Samsung, SK Hynix, and American tech giants signed cooperation agreements worth $950 billion.
Anthropic directly signed supply agreements with Samsung and SK Hynix. Nvidia and SK Group signed cooperation exceeding $500 billion, securing long-term priority supply rights for HBM. Samsung supplies Broadcom with $200 billion worth of chips.
The Korean giants have fully consumed the fattest orders of the AI era.
SK Hynix just went public on Nasdaq on July 10, raising $26.5 billion, setting a record for foreign companies listing in the US. Changxin listed on the STAR Market on July 27, raising 57.9 billion yuan, the largest IPO in STAR Market history.
Two IPOs less than two weeks apart.
Capital is telling the world with real money: the storage track has officially entered a three-way battle.
Data doesn't lie.
In Q1 2026, global DRAM market share: Samsung about 39%, SK Hynix 29%, Micron 22%, Changxin 8%.
Changxin increased from 4.7% a year ago to 8%. Northeast Securities predicts its long-term share could rise to 30%.
There are only four global DRAM manufacturers with full IDM capabilities: Samsung, SK Hynix, Micron, and Changxin.
It used to be a two-horse race; now it's a three-way contest.
But the question arises—Is the AI storage cake big enough for three to share?
Anthropic's orders went to the Korean giants, A-share funds went to Changxin. Both sides are expanding production and burning cash.
Samsung and SK Hynix signed $950 billion long-term contracts. Changxin's DRAM capacity is expected to approach Micron's by the end of 2026.
The cake is growing, but the number of knives cutting it is also increasing.
JPMorgan expects global semiconductor revenue to grow over 90% year-on-year in 2026, reaching $1.5 to $1.6 trillion. Industrial Securities estimates a global DRAM supply-demand gap of about 7.22% in 2026, with tightness continuing into 2027.
The gap remains, but whoever captures the largest share before the gap closes will be the king of the next decade.
Finally, something useful for the crypto community.
AI computing power-related crypto assets have a narrative based on computing power scarcity. The core bottleneck of computing power is storage—HBM, DRAM, these determine how fast AI chips can run.
Previously, it was a "duopoly narrative"—Samsung and SK Hynix monopolized high-end storage, setting computing power costs.
Now it has become a three-party structure.
Once Changxin's capacity is massively released, how will DRAM prices move? Morgan Stanley predicts contract prices will peak in Q4 2026. Once prices loosen, computing power costs will fall—is this good or bad for AI computing power tokens?
This is a question worth pondering tonight.
Anthropic fed Korea, A-shares lifted China.
But the real winners are never the storytellers—
they are those who see the flow of funds clearly before the landscape reshapes.
$SAMSUNG $SKHY $MU
#长鑫科技上市，全球存储竞争添变量 $SNDK storage sector is impacted by rising inflation expectations, with short-term valuations facing deleveraging and macro rebalancing pressure. Rising international oil prices boost inflation hedging sentiment, and the probability of a Federal Reserve rate hike jumps, suppressing overall risk appetite for tech stocks, causing concentrated withdrawal of long positions. If the July Federal Reserve decision shows a more hawkish stance than expected, valuation contraction pressure will further transmit to the NAND long-term pricing market. If a hawkish policy is implemented and global cloud providers' capital expenditures exceed expectations strongly in the second half of the year, the current trading desk's price-smashing logic will fail.
#新手必看：这里有你需要的一切 #贝莱德等九机构组建安全联盟 #英伟达拟为OpenAI提供2500亿美元担保The true long and short logic of Changxin Technology
Bullish logic:
① A scarce large-scale DRAM manufacturer in mainland China
② Approximately 8% global market share, with room for growth
③ Domestic substitution and AI computing power demand provide long-term support
④ DDR5, LPDDR5X, and future HBM bring product upgrade potential
Risk logic:
① DRAM prices are clearly cyclical
② Top five customers account for about 68% of sales, indicating high customer concentration
③ Gross margin and process technology still lag behind the top three international manufacturers
④ High capital expenditure, depreciation, and equipment export restrictions may affect capacity expansion
⑤ HBM still needs to pass technical, yield, and customer certification verification
My judgment is: Changxin Technology's industrial value has long-term scarcity, but the high valuation on the first day of listing has already priced in some medium- to long-term expectations. In the short term, it looks more like a game of funds, chips, and sentiment; in the medium to long term, we need to wait for verification of profit quality and HBM progress.
Additionally, on-chain CXMT contracts are not equivalent to holding 688825 stock. The two differ in trading hours, liquidity, price sources, delivery mechanisms, and investor rights, so risk-free arbitrage cannot be directly performed.
The above is only market research and does not constitute investment advice.
#长鑫科技上市，全球存储竞争添变量 ETH's +4% against BTC's +1.5% today is not a random divergence. It reads like a short-squeeze on crowded ETH underperformance positioning, amplified by FOMC proximity, where risk gets repriced in both directions fast.
Korea capital shift and FOMCRateWatch trending together suggest institutional money is repositioning before the meeting, not after. That kind of pre-FOMC bid reverses hard if the statement surprises hawkish. Worth monitoring, not chasing.
NFA, just my read.
#OKXOrbitThe market just told us: beats alone don’t cut it anymore.
Alphabet dropped $119.8B Q2 revenue and Cloud kept growing. $GOOGL still tanked 4% after hours.
Why? Eyes moved to the future. Capex guidance hiked to $195B–$205B for 2026, up from $180B–$190B. Free cash flow went negative. AI is huge, but Wall Street is asking: who’s paying for it?
Google + Microsoft + Meta + Amazon are set to spend $725B combined in 2026. That’s +77% YoY.
Tesla was quiet. Still holding 11,509 BTC since 2022. Took a $112M BTC-related loss, but didn’t sell. No panic. No buys. Just HODL.
What this means for crypto:
1. ETF inflows keep supporting $BTC
2. Crypto still tracks Nasdaq 100. Big Tech earnings = crypto sentiment now
3. Microsoft, Meta, Amazon up next. Their guidance will move both stocks and crypto
Trader edge: stocks sleep, crypto doesn’t. With OKX tokenized US stocks trading 24/7 in $USDT, $XGOOGL and $XTSLA stay live through earnings and weekends.
Will the next Big Tech reports fuel crypto or drag it down?
#DailyOrbit @OKX Orbit
#CXMTMemoryIPO
#FOMCRateWatch $BTC 周一早评：反弹只是情绪修复，并非趋势反转
大饼周一反弹至65400，核心推手是特朗普暂停对伊朗的军事打击，中东地缘风险降温，此前避险溢价回落，行情提前止跌回弹，属于超预期的情绪修复。
但这次反弹并不是买盘回归：现货ETF连续两日大额净流出超4.65亿美元，机构在撤退；交易所稳定币流入跌至数月低点，场内增量资金不足；反弹成交量萎靡，缩量上涨根基不稳。
上方关键阻力在65500-65800区间，没法放量站稳就只是短期震荡。本周美联储议息会议是核心变量，市场加息概率约35.8%，若释放鹰派信号，BTC大概率回踩64000甚至62500支撑。
原本预判的62500见底逻辑没变，只是地缘消息提前了反弹时间。情绪修复不等于趋势反转，本周消息面繁杂，短线多空博弈难度大，建议先观望，短期波动不改熊市末期的中期大方向。财报夜拆雷：谷歌在数钱，特斯拉在数“还剩几个月”
昨晚美股收盘后，我盯着谷歌和特斯拉的盘后走势，差点把咖啡喷在屏幕上。
一个涨4%，一个跌3%。冰火两重天。
但说真的，这俩公司的财报，就像两张完全不一样的高考答题卡——一张在认真算答案，另一张在空白处画变形金刚。
---
谷歌：无聊，但有钱
先聊谷歌。847亿营收，超预期。广告回血，Youtube增速干到21%。
没什么惊喜，但稳得像中年人的保温杯。
电话会上那帮分析师不死心，追着问云业务AI到底赚了多少钱。谷歌CFO的回答翻译成人话就是：“我们确实在猛砸钱买芯片，利润暂时被吃掉了，但你们别急。”
市场居然买单了。盘后涨4%。
为什么？因为大家突然反应过来：这哥们儿是全班那个不偏科的优等生——搜索是保送名额，Youtube是特长加分，云业务就算暂时瘸腿，也架不住家里有矿。
唯一让我心里犯嘀咕的是，资本开支还要继续扩大。这意味着谷歌的AI故事，短期内依然是“烧钱模式”，不是“印钱模式”。22倍的PE，不算贵，但也别指望它像meme股那样天天给你涨停。
---
特斯拉：除了储能，全是问题
再看特斯拉，我都不想说了。
汽车毛利率跌破14%。朋友们，这可是特斯拉啊，当年那个毛利率吊打BBA的怪物，现在跟普通合资厂坐一桌吃饭了。
交付量44.4万辆，环比几乎原地踏步。降价都降了个寂寞。
最要命的是FSD。全村的希望，AI落地的排头兵，订阅率卡在18%死活上不去。马斯克在推特上吹了那么多天的V12，结果用户根本不买账。
华尔街那帮人最狠，他们不看故事，看数据。数据出来，信仰崩塌，盘后直接再砸3%。
68倍的市盈率，卖的是“未来机器人霸主”的梦想。但你现在的表现，分明就是“卖着四年前老款车的车厂”，外加一个卖得不错的充电宝业务（储能确实亮眼，翻了一倍多，但体量还太小）。
我甚至有点毒舌地想：如果去掉“AI”两个字，特斯拉现在的估值，能不能有现在的一半？
---
英伟达：被猪队友带沟里的真学霸
英伟达这周也跟着跌了8%，纯粹是躺枪。
H100还在缺货，H200排到明年，台积电的产线都快踩冒烟了。基本面一点问题没有。
但市场不管，先跌为敬。理由是：如果谷歌和微软自己都快养不起云业务的利润率了，那他们还会疯狂买你的芯片吗？
这是个好问题，但至少目前，没有数据支持这个逻辑。英伟达现在的处境，有点像学霸被两个学渣拖累了班级平均分，老师还得找他谈话。冤不冤？冤。但资金要避险，先跑为敬，这是人性。
---
说点真心话
这轮下跌，不是AI死了，而是市场从“喝大酒吹牛B”切换到“拿计算器算账”了。
退潮之后，谁在裸泳？
· 谷歌是那个穿着泳裤、体型微胖的中年人，不好看，但淹不死。
· 英伟达是那个正在冲浪的肌肉男，浪还在，只是风变小了。
· 特斯拉？他可能穿着一条画满火箭的泳裤，但裤腰带松了。
如果非要我下注，目前这个价位，我更愿意捏着鼻子买谷歌。虽然无聊，但踏实。特斯拉需要证明的不是产能有多强，而是那套FSD能不能真正变成收费刚需，而不是一个昂贵的玩具。
AI下半场，市场只认一种人：能把算力变成真金白银的人。
讲故事的时代，过去了。
--- The U.S. annual interest on national debt is as high as $1.2 trillion, and this year's fiscal deficit is expected to exceed $2 trillion. Don't naively think the U.S. can't hold on; behind this lies profit-making and vested interests.
The interest is paid by American taxpayers, while the interest is paid by Treasury holders: overseas holdings hold 9 trillion US Treasuries, Japan holds the largest shareholding, China continues to reduce holdings, and most of the rest is held by domestic banks, funds, and pension funds.
Normally, buying government bonds means bearing interest rate volatility risk, but now that long-term US Treasury yields have broken through 5%, stablecoin issuers have played a trick. Take TEDA as an example: at the end of last year, 83% of its reserves were U.S. Treasuries, totaling over $122.3 billion. Users who buy 1 USDT only get the face value, and the 4%-5% interest generated by government bonds is all pocketed by the issuer. Last year, the Genius Act directly stipulated that stablecoin issuers could not pay interest to holders, confirming this profit-taking.
Exchanges tried to circumvent the rules through event rewards, but Bank of America immediately intervened, claiming that stablecoin interest payments would lead to an outflow of 1.3 trillion in bank deposits, repeatedly applying pressure under the Crypto Clarity Act. What banks truly fear is not the crypto industry, but the withholding of interest that should have been distributed to ordinary investors.
Holding stablecoins essentially means indirectly holding US Treasuries, but not getting corresponding returns. Many people hope the bill will blindly support the implementation of a bull market. Although I also hope the bill passes, the distribution of benefits is really unfair—the higher the yield on U.S. Treasuries, the more profits ordinary investors take. #参议院CLARITY法案下周或表决: Favorable Moments or Shortcoming? 好家伙，CoinGecko 的二季度报告显示：Crypto 总市值下降 12.6%，中心化交易所现货成交量更是下降 27.9%。 价格跌，交易也冷，说明市场缺的不是新故事，而是愿意持续进场的资金。 我的判断是，如果现货成交量后续仍没有明显恢复，市场较大概率还是以局部热点和短线反弹为主，很难迅速回到普涨节奏；反过来，成交量与稳定币资金同步回升，才可能成为风险偏好真正修复的信号。👀 仅作市场观察，不构成投资建议。如果RSI、MACD和成交量三项技术指标同时走弱，那么山寨币市场可能正在进入一个系统性风险释放阶段。
问题在于，这种技术面恶化是否已充分反映在价格中，还是仍有进一步下行的空间？
原文基于对超过100个山寨币的技术扫描，给出了五条关键信号，并指向一个偏空结论。这些信号包括：
- 超过65%的山寨币出现RSI顶背离，即价格创新高但RSI走低
- 超过70%的山寨币MACD柱状图正在收窄，显示上涨动能衰竭
- 超过75%的山寨币成交量低于20日均线，表明买盘缺乏燃料
- BTC市场占有率从54%升至56.8%，资金正在从山寨币回流比特币
- 仅8个山寨币呈现正向成交量背离，占比不到8%
上述信号共同指向一个结构性变化：资金正在从广泛的山寨币市场撤出，集中到比特币和极少数强势项目。这种仓位行为的改变，直接传导至ETH和SOL等主流山寨币，使其承压。风险偏好显著下降，投资者倾向于持有BTC而非追逐高贝塔资产。
偏多路径的条件：如果BTC市占率停止上升并回落至54%以下，同时成交量重新站上20日均线，那么上述背离信号可能被证伪，山寨币存在修复机会。但当前数据显示这一条件尚未满足。
偏空风险的条件：如果BTC市占率继续向58%甚至60%迈进，且成交量持续低迷，那么山寨币可能面临更深度的回调。原文提到的92个山寨币RSI已跌至35-48区间，成交量萎缩70%，CMF为负，说明资金流出尚未结束。
核心结论是，市场正在重定价山寨币的风险溢价。技术指标的集体走弱，叠加资金向BTC集中，意味着短期山寨币的持有成本上升、流动性下降。对于以BTC为锚的持仓者，这可能是一个相对安全的避风港；但对于重仓山寨币的策略，需要警惕进一步的下行风险。
主要风险在于，技术指标可能滞后于价格，且背离信号在极端行情中可能失效。建议交叉验证链上数据，例如交易所净流量和稳定币供应比率，以判断资金是否真正离场。
如果成交量无法恢复，山寨币的弱势可能延续至下一轮催化剂出现。
$BTC $ETH $SOLPCE 反彈後美股怎麼走
通脹粘性服務業功不可沒
通脹數據是美聯儲決策的錨。
核心 PCE 2.7%。服務通脹 4.1% 仍是主要推動力。
商品通脹 -0.3%。能源和耐用品價格回落，給通脹降溫做出貢獻。
工資增速 4.5%。高於通脹 2.7%，實際工資轉正。
組合配置永遠比單個標的判斷重要。
組合配置永遠比單個標的判斷重要。
📌 為什麼要把 PCE 放進資產框架
PCE 不是一個直接的買賣按鈕，它更像流動性和利率預期的背景變量。核心服務通脹如果持續有黏性，降息節奏可能放慢；商品價格回落則可能給政策留下空間。兩者方向相反時，市場往往先交易預期，再等待後續數據確認。
🧭 我會怎樣跟蹤
第一，看核心 PCE 的三個月和六個月趨勢，不只看單月變化。第二，看工資、住房和能源是否出現同向拐點。第三，看美債收益率、美元和風險資產是否對數據作出一致反應。數據和價格不一致時，我會先降低確定性。
⚠️ 風險提醒
市場預期會在正式數據前反覆變化，任何降息概率都不是承諾。宏觀數據也可能被修正，不能把一個指標包裝成確定答案。
🎯 最後的執行框架
把宏觀判斷用來調整風險預算，而不是用來預測每一個短線高低點；保留流動性，等政策與市場價格真正共振。
我會把這個話題拆成三層來看。第一層是可以直接觀察的數據，先記錄數值、時間和方向，避免只截一張圖就下結論；第二層是市場如何反應，數據改善但價格不動，和數據轉弱而價格仍然上漲，含義完全不同；第三層才是自己的操作，先寫下最大可承受損失，再決定是否需要調整倉位。這個順序看起來慢，但能減少被單一標題帶著走。
對我來說，通脹分項、利率預期和美元流動性要放在同一張表裡對照。每次更新只改變有新證據的部分，不能因為一個數字變化就把整個判斷翻轉。若三個觀察方向彼此矛盾，我會把結論降級為「等待確認」，而不是硬湊出一個看多或看空的故事。市場中最容易被忽略的成本，是過早確定之後不願意承認假設已經失效。
執行上我會先用觀察倉測試，等成交量、價格和基本面至少有兩項同向，再考慮增加曝險；若波動擴大或流動性變薄，則先縮小倉位。任何回測、歷史案例或 KOL 觀點都只能用來建立假設，不能代替當下的風險檢查。這篇內容是我的研究筆記，不是保證收益的買賣指令。
我會在下一次更新時重新檢查四件事：消息是不是仍然有效、價格反應有沒有確認、流動性是否足以執行，以及原本的風險假設有沒有被破壞。若只是社交媒體熱度上升，卻看不到成交量或資金的配合，我會把它當作待觀察訊號；若數據方向改變，也會同步修改原先的劇本，而不是為了維持面子繼續持有。
這種做法的好處是把「看法」和「行動」分開。看法可以保留多個可能性，行動則必須有清楚的觸發條件。對短線交易，我會設定時間上限；對中長線配置，我會檢查基本面和資金成本。無論最後結果如何，都把進場理由、退出理由和實際滑點記錄下來，下一次才有真正可以改進的復盤材料。
如果資料來源之間互相矛盾，我會先標記衝突，等原始公告或下一個時間點確認，不用社交媒體的情緒替代證據。這也意味著有些時候最好的操作是空倉等待，因為沒有交易本身也是對不確定性的管理。3.3 trillion! China's storage giant outperforms ICBC in one day, but the South Korean stock market crashes
On July 27, no new stock in A-shares history has ever been as crazy as today.
Changxin Technology, with an issue price of 8.66 yuan, opened at 49.5 yuan, soaring 471%. The opening market value reached 3.31 trillion yuan, directly surpassing ICBC, topping A-shares. Intraday it surged over 530%, with market value hitting 3.61 trillion yuan—trampling Intel underfoot.
The turnover in the first hour of listing broke 100 billion yuan, the first A-share stock in history to exceed 100 billion yuan in single-day turnover. The turnover rate was 53%, with half of the circulating shares changing hands in the first hour. One winning lot earned 20,000 yuan.
In ten years, from the “506” project in the suburbs of Hefei to the world's fourth largest DRAM manufacturer. Domestic storage crowned today.
But on the other side of the story, at the same moment in South Korea—
The KOSPI index opened up 1.7%, then plunged straight down. Samsung Electronics opened 3% higher but turned negative. SK Hynix opened 3.13% higher but directly reversed.
The $950 billion semiconductor cooperation agreement—Samsung + Broadcom 200 billion, SK + NVIDIA 750 billion—could not stop foreign and institutional investors from net selling 40 billion Korean won.
Good news exhausted. Again, good news exhausted.
A Chinese company goes public, and the South Korean stock market crashes first. Can you believe this scene?
A week ago, Anthropic just signed supply agreements with Samsung and SK Hynix. NVIDIA acquired 4.5% of Naver for $1 billion. Everyone said the Korean giants were secure, and the AI storage cake was theirs.
Then Changxin arrived.
In the global DRAM market, Samsung holds 38%, SK Hynix 29%, Micron 22%, three companies monopolizing 90%. Changxin? One year ago 3%, in Q1 this year already reached 8%. The world's fourth. Net profit is expected to be 50 to 57 billion yuan in the first half of 2026, with a year-on-year surge up to 2544%.
This is not "joining the game." This is flipping the table.
For the crypto world, this matter is much bigger than you think.
SK Hynix's tokenized stock $SKHY is already trading on Solana. Micron's tokenized version is also on Ethereum. If you have allocated storage chip assets through RWA—
your investment narrative must be rewritten from today.
"Duel of the two giants" turns into "three-way melee." The DRAM capacity pattern changes from a three-company monopoly to a four-player contest. Some predict Changxin's capacity will approach Micron's by the end of 2026.
What does this mean?
It means the pricing power of Samsung, SK Hynix, and Micron will be diluted. It means the profit margin in the high-end HBM market may be squeezed. It means the valuation logic of your storage-related crypto assets—whether tokenized stocks or projects in the AI computing power track—needs to be recalculated.
The world's most attractive storage target is listed at your doorstep, and you can only watch.
Storage chips are the hardest assets in this AI era. HBM price increases, DRAM shortages, AI server memory capacity is 10 times that of traditional servers. Samsung, SK Hynix, and Micron allocate 80% of advanced capacity to AI storage. The entire track is in short supply.
Now, Chinese players officially enter the pricing system.
$SAMSUNG $NVDA $SKHY
#长鑫科技上市，全球存储竞争添变量 美国自上周五起暂停对伊朗空袭，市场7月27日开盘直接定价：
布伦特原油 -6%，至$91/桶附近
WTI跌至$84以下
纳指期货高开1.4%
BTC重回$65,000上方
据CBS，暂停轰炸与阿曼官员上周五赴德黑兰会谈直接相关。伊朗陆军同步表示已暂停回应行动。预测市场对"8月31日前达成停火协议"的定价升至75%。
剧本很熟悉：地缘缓和→油价跌→风险资产涨。每次都这样，这次也没例外。
我个人觉得，75%的停火定价是目前最值得盯的数字——不是因为它一定会发生，而是因为一旦跌破，反向交易会很猛。上次停火协议不足24小时就破裂，市场有记忆，但每次还是先信再说。
霍尔木兹这条路，开了又关，关了又开。这次能撑多久？
#美军暂停对伊空袭，国际油价开盘大幅下跌 The market is now betting on which coin will get the ETF 👀 entry ticket first
Have you ever thought that when ETF expectations are hyped up in advance, the real game doesn't happen on the day of approval, but before the news is realized?
Recently, the US has quietly advanced the approval process for several crypto ETFs, from XRP to DOGE to SOL, with nine officially launched. There were also 13 players in line, including old friends like ADA, LINK, and XLM.
But what I find interesting is that many people focus only on "who will be criticized" and overlook another aspect—the cross-market linkage logic behind these coins.
For example, LINK and HBAR seem to represent a "strong ecosystem + strong institutional attention," but their trends are quietly following the pace of US tech stocks. When Nasdaq pulls back, no matter how strong the ETF expectations for these coins are, liquidity can easily drag them down. Meme stocks like DOGE are more betting on sentiment premium, with much less relevance to the US stock market.
In other words, the current narrative of betting on a particular coin's ETF is not just about its fundamentals, but also about whether its market style is favored by macro capital.
From a bullish perspective:
- If macro liquidity improves (such as rising expectations of rate cuts), coins that have already submitted applications and have a solid ecosystem may be the first to be bought by concentrated funds, because the ETF narrative itself is a "tradable story."
- For companies like ADA and LINK, which have already submitted applications, once there is clear progress, price elasticity will be significant.
But risks also lurk in the shadows:
- Currently, market expectations for ETFs have been partially priced in advance. If the approval pace is slower than expected or delayed by the SEC, these coins may experience a "good news exhausted" pullback.
- Cross-market linkage means that if US tech stocks experience a systemic decline, coins with strong ties to US stocks may be dragged down, and no matter how strong the ETF narrative is, it can't withstand liquidity contraction.
So the current stage is more like a "game divergence period": it's not simply chasing gains or washing stocks, but the market waiting for a clear signal to confirm whether the narrative holds.
My judgment is: in the short term, focus on watching more than making moves, and focus on the resilience of ETF-related coins during US pullbacks. If they can hold firm, that's when they truly deserve attention.
(The above is just my personal thoughts while watching the market and does not constitute any trading advice.)
$ADA $LINK $HBAR #ETF叙事 #跨市场观察At least 29 overseas cryptocurrency exchanges in South Korea have reportedly become unavailable for downloads on their local Google Play stores.
South Korea has always been one of the most active cryptocurrency trading markets globally, with impressive investor participation and trading volume, often becoming a major force influencing market sentiment. Once an exchange app is restricted from downloading, it affects not only new user registrations but may also involve subsequent updates, account login, asset operations, and service support for existing users. If restrictions continue to expand, it could impact user growth, capital flow, and market confidence on exchanges.
@OKX Chinese: Is this news true?On the eve of the FOMC, BTC is completing a healthy leverage reset
BTC quickly rebounded after hitting $63,666, rising back to around $64,500 at the time of writing. In the past 24 hours, the entire network liquidated $323 million, with long liquidations accounting for 84%. This is a typical leverage clearing rather than a panic sell-off.
1. The core contradiction of this decline lies in macro factors
The 10-year US Treasury yield broke through 4.71%, hitting a new high for the year; oil prices rose above $100/barrel due to Middle East tensions. CME data shows the probability of a rate hike in September has surged to 82%.
A bigger variable is this week—at 2:00 AM Beijing time on July 30, the Federal Reserve will announce its interest rate decision. The market expects rates to likely remain unchanged (3.50%-3.75%), but the key lies in the post-meeting statement and Chair Powell's press conference.
Bloomberg predicts that Dallas Fed President Logan and Cleveland Fed President Mester may vote against, favoring an immediate rate hike. If there are two dissenting votes, it would signal a rate hike in September.
On-chain data shows different signals. Gate.com analysts point out that the number of long-term Bitcoin holders currently at a loss has exceeded the level during the FTX crash, approaching the 2018 bear market level, with Bitcoin priced around $50,000. This implies two directions: long-term investors’ unrealized losses deepen, but true bottoms often form after chips have fully changed hands.
2. Technicals: Key ranges are being repeatedly tested
BTC is currently oscillating narrowly between $64,300 and $64,900, with short-term momentum weak but structure intact.
Key levels:
Resistance above: $64,800-$64,900 (1-hour long-short dividing line), $65,200-$65,400 (4-hour moving average resonance pressure), $66,600-$66,900 (daily strong resistance)
Support below: $63,700-$63,800 (short-term buying concentration), $63,100-$63,300 (50-day moving average and mid-term lifeline), $61,200-$62,500 (extreme trend support)
Core logic in a range-bound market: defend support and closely watch resistance. Before the FOMC, a narrow range-bound continuation with no trending market is highly probable.
3. Trading strategy
Remain cautious before the meeting.
Long side: If BTC retraces to $63,700-$63,800 with signs of volume contraction and stabilization, consider light long positions with a stop loss at $62,800, target $65,200-$65,500, risk-reward ratio about 2:1.
Short side: Avoid for now. Current price still has room before resistance, and ETF net inflows have continued for three consecutive weeks without breaking trend, making counter-trend shorts unfavorable in risk-reward.
4. Key variables
Three FOMC scenarios: Hawkish (high probability) → BTC falls back to test $63,000; Neutral → maintains $63,000-$65,000 range; Dovish (low probability) → pushes $64,850-$65,300.
Final notes
The core of this decline is leverage clearing, not fundamental deterioration. The quick recovery near $63,600 shows buying remains. The real test is the FOMC meeting; before that, the $63,700-$64,900 range will likely continue to consolidate.
The direction will emerge. No rush. Three Benchmarks to Test the Quality of AI: Understanding the Trends of Microsoft, Meta, and Amazon #财报观察员：微软Meta亚马逊能稳住AI叙事吗？
1. The True Quality of AI Narratives from the Three Companies
1. Microsoft: The Only One with "Cash Flow + AI Revenue Dual Verification," Strongest Narrative Resilience
Core Financial Data
1. AI annualized revenue exceeded $37 billion, a year-over-year surge of 123%, the only company among the three to separately disclose AI-specific revenue and maintain triple-digit growth; Azure cloud backlog orders reached $462 billion, nearly half from long-term enterprise AI contracts, demonstrating strong customer stickiness.
2. Annual capital expenditure around $105 billion; despite continuous investment in OpenAI compute clusters, free cash flow over 12 months was $66.9 billion, a significant year-over-year increase, with AI investments fully covered by internal cash flow, no need to overdraw the company’s foundation.
3. Product closed-loop barriers: Office Copilot enterprise paid penetration rate exceeded 41%, B2B customers willing to pay an additional 30% for AI features, a rare industry case of direct AI price monetization.
Narrative Advantages and Risks
✅ Supporting logic: AI is not just a cloud ancillary service but an independent second growth curve; the software-native light-asset nature offsets compute investment costs, with a very thick cash flow safety cushion, and the market is willing to give a long-term valuation premium.
⚠️ Potential cracks: 45% of AI cloud contracts are tied to OpenAI, indicating high customer concentration; Azure growth slightly slowed to 39%, and pure compute rental gross margin continues to be compressed by GPU procurement costs.
Conclusion: The most stable AI narrative among the three; as long as AI annualized revenue maintains growth at the hundred-billion level, capital expenditure pressure will not shake the fundamentals.
2. Meta: No ToB AI Infrastructure, Relies on Advertising Efficiency for Revenue Growth, Narrative Follows a Differentiated Segmentation Path
Core Financial Data
1. Annual capital expenditure raised to $115-$135 billion, all invested in self-developed large models, AI recommendation algorithms, and VR compute clusters; total advertising revenue up 33% year-over-year, with AI-optimized feed contributing 70% of incremental revenue.
2. No external AI compute rental business, does not make money by selling GPUs; AI is fully internalized to serve the advertising main business: AI recommendations reduce customer acquisition cost by 22%, ad conversion rate up 18%, directly boosting ad profits.
3. Cash flow buffer: Advertising business continuously and stably generates cash, with quarterly operating cash flow exceeding $32 billion, sufficient to cover AI R&D and compute investments, no risk of negative cash flow turnover.
Narrative Advantages and Risks
✅ Supporting logic: Takes a completely different AI path from Microsoft and Amazon, avoids heavy-asset cloud infrastructure, turns AI into a tool to reduce advertising costs and improve efficiency, commercialization path with zero external dependency, shortest profit realization cycle.
⚠️ Potential cracks: Consumer large model Threads and Meta AI user growth slowing, consumer AI has not yet formed independent revenue; if the advertising industry cycle declines, AI optimization-driven increments will shrink accordingly, lacking a second growth curve.
Conclusion: Short-term narrative is solid, but long-term ceiling is tied to the global advertising market, making it difficult to develop an independent AI valuation trend.
3. Amazon: Most Aggressive AI Infrastructure Investment, Heaviest Cash Flow Pressure, Highest Narrative Risk
Core Financial Data
1. 2026 annual capital expenditure approaching $200 billion, highest among the three, mostly for AWS AI data centers and Trainium self-developed AI chip production lines; free cash flow in the past 12 months plunged from $25.9 billion to $1.2 billion, market institutions warn of possible negative cash flow turnover in 2027.
2. AWS cloud revenue grew only 28% year-over-year, below the market expectation of 30%; although AI backlog orders are $364 billion and self-developed chips locked in $225 billion long-term contracts, AI compute rental gross margin continues to decline, many orders are low-price lock-in, difficult to convert to profit in the short term.
3. Business fragmentation: Retail segment has almost no AI monetization, AI growth fully relies on cloud business; if enterprise customers reduce IT budgets, AI revenue will be directly pressured.
Narrative Advantages and Risks
✅ Supporting logic: The world’s largest AI compute supplier, Trainium self-developed chips free from Nvidia dependence, with significant long-term compute cost advantages; institutional funds are betting on profit realization after 2027.
⚠️ Core cracks: Currently in a "high investment, low return" cycle, cash flow continuously consumed by compute infrastructure; the market has lost patience, stock price has repeatedly dropped after earnings due to downward capital expenditure guidance, making it the most fragile AI narrative among the three.
Conclusion: Short-term narrative faces loosening risk; must wait for AI chip gross margin recovery and free cash flow stabilization to reestablish AI growth logic.
2. Three Golden Benchmarks to Judge Whether AI Narratives Can Sustain
Most market reviews only compare AI revenue growth rates, ignoring the underlying constraints of capital expenditure and cash flow. To truly test the quality of AI narratives, look at three points:
1. Benchmark One: Whether AI revenue is independently incremental rather than cloud business bundled and inflated
Microsoft separately breaks down AI annualized revenue; Meta relies on advertising efficiency to clearly show AI increments, making their narratives more credible; Amazon classifies all compute rentals as AI, unable to distinguish traditional cloud from AI new orders, raising suspicion of narrative inflation.
2. Benchmark Two: Whether free cash flow can cover annual AI capital expenditure
The industry has entered a heavy-asset era; AI is no longer a light-asset software business. If cash flow is insufficient to cover compute investments, it means the company is continuously overdrawing future profits for short-term AI stories. Once financing tightens, the narrative collapses directly. Microsoft and Meta meet cash flow standards; Amazon has a clear gap.
3. Benchmark Three: Whether AI has pricing power rather than relying on low-price order grabbing
Office Copilot’s ability to charge separately for AI features is a sign of sustainable commercialization; AWS and cloud providers generally compete for customers with low-price compute, using price wars to gain scale, which will continuously depress profit margins, raising doubts about AI growth quality. #长鑫科技上市，全球存储竞争添变量
That's how impressive it is: ChangXin Memory Technologies officially debuted on the STAR Market today, opening at ¥49.5, up 471%, with its market cap briefly hitting ¥3.7 trillion, directly becoming the top stock in the A-share market. One standard lot earned ¥20,000, and the first hour's trading volume broke ¥100 billion, making it the first A-share stock in history to surpass ¥100 billion in single-day trading volume. Ten years of hard work, truly shocking.
But interestingly, while ChangXin soared, other storage stocks in the A-share market all fell. Gigadevice dropped 5.5%, Protronics fell over 6%, and Bawei Storage and Jiangbolong also declined. In the same sector, the leader's listing drained the smaller players. The funds are limited and all chased ChangXin, so other stocks were naturally sold off. Moreover, ChangXin is a pure DRAM foundry, a type of stock previously absent in the A-share market; other storage stocks are either module manufacturers or distributors, so the logic differs.
ChangXin's global DRAM market share rose from 4.7% in Q4 last year to 7.6% in Q1 this year, surpassing Nanya Technology to become the world's fourth largest. It is expected to net over ¥50 billion in the first half of this year. However, 98% of its revenue comes from traditional DRAM—commodity-grade chips used in servers and smartphones. The most profitable HBM (High Bandwidth Memory), used stacked in AI servers, is almost nonexistent for ChangXin. The three major giants have a technology gap measured in years in this field. Micron earns $28.2 billion in one quarter with an 86% gross margin. ChangXin is still far from earning that kind of money.
More critically, capacity is key. ChangXin raised ¥57.9 billion this time, nearly double the original plan. All this money will be invested in production line upgrades and next-generation DRAM R&D. Morgan Stanley predicts ChangXin's monthly wafer capacity could reach 388,000 by 2028. Counterpoint analysts say it could exceed 300,000 by the end of this year, close to Micron's level. But the technology lags by two generations, and the time gap is three years. Even if capacity catches up, the products are still from the previous generation. Additionally, EUV lithography machine export controls to China are tightening, making advanced process equipment the biggest bottleneck.
The long-term impact of this is much greater than the short-term. The memory chip industry’s boom always ends with the same script—during the upcycle, everyone expands production wildly, new capacity is released in clusters, supply-demand reverses, and prices crash. ChangXin’s listing has armed China’s DRAM industry with a stockpile. In the short term, the high-end profit zones of companies like Micron are temporarily safe, but in three to five years, if ChangXin truly catches up in capacity and technology, the global memory pricing power structure will have to be rewritten. It’s not a matter for today, but today marks the beginning.🚨 $FWA LOOKS LIKE A CASINO — BUT THE NUMBERS ARE WHAT REALLY MATTER.
Here’s the simple breakdown:
«Fake World Assets is basically an on-chain gacha machine built on ETH.»
You put in roughly 0.117 ETH for a random NFT position.
Then you have two choices:
🎰 Keep the NFT you pulled
💰 Or sell it back for 85% of its ETH backing
And these aren’t random junk NFTs either — the system can involve NFTs from major collections.
The interesting part?
There’s no NFT price oracle.
No floor-price feed. No external pricing mechanism.
The depositor sets the ETH backing, and that single number determines the sale price, selection odds, and stake.
But here’s where it gets wild:
🎲 Drawing has roughly -21% EV.
Across 1,981 real settlements, the average dump was around -18.1%.
That’s an incredibly heavy rake — and it’s openly disclosed in the docs.
LPing tells a different story:
📈 Roughly +8.9% per cycle
🖼️ You keep your NFT about 94.5% of the time
But there’s a catch…
LPs are essentially taking a quiet short position on NFT floor prices.
And then there’s the $FWA token.
Right now, it has zero value accrual.
Buybacks have paid out $0 since launch.
Meanwhile, emissions are running at 2% of total supply per day — and they're scheduled to end on August 4 at 19:01 UTC.
🔥 That date could be the real turning point.
Current protocol revenue is reportedly around $289K/day on day 7, which is roughly 2.3x Collector Crypt while sitting at just 11% of its valuation.
But here's the real question:
Can that revenue actually last after the emissions end?
Because if the revenue is sustainable, $FWA could get very interesting.
If it isn't…
The tokenomics could tell a completely different story.
$FWA 👀
#DailyOrbit 🔥 $ETH Breaking: ETH Current Price 1957 Strategy Summarized! The 1957 resistance level determines strength; avoid blindly chasing the rally
🔸 Core Key Price | Current Price 1957
Short-term resistance: 1957
Core watershed: 2000 integer threshold
Short-term support: 1890
Trend Lifeline: 1840
Market boundaries:
With increased volume, it held above the 1957 level, with the bulls continuing their recovery and challenging the 2000 watershed upward;
Multiple attempts to break through 1957 with no volume under pressure and a long upper shadow close, with bullish momentum exhausted. It is highly likely to pull back to the 1890 support and fluctuate to digest floating shares; Effectively breaking below 1840, this round of rebound structure has been broken.
✅ Clear practical approach
Portfolio Partners: Gradually reduce positions and defend near the 1957 resistance range. Do not hold heavy positions and stubbornly break through 2000 directly. Beware of concentrated profit-taking after a pulse surge.
Watching his teammates: Not chasing the current pressure level to sprint! Two prudent plans
(1) Wait for a pullback near 1890 to stabilize, then choose the right time to buy on dips;
(2) Wait for volume to stabilize above 1957, confirm a valid breakout on the hourly chart, then follow after pullback.
Unlimited volume sprint above 1957, blind chasing and gambling is prohibited!
⚠ A rational reminder
Ecological narratives determine long-term value, while trading volume determines the height of short-term rebounds.
Volatility at the resistance level has intensified, with frequent spike rallies. Leveraged traders strictly control their positions and cut losses; avoid one-sided heavy positions and bet on breakouts!
💬 Interactive Q&A: Thought process organized! Are you optimistic that ETH will break through 2000 and continue the rebound, or will it hit resistance at 1957 and fall back into consolidation?
$ETH
⚠ This is only a market strategy exchange and does not constitute any investment advice
Main line targets within the market
$JELLY $OPG $SLX $LAB $CORE $BSB $ALLO $CHIP
⚠ Stability priority: ALLO, BSB🚨 This might be one of the most misunderstood crypto projects right now.
Here's the TL;DR on $FWA (Fake World Assets) 👇
🌀 Think of it as an on-chain gacha machine built on Ethereum.
You pay around 0.117 ETH for a random NFT position. Once you reveal it, you can either:
• Keep the NFT 🖼️
• Sell it back instantly for 85% of its ETH backing 💰
The twist? There are no price oracles or floor-price feeds.
The ETH backing chosen by depositors determines everything:
The NFT's sale price
Your odds of drawing it
The amount at stake
It's an unusually simple system.
🎰 For players, the math isn't great.
Across 1,981 real settlements, the average sell-back loses about 18.1%, making each draw roughly -21% expected value. The protocol is upfront about this in its documentation.
💧 For liquidity providers, the story is different.
LPs earn roughly +8.9% per cycle and keep their NFT about 94.5% of the time—but they're effectively taking the risk if NFT floor prices fall.
🪙 As for the $FWA token...
Right now, it has no direct value accrual. Buybacks have totaled $0 since launch.
The key date is August 4, 19:01 UTC, when emissions of 2% of total supply per day come to an end. That could become a major turning point.
📈 The protocol is reportedly generating around $289K in daily revenue (day 7), about 2.3× Collector Crypt's revenue despite having only ~11% of its valuation.
The real question isn't whether it's making money today—it's whether that revenue is sustainable once emissions end.
What do you think: hidden gem or cleverly designed casino? 👇#DailyOrbit Aave V4 is now available on Avalanche.
This is the first time since V4 completed Ethereum deployment to expand to other public chains.
The first batch of structures to launch includes a core liquidity center, as well as the main market, AVAX-related asset market, and forex market.
What makes this expansion noteworthy is not just that it adds support for another chain.
V4's new structure allows different lending markets to share underlying liquidity while using their respective collateral and risk rules.
If this mode runs stably,
The competition in DeFi lending will gradually shift from "which chain has the most funds" to "who can more efficiently manage multi-chain liquidity."
$AAVE $AVAXStop obsessing over FOMC historical data to predict rises and falls; the core truth has never been fully revealed.
Right now, the internet is flooded with contradictory Federal Reserve meeting statistics. Both bulls and bears each use a set of data to convince retail investors, making it seem reasonable, but in reality, it's all fragmented and one-sided information.
Some statistics show that in the past nine FOMC meetings, eight sessions saw sell-offs one week after, with an average seven-day drop close to 11%;
Another set of data shows that in the past seven meetings, five resulted in upward trends, with an average gain as high as 17.6%;
Long-term cycle data is completely different, with an average 0.9% rise 5 days after the meeting, 3.9% after 10 days, and a direct 11.1% increase after 20 days.
Conflicting data contradict each other, so simply predicting the market based on meeting results is fundamentally unfeasible.
What truly determines market direction is never the FOMC decision itself, but the market sentiment before the decision is released.
When greed is at its peak, the Fed meeting becomes the perfect excuse for major players to offload positions; when the market is deeply fearful, the meeting instead acts as a catalyst for price rallies.
Comparing all current market signals, the present market environment is very clear:
BTC price has stabilized above the key 65000 moving average, with technical structure support;
The Fear & Greed Index is 39, indicating the market is in a fear zone, with no overheated bubble from chasing highs;
Oil prices continue to fall, significantly easing inflationary pressure;
Employment data remains strong, stabilizing the macroeconomic fundamentals;
Market expectations for rate hikes have surged from 13% to 38%, with bearish sentiment already priced in.
Here is a practical strategy tailored for this Fed decision:
You can gradually build a base position below 65000, with stop-losses uniformly set at the previous daily low; this setup offers a very favorable risk-reward ratio.
Do not heavily bet on one-sided moves in the next two days; wait for the decision to drop early Thursday morning before making any decisive moves.
Just focus on the first 15-minute candle after the announcement to determine subsequent actions:
If the official statement mentions phased progress on inflation and easing expectations rise, BTC is likely to surge directly to test the 68000-69000 range;
If there is an unexpected hawkish rate hike, short-term volatility will spike sharply, but since the market is already in a fear zone, this sudden bearish shock could create a rare golden buying opportunity.
Many people lose big every Fed meeting because they blindly rely on historical data and ignore current market sentiment. Don’t fall into the same trap this time.
Are you choosing to wait and watch for the outcome, or have you already started building positions gradually? #美联储周四凌晨公布利率决议 🚨 Everyone's watching AI chips... but the real battle might be happening in memory.
China just made its biggest move yet.
CXMT (ChangXin Memory) debuted on the STAR Market with a 3.31 trillion yuan valuation, instantly becoming the largest stock on China's A-share market. 🔥
That means the global memory race is no longer just Samsung vs. SK Hynix.
Just last week, Anthropic locked in memory supply deals with Samsung and SK Hynix, while Nvidia strengthened its AI partnerships in Korea.
Now, China has officially entered the conversation with a publicly traded memory giant. 👀
The market reacted fast.
KOSPI surged more than 1.7% at the open before reversing, as investors began pricing in the possibility of a third major DRAM player. 📉
From here, keep your eyes on two things:
📌 DRAM contract prices
📌 CXMT's capacity expansion
If supply ramps faster than demand, pricing power could come under pressure—even for today's leaders.
The big question is simple:
Can AI demand support three global memory giants, or is a price war inevitable? 🤔
How are you playing this theme—Korean chip stocks, AI names, or China's A-shares? 👇
#CXMTMemoryIPO
#DailyOrbit 周末很热闹，“交易所排队倒闭”成了热门话题。没有政策压迫，没有黑客攻击，忽然一家又一家的关门，啥原因？
我不懂交易所运营，只是作为一个十多年的老币民直观地觉得，应该是不赚钱，甚至亏钱了。说一千到一万，开门做生意，是要赚钱的，每天一睁眼，辣么多人等着发工资，辣么多日常开支，不赚钱怎么撑得住？
我觉得交易所应该有两大块收入，一个是赚交易者的手续费，现在这个钱不好赚了，熊市太久了，进来就是被套被割，赚钱效应太低，加之股市高歌猛进，新人不来，老人又被消灭或者吸引去了股市，没人交易自然也别想赚手续费了。
交易所的另一块收入，应该是赚上币费之类的发币项目的钱，熊市币圈项目难拿投资，发币的少了，另外头部大所持续大搞Alpha，成为新项目上币的首选途径。所以中型所也赚不到什么项目发币的钱了。
既赚不到手续费，又赚不到上币费。而中型所家大业大，天天吃老本怎么撑得住？放眼熊市还看不到头，所以，只剩下一条路，关门止损。
按照以往的经验，每每交易所纷纷扛不住歇菜的时候，往往就是熊市见底的信号，也不知道这老经验这回还能不能显灵，让我们一起祈祷吧。 #英伟达拟为OpenAI提供2500亿美元担保
2500亿美元——英伟达正在谈判为OpenAI提供一笔巨额融资担保，用于租赁软银在俄亥俄州开发的10吉瓦数据中心项目。 这不仅是AI热潮中最大的一笔金融交易，也标志着英伟达的角色从“AI卖铲人”升级成了“AI基建银行”。
英伟达正在用资产负债表，为AI的未来买单——同时确保未来的订单继续流向自己。
---
① 交易细节：2500亿担保，撬动5000亿项目
这笔担保将帮助OpenAI租赁软银在俄亥俄州南部开发的10吉瓦数据中心园区。项目总成本可能超过5000亿美元，其中英伟达的担保将覆盖数据中心租赁与建设所需的债务，但不包含其中部署的英伟达芯片。此外，英伟达还在讨论另一笔为OpenAI采购芯片提供融资的交易，规模可能高达3500亿美元。
这笔交易的核心逻辑是：OpenAI作为一家尚未盈利的私人企业，缺乏投资级信用评级。英伟达的信用背书就像一张“入场券”，让软银能获得更多融资来推进项目建设。而OpenAI则承诺租下这些算力——形成“英伟达担保→银行放贷→软银建数据中心→OpenAI租用→采购英伟达芯片”的闭环。
② 项目有多庞大？
项目选址俄亥俄州南部派克县的Portsmouth铀浓缩厂旧址——冷战期间为核武器计划提供武器级铀，2001年关停。软银旗下SB Energy将在此建设10吉瓦发电设施，其中至少9.2吉瓦为天然气发电。10吉瓦大约相当于一座大型核电站的全部输出功率，能满足约800万户美国家庭的用电需求。
项目第一阶段预计2028年完工，提供约800兆瓦电力。整座园区满载预计持续建设至2030年代中后期。
③ 英伟达的角色升级：从芯片供应商到基建银行
此前市场曾传出英伟达缩减对OpenAI千亿级资金支持的消息。早期规划中，英伟达曾考虑1000亿美元的直接投资计划，但因OpenAI推进上市进程而搁置。如今从直接股权注资转向2500亿美元项目融资担保，英伟达不仅大幅提升了资金承诺量级，更在支持方式上实现了实质性升级。
这种模式能规避OpenAI上市前的估值博弈与股权稀释风险，以较低资本占用撬动庞大基建项目，同时通过深度绑定确保OpenAI未来庞大的算力采购订单持续流向英伟达。 这不是投资，是锁定。
④ 质疑：循环融资与泡沫风险
批评人士指出，英伟达通过投资AI公司、为基础设施建设提供支持，可能正在人为制造更高的市场需求，进一步推高AI产业估值泡沫风险。
⑤ 对加密市场的传导
英伟达对OpenAI的2500亿美元担保，本质上是AI基建的进一步加速。对加密市场的影响是间接但深远的。从算力需求看，AI数据中心的扩张正在消耗全球大量电力资源，比特币矿工面临的能源竞争只会加剧。从叙事传导看，科技巨头持续加码AI基建叙事，对AI+Crypto赛道的市场情绪是间接支撑，但对存储芯片等硬件端的短期影响有限——长鑫科技上市、存储股波动仍是更直接的变量。
从“卖铲子”到“当银行”，英伟达正在重新定义自己在AI产业链中的位置。这一模式能否持续，取决于AI的最终需求是否真能填满这些即将建成的数据中心！ 比特币铭文赛道的开山币种ORDI，掀起过一轮轰轰烈烈的铭文大牛市。无数人依靠铭文叙事实现盈利，也吸引大批量散户跟风入场。如今热潮褪去，市场目光不断转向AI、新公链热点。比特币铭文新增流量增速放缓，很难复刻当初全民参与的盛况。依旧坚守等待二次风口的人不在少数，但风口从来不会按照人的预期到来，存量博弈环境下，复刻历史涨幅难度极大。 「巨鲸发现」千万级原油巨鲸套利失败，美、布原油反向开仓亏达45万美元
0xa314 开头巨鲸在原油大涨后，将仓位调整为做多 WTI、做空布伦特，形成一笔总规模约 1500 万美元的跨品种配对交易，或预期 WTI 延续此前相对强势。
7 月 14 日至 23 日，Hyperliquid 上 WTI 与布伦特分别上涨约 16.4% 和 13.1%，WTI 明显跑赢。该巨鲸随后建立约 9.1 万份 WTI 多单，并将布伦特空单扩大至约 10 万份。
但地缘风险降温后，两种原油同步回落，WTI 跌幅略高于布伦特，配对交易未能兑现。由于 WTI 多单建仓成本较高，其亏损亦超过布伦特空单收益。
截至发稿，该巨鲸持有 8.6 万份 WTI 多单及 9 万份布伦特空单，两腿总价值约 1514.64 万美元。其中，WTI 多单浮亏约 61.16 万美元，布伦特空单浮盈约 19.80 万美元；叠加此前已实现的约 4.59 万美元减仓亏损与 1.5 万美元资金费率，该轮美布原油价差交易累计亏损约 44.44 万美元。
该地址非专做原油，记录中，其最初集中交易 NVDA、DRAM、MU、SNDK 等半导体及存储标的，同时参与 SP500、XYZ100 等股指合约，历史盈利 860 万美元#美联储周四凌晨公布利率决议
美联储这周开会，加息概率从一周前的13%飙到了38%。
① 数据
7月28-29日，美联储将召开议息会议。一周前，市场还觉得加息概率只有13%，现在CME数据显示加息25个基点的概率已经涨到36.3%-38%。
更关键的是9月——维持利率不变的概率只剩19.6%，加息25个基点的概率55.2%，加息50个基点的概率25.2%。市场已经完全定价9月会加息。
但彭博对76位经济学家的调查显示，所有受访者都预计7月维持利率不变。经济学家和市场交易员，对同一件事给出了完全相反的判断。
② 为什么加息概率突然暴涨？
油价。 美伊冲突期间布伦特一度逼近94美元。油价涨，通胀预期就涨。通胀预期涨，美联储就被迫更鹰。
沃什的风格。 这位新主席上任后已经明确表态——不提前给指引，一切看数据说话。他不像前任那样给市场明确信号，市场就只能用概率来猜。
内部已经在分裂。 6月点阵图显示，9位官员预计年底前至少加息一次，其中6人预计加息两次。文艺复兴宏观的首席经济学家甚至直接放话：“为什么不现在加息？”
③ 市场在担心什么？
通胀可能卷土重来。 6月CPI确实降到了3.5%，但那是油价还没涨上来的时候。7月油价涨了这么多，下个月的CPI大概率不会这么好看。
就业市场太强。 初请失业金18.7万人，七年最低。美联储想降息，就业数据不给理由。
④ 对加密市场意味着什么？
BTC已经站回65,000，恐慌指数回升到30。如果美联储维持不变但声明偏鹰，市场可能会先跌后反弹。如果意外加息，BTC可能快速回踩62,000-63,000。如果维持不变且措辞温和，BTC可能挑战67,000-68,000。
今晚到明天，纳指期货和美债收益率的波动会比K线更早告诉你方向。市场在看沃什怎么说——是会“维持不变但保留加息选项”，还是会释放更明确的信号。答案周四凌晨见分晓。#财报观察员：微软Meta亚马逊能稳住AI叙事吗？
Recently, many people have been waiting for the earnings reports of Microsoft, Meta, and Amazon. However, I feel that this time the market is less concerned about whether the EPS beats expectations.
Instead, the key question is: Has AI actually started to generate profits?
Over the past year, the AI narrative has driven tech stocks to new highs, but the market's tolerance is declining.
A few days ago, although Alphabet delivered decent results, the market reacted negatively due to the raised expectations for AI capital expenditure (CapEx), sparking concerns about rapid investment with slow returns, which clearly pressured the stock price.
This means the earnings reports of Microsoft, Meta, and Amazon face greater pressure than before.
Microsoft needs to prove that the revenue growth from Azure AI is enough to cover the expanding compute investments; Meta must convince the market that its multi-billion-dollar data center investments will eventually translate into advertising revenue and profits; Amazon has to demonstrate that AWS and AI services can continue growing, not just that capital expenditures keep rising.
I believe this earnings season will be a watershed moment for AI investment logic.
Previously, the market believed in "invest first, profit later."
Now, the market wants to see "after investment, profits have already started."
If all three companies can simultaneously prove that AI investments are continuously converting into revenue, then the entire AI industry chain—GPU, storage, networking, optical modules, and even the Physical AI direction I have been following—stands to benefit further.
But if revenue growth cannot keep pace with capital expenditures, the market may further compress valuations in the AI sector, and even good earnings reports may not lead to stock price increases.
So this week, when I look at earnings reports, my first focus won’t be on EPS but on whether management answers one question: Has AI been falsified, or can it truly bring massive profits? $META One of today's standout tokens: 💥 Euler rose sharply after the information was listed on Upbit KRW, one of South Korea's largest crypto exchanges. Appearing on Upbit often generates big attention thanks to: Korean retail cash flow.
📊 New liquidity from the KRW trading pair.
⚡ FOMO sentiment when the token is accessible to a larger number of users. 🏦 The story behind $EUL Euler is a DeFi protocol focused on: 💰 Lending & borrowing.
⛓️ On-chain asset management.
🔧 Decentralized finance mechanism币安 RLUSD 年化收益率高达 22.25%，但又担心 0.1% 的溢价吃掉收益？
分享一个不用追高买入 RLUSD、也不用担心活动结束后 RLUSD 回落，同样可以参与币安 RLUSD 活动的策略。
核心工具：Euler
目前可以将图中的三种稳定币作为抵押品，在 Euler 借出 RLUSD，再把借到的 RLUSD 存入币安参与活动。
求稳的话，可以优先选择 USDC 作为抵押品。
Euler 和 Morpho 有一个比较明显的区别：
抵押进 Euler 的稳定币本身也可以继续产生收益。
因此，扣除 RLUSD 的借款成本之后，抵押端仍然可以获得约 0.85%—2.65% 的净收益，同时借出的 RLUSD 还能参与币安的活动。目前可用流动性有6.2M 资金容量也不错。
这样做的好处是：
- 不需要顶着 RLUSD 的高溢价买入；
- 不需要承担活动结束后 RLUSD 回落、吃掉理财收益的风险；
- 抵押资产本身还能继续赚取收益
代价就是多了一层Euler的合约风险，以及降低了一点资金效率。
以 USDC 为例，Euler 的 Max LTV 为89%，也就是说，抵押10万美元USDC，最多只能借出约8.9万美元RLUSD，无法做到100%的资金利用率。
不过，由于抵押品和借款资产都是主流稳定币，因此清算风险相对可控。
这套策略本质上是：
牺牲一部分资金效率，换取不追高RLUSD、不承担溢价回落风险，同时保留活动收益。
自用策略分享，不构成投资建议，请一定DYOR #长鑫科技上市，全球存储竞争添变量
Today, Changxin Technology went public, soaring 471% at the opening, feeling like the top of the A-share market.
I want to say this is a milestone step for domestic semiconductors, but it is still far from shaking the pattern of the three giants.
The core significance of Changxin's listing is that it has obtained sufficient ammunition to pursue process technology and expand production capacity, truly breaking the global DRAM "three-company monopoly" into a new pattern of "three strong and one weak." Short-term sentiment is very heated, and the first-day valuation has already overdrawn many performance expectations, so don't blindly chase the high; in the long term, general storage domestic substitution is a definite trend, but the high-end HBM track still lags by 2-3 years, and the real tough battle is yet to come.
Because whether in terms of market value or technology, there is still a considerable gap.
Changxin's listing has rewritten the global storage competition pattern. The long-term logic of domestic substitution is very solid, but the short-term sentiment premium is too high, and the technology gap still needs time to be bridged Crude oil risk premium squeeze and strong employment data intertwine, with the core market conflict on the eve of the Federal Reserve interest rate decision shifting to the macro battle between "cooling inflation" and "prolonged high interest rates."
Geopolitical easing expectations drove crude oil futures $CL to plunge more than 5% intraday, and the decline in energy prices directly lowered short- and medium-term inflation expectations. Meanwhile, initial jobless claims last week fell to 187,000, indicating the labor market remains resilient and strong.
In terms of driving factors, the Federal Reserve FOMC meeting statement forms the primary trading theme, while tech giants' earnings and capital expenditures serve as the secondary liquidity indicator. The crypto fear and greed index rose back to 30, coupled with BTC surpassing the $65,000 mark, indicating increasing market pricing for a soft landing.
The upside scenario requires the meeting to release dovish signals and tech giants' earnings to exceed expectations. If the Fed confirms the downward inflation trend, U.S. stocks and crypto markets will see cross-market liquidity recovery, but attention should be paid to the approximately $900 million FTX compensation starting July 31, which may cause short-term pressure on market chips.
The downside scenario depends on labor market resilience triggering hawkish signals or sudden geopolitical fluctuations. If the Fed emphasizes maintaining high interest rates longer than expected, oil prices $CL will rebound again, and a stronger dollar and U.S. Treasury yields will trigger cross-market risk aversion, with tech stock pullbacks dragging crypto assets down simultaneously.
The failure conditions for these two scenarios lie in a full return of geopolitical risk premiums or an unexpectedly implemented Fed policy shift. Once crude oil regains upward momentum and recovers losses, the cooling inflation logic fails; if jobless claims continue to be lower than expected, delaying rate cut expectations entirely, the macro pricing framework will be reconstructed.
In the next 7 days, key focus should be on the Fed interest rate decision statement early Thursday, changes in tech giants' earnings and capital expenditures, and the actual market absorption capacity of FTX compensation funds flowing in starting July 31.
#新手必看：这里有你需要的一切 #英伟达拟为OpenAI提供2500亿美元担保#美联储周四凌晨公布利率决议 很多人都在猜，这次是维持利率还是加息。
但我认为，比结果更值得关注的是Warsh时代的第一套货币政策逻辑。
鲍威尔时代，市场习惯了通过讲话、点阵图和前瞻指引去提前交易预期；而Warsh上任后，最大的变化就是弱化前瞻指引，让市场重新回到”数据决定政策”。
这意味着，未来市场波动可能会更大。
因为当央行不再提前告诉你答案，资产价格只能不断根据新的经济数据重新定价。
对于BTC、美股AI板块以及风险资产而言，这不仅是一场利率会议，更是一次市场定价机制的转换。
我的交易哲学一直很简单：
交易的不是新闻，而是市场如何消化新闻；交易的不是观点，而是预期如何发生变化。
所以这次会议，我不会急着预测方向，而是重点观察三个信号：
* Warsh如何评价当前通胀；
* 是否继续坚持弱化前瞻指引；
* 市场会不会因此重新定价未来几个月的流动性预期。
真正的机会，往往来自市场认知发生变化的那一刻，而不是利率公布的那一分钟。$ETH Is Changxin's IPO bad news for Samsung, SK Hynix, and Micron? #长鑫科技上市，全球存储竞争添变量
Short term, it's not bad news; long term, definitely yes!!
Today many are discussing Changxin's IPO surge, but I'm more concerned about another thing: will Samsung, SK Hynix, and Micron lose sleep over this?
Changxin's IPO won't reshuffle the global DRAM market overnight. In the AI era, HBM remains the most profitable business for Samsung, SK Hynix, and Micron, and Changxin still has a clear gap to close.
So in the next two to three years, their real competitors remain each other.
But the capital market looks to the future.
The large amount of funds raised by Changxin this time is not just for expanding production but also means more chips for future R&D investment, advanced processes, HBM, and other high-end storage fields.
If global storage competition used to be a "Three Kingdoms Kill" game,
it has now officially become "Three Kingdoms Kill + a rapidly upgrading new player."
I have always focused on Physical AI, so I pay special attention to the storage industry.
Because in the future, whether AI servers, robots, or autonomous driving, they all fundamentally rely on DRAM and HBM.
Whoever can capture the next generation of high-end storage will reap the biggest dividends from the next round of AI infrastructure.
So in my view, the greatest significance of Changxin's IPO is not how much it rose today, but that the global storage industry has for the first time seen a competitor truly worth Samsung, SK Hynix, and Micron's long-term vigilance.
In the next three to five years, I won't change my judgment based on one company's stock price fluctuations, but I will keep watching one indicator:
When will Changxin truly enter the HBM market?
If that day comes, I believe the valuation logic of the global storage industry may be rewritten Guys, CVX rose 5.55% today, currently at $1.387, continuing to recover from the historical low of 1.04 in early July. The logic behind the rise: CVX automatically reinvested at 29% annualized, AbcCVX staking up to 2.5x bonus—staking yields have increased significantly, holders are incentivized to transfer tokens into locked, and the circulating market is passively tightened. Analysts clearly state: once the lock-up curve continues to climb, supply-side scarcity will first be reflected in prices. The total supply cap is 100 million, with a current circulating ratio of about 92%; The original unlock cycle for teams and early investors has ended, but the protocol will continue to slowly add new mints alongside LP mining. Key price levels: Resistance $1.40-$1.42 (break out to $1.45-$1.50), support $1.30-$1.33 (break to $1.20-$1.23). Risk: The sustainability of protocol returns depends on the stability of Curve's TVL, and whether the staking APR can be maintained is key. Recent key indicators to watch—staking APR trends and staking growth rate. For assets that have dropped over 97% from their ATH and whose fundamentals are improving, focusing on value lock growth and staking APR is more important than focusing on candlesticks. Personal market view analysis and market information compilation, not investment advice. $BTC $ETH $CVX #长鑫科技上市, global storage competition adds variables #美联储周四凌晨公布利率决议 #财报观察员: Can Microsoft, Meta, and Amazon hold the AI narrative? $CRWD
Price action is trading around 185.57, holding above dynamic MA5 (184.49), while sitting below MA10 (190.08) and MA20 (193.27).
EP
183.00 - 185.50
TP
190.08
193.27
205.00
SL
179.00
Price pulled back significantly from its recent high of 219.18 and is attempting to bounce off lower support levels around 180.00. Reclaiming MA5 (184.49) provides initial signs of stabilization for a move toward MA10.
Let's go $CRWD
#CXMTMemoryIPO #FOMCRateWatch #AIEarningsWatch ₿ $BTC Volume Profile Update 📊
Is anyone still watching the volume profile? 👀
The $64K level where I mentioned taking profits was not random — it was the Point of Control (POC), the area with the highest traded volume in that range.
Combined with additional confluences, it provided a strong technical reaction zone. 🎯
Volume profile remains one of the most useful tools for identifying:
📌 High-interest price areas
📌 Key support & resistance zones
📌 Potential reaction levels
Price often respects where the most business was done.
Don’t just watch candles — understand where the volume is sitting.
#BTC #Bitcoin #Trading #TechnicalAnalysis #Crypto📰 美伊谈霍尔木兹妥协！特朗普军事牌还攥着，BTC $65,382能冲吗？
事件概述
说白了就是，美国和伊朗正在接触，想把霍尔木兹海峡的通行问题谈拢。全球三分之一的海运石油从这过，谁控制了这里谁就掐住了能源命脉。但特朗普这边军事选项一直没拿掉，相当于一手谈一手亮剑。对币圈来说，至少双方在谈而不是直接开干，短期情绪上是好事。
深度解读
为什么这条新闻重要？
老铁们，霍尔木兹这事不是今天才有的。今年以来伊朗在该区域动作不断，市场一直在怕一个事：万一真封锁海峡，油价飙升，通胀反弹，美联储降息就彻底没戏了。降息预期一旦崩了，币圈流动性直接断崖。
现在美伊开始谈妥协方案，说明两边都不想真打。伊朗经济被制裁搞得够呛，美国也不想在大选年搞出中东全面战争。这是务实的信号。
讲真，今天盘面已经说话了。BTC回到$65,382涨了1.36%，ETH更猛直接$1,955.75涨了3.70%，资金在用脚投票。地缘风险降温 = 避险情绪回落 = 风险资产回暖，这条逻辑链很清晰。
但注意，特朗普说军事选项还在桌上。这不是废话，这是给谈判施压同时给自己留退路。如果谈崩了随时能升级。所以这个利好是有保质期的，别无脑冲。
对市场的影响
短期看，避险资金从黄金和美债里撤出来，部分会进风险资产。ETH今天3.70%的涨幅说明聪明钱已经在抢跑了，跑得比BTC还快，这通常意味着市场风险偏好在快速修复。
中期来看，如果谈判出实质进展，比如伊朗承诺不干扰航运，油价回落 → 通胀预期下降 → 降息预期回来 → 币圈迎来真正的喘息窗口。这条链路是通的，也是我在盯的主线。
但如果谈崩呢？说实话概率不高但不是零。特朗普的操作风格你们懂的，今天谈明天翻脸。真到那一步，BTC可能要回踩$62,000-63,000区间找支撑。
参考2020年初美伊那波冲突，BTC短线上下震荡超过5%，但地缘事件只是短期扰动，不改大趋势本质。这次大概率也是同款剧本——恐慌砸坑，然后V回来。
操作思路
🎯 影响预判
- 币种：BTC / ETH
- 方向：利多📈 预测涨
- 时长：BTC 12小时 / ETH 24小时
💡 短线思路很明确：别空。BTC回调到$63,500-64,000区间如果不破就是接货位，止损放$62,500下方。ETH今天这个走势非常有劲，站稳$1,955.75上方的话下一个目标直接看$2,000整数关口。但仓位控制在五成以内，谈判这东西一天一个说法，留好子弹防黑天鹅。
认同比特币这波反弹逻辑的，点个赞让我看看有多少人跟上节奏了
$BTC $ETH #BTC #ETH
#地缘政治
⚠️ 不构成投资建议，预测仅供参考This AGLD dish is basically a forgotten scrap in the kitchen—the plate is cold—but I smelled the aroma of twice-cooked pork in the oven! 🔥 The one-hour RSI of 35.66 is like the precise final three seconds of sous vide cooking—if you don't go overheated, the meat will get tough. The lower band of 0.1477 is slippery against the cutting board, and the 4-hour lower band at 0.1457 is the anti-slip mat of the prep station—the market players are crouching by the stove to steal your stop-loss orders!
The main course (mainstream coins) now looks like overnight broth—steady but without surprises; AGLD's "ginger and scallion stir-fried liver kidney" — price dropped 2.11%, but the KDJ (here referring to RSI) is at 35.66, like the raw green stage before chili peppers are fried in oil, only fragrant after stir-frying. Enter the market at 0.1431, like a sliced kidney flower that needs to be quickly slit. The first target is 0.1563 for the crisp after the slices are cooked, and the second target is 0.1580 for the final pouring of hot oil, sizzling. Stop loss at 0.1275—that's like a wet rag under your cutting board—never let blood get on the spicy pot.
What about the spiciness? This time, I only put in "mild spiciness"—leverage up to 1.5 times, with a position of 5% of total capital, like sprinkling a handful of Sichuan pepper powder to enhance the aroma but not spiciness. The 4-hour Boolean upper band of 0.1580 is the rim of the pot; I don't bother touching the spatula, it's too hot to handle. Remember: RSI 1D is still at 41.94, like marinated ribs that haven't fully absorbed the flavor. Wait for the hourly chart to flip and then turn up the heat.
I'm about to put this dish into the pot, but let me remind you—the extra spicy chili jar is always on the table. Whether you pick it up depends on how steady your hands are.日内高点3.76u，日内低点3.3u，现价3.5u，24小时最大跌幅11.2%；全线跌破3.5u支撑位，回调修复，多头承接力较强。 7月27日回调五大核心利空逻辑 1. 短期炒作热度彻底消退，纯情绪拉盘无基本面支撑（核心导火索） 前一日上涨完全依靠海外加密博主统一喊单吸引散户追高，7月27日无任何合作、版本更新等实质性利好，网红流量热度快速消散，增量资金瞬间断流。项目仅靠跳舞挖矿叙事，平台用户增长停滞，无真实代币消耗场景，失去买盘支撑出现回落。 2. 高位巨鲸集中获利了结，低流动性放大回调幅度 流通盘占比不足14%，前十大巨鲸持有流通盘超65%，拉升后浮盈空间充足。大户集中大额抛售，盘面订单深度薄弱，少量卖单即可快速击穿关键支撑位，引发散户恐慌跟风卖出，形成短期踩踏行情。 3. IP版权负面持续发酵，市场信任度下滑 舞蹈IP版权方公开澄清未与BEAT项目达成官方合作，项目方虚假宣传叙事被戳破，投资者担忧后续项目下架、维权风险，大量持仓用户恐慌减仓离场，进一步加剧抛压。 4. 代币远期解锁预期压制，资金规避长期通胀风险 市场提前计价未来大额代币解锁，社区、团队份额分4年持续线性释放，长期#长鑫科技上市，全球存储竞争添变量
ChangXin Technology IPO: Under the AI computing power wave, a domestic variable emerges in the storage sector
While Google, Microsoft, and Meta are all revealing their AI capital expenditures, the storage sector welcomes a heavyweight player: ChangXin Technology officially listed on the STAR Market, with a first-day market value surpassing 3.3 trillion, becoming an unignorable "fourth pole" in the global DRAM market.
The AI boom has driven hype around computing power, servers, and optical modules, but ultimately cannot bypass the underlying storage chips. ChangXin's listing precisely adds the most critical domestic variable to the global storage landscape in the AI era.
The AI-driven storage market surge, ChangXin capitalizes on a differentiated dividend
In this round of AI computing power explosion, the first beneficiaries were HBM (High Bandwidth Memory), with Samsung and SK Hynix earning huge profits and shifting capacity toward high-end HBM.
However, demand for general-purpose DRAM is also growing—basic memory needs for servers, consumer electronics, and automotive electronics continue to rise. After the three giants moved capacity upward, a supply gap appeared in the mid-tier market. ChangXin accurately positioned itself in this window, rapidly scaling DDR5 and LPDDR5 products, with market share climbing from under 3% to 7.67%, and performance booming accordingly.
This is equivalent to the leading players competing for the high-end, high-margin market, leaving the mid-tier cake to ChangXin—an opportunity granted by the times.
66.6 billion yuan fundraising, expansion + R&D, targeting the next round of competition
This IPO raised up to 66.6 billion yuan, with core investments in two directions:
1. Expansion of 12-inch wafer fabs: continue increasing general DRAM capacity to capture more mid-tier market share, aiming for over 10% market share by 2026;
2. R&D of advanced processes and 3D DRAM: unable to adopt EUV due to equipment constraints, focusing instead on breakthrough 3D DRAM differentiation technology, attempting to overtake on next-generation storage technology to meet future AI storage demands.
Chain impact on AI tech stocks
Upstream equipment and materials chain: ChangXin's expansion will drive orders for domestic semiconductor equipment and material suppliers, further strengthening the domestic substitution logic;
Downstream server/terminal manufacturers: having an additional local supplier means more controllable storage procurement costs and a safer supply chain, a long-term positive for domestic cloud providers and consumer electronics manufacturers;
Storage cycle fluctuations: new capacity deployment may intensify price competition in general DRAM, potentially increasing industry cycle volatility, a double-edged sword for pure storage stocks.
To conclude honestly:
In the AI era, storage is as fundamental as computing power infrastructure. ChangXin's listing is not an end but the beginning of domestic storage's participation in global AI industry competition. The gate for high-end HBM is not yet passed, but at least it is now at the table.BTC 与山寨分化加剧，市场宽度已压缩至极端水平
当前市场是否已进入仅剩少数标的能跑赢的结构性行情？
原文数据表明：在追踪的 100 只山寨币中，上涨/下跌比例仅为 0.3，即每 1 只上涨对应约 3 只下跌。仅 8 只标的出现正向成交量背离，其余 92 只处于持续流出状态。这 8 只分别为 JELLYJELLY、OPG、SLX、LAB、BSB、ALLO、CHIP、ZKP，而 BEAT、EDGE、COAI、TRUMP 等 92 只则普遍承压。数据本身是静态截面，未说明时间窗口，但已足够揭示当前定价结构。
- 市场结构的核心变化：山寨币整体呈现资金净流出，宽度极窄，说明增量资金并未扩散，而是集中在极少数标的。这种环境通常意味着：BTC 和 ETH 若未能带动整体情绪回升，则山寨板块的尾部风险可能进一步释放。
- 传导逻辑：BTC 若维持震荡或小幅上行，可能进一步吸收有限的风险偏好，导致山寨资金继续向少数高动量标的集中，形成"赢家通吃"格局。ETH 若无法突破关键阻力，则山寨轮动逻辑更难以成立。
- 偏多路径：若 BTC 突破近期震荡区间并放量，可能带动流动性溢出至 ETH 及上述 8 只正向背离标的，形成局部反弹。条件是 BTC 日线收盘站稳并伴随现货 ETF 净流入放大。
- 偏空风险：若 BTC 回落或横盘，山寨整体将继续承压，上述 92 只标的可能加速下行。风险信号是 BTC 未突破前高，且山寨上涨/下跌比例持续低于 0.5。
结论：当前市场宽度不支持全面做多山寨，仅少数标的具备独立动量。后续应重点观察 BTC 能否有效突破，以及 ETH 是否出现跟涨迹象，以此判断窄幅结构是否会扩散或收敛。
风险提示：宽度数据是滞后指标，极端值可能预示反转，但也可能延续。Cashtags: $BTC $ETH $JELLYJELLY $ZKP $SLX。#财报观察员：Can Microsoft, Meta, and Amazon Stabilize the AI Narrative? Microsoft, Meta, and Amazon Report Next Week—Is the AI Narrative Losing Ground? Seven Tech Giants Lost $800 Billion Overnight, Crypto AI Concepts Kneel First in Respect
Google just raised its 2026 capital expenditure cap by another $15 billion, with Q2 free cash flow turning negative at $5.9 billion. As a result, on July 23, the seven tech giants evaporated $797 billion in a single day, with Google down 7% and Tesla down 14.5%.
Next week, Microsoft, Meta, and Amazon will release their earnings. The market isn’t just looking at whether revenue beats expectations, but at a single yardstick: You’re spending $190 billion/$145 billion/$200 billion a year—how much gross revenue is Azure/AWS/advertising AI actually bringing back?
Here are the three companies’ key cards:
• Microsoft: FY2026 CapEx guidance around $190 billion, Azure growth still 39%-40%, AI annualized ARR $37 billion (doubling year-over-year), but CapEx-to-revenue ratio is already three times that of the last cloud cycle, with free cash flow forecast cut in half to $25.3 billion
• Meta: CapEx raised to $125 billion–$145 billion, full suite DAP growth hitting the lowest since 2021, AI spending mainly on recommendation ads and self-developed chips, monetization path the most unclear
• Amazon: AWS growth back to 28% (highest in nearly 3 years), but quarterly CapEx $43.2 billion, up 85% year-over-year, annual around $200 billion, free cash flow forecast to remain negative for the full year
The logic is simple:
Chip stocks (NVDA/MU) sell the shovels; cloud giants use the shovels. Now the shovels are selling like crazy, but the users’ cash flow is bleeding—this is the real pricing behind Microsoft -2.3%, Amazon -4.2%, Meta -3.8% on July 24.
What this means for crypto:
Recently, $BTC has been stable, and AI Agent/DePIN tokens have been flying by riding the US stock AI hype, essentially an overflow of the big brothers’ narrative premium. If the big three continue to cut valuations after earnings, the first wave of AI clone concepts to be killed will be those with no revenue, pure narrative, and high FDV. Conversely, if the three companies’ guidance shows Azure/AWS growth plus Copilot paid user numbers holding up, capital will flow back from chips to platform stocks, risk appetite will return, benefiting BTC and mainstream first, and AI clones will get a second wind.
My own market sense:
If BTC doesn’t break 62k and ETH doesn’t break 1600, don’t rush to cut AI clones; but if any of the three continue to raise CapEx or give weak cash flow guidance, don’t get stubborn—delever first. Changxin, creating the impossible in A-shares
122.1 billion
Increase of 471%
Market value successfully tops A-shares
【Pump, Engine】
Short-term withdrawal of some market liquidity, putting pressure on storage concept stocks and some tech stocks
Medium to long-term injection of certain orders and a new valuation anchor into the domestic storage industry chain
Other information
Status:
Changxin is China's largest and most comprehensively laid out integrated DRAM storage chip company, covering DDR4 and DDR5. By Q4 2025, it is projected to be first in China and fourth globally, with a 7.67% share.
However, the top three globally—SK Hynix, Samsung, and Micron—account for over 90%, and Changxin remains a pursuer.
"Financial backers":
The top five shareholders are mainly Hefei, Anhui state-owned assets, and the National Integrated Circuit Industry Investment Fund Phase II. Industrial capital includes Gigadevice, Midea, Xiaomi, Alibaba, and also the National Adjustment Fund.
A trinity of state-owned, industrial chain, and market-oriented capital supports the flagship example of independent and controllable domestic storage.
⚠️ A market value of 3.31 trillion corresponds to revenue of 110 to 120 billion yuan in the first half of 2026, with cumulative losses just expected to be wiped out.
The surge on the first day is more driven by scarcity and the narrative of domestic substitution sentiment pricing rather than current fundamentals.
The milestone significance outweighs performance fulfillment; the real focus is whether it can catch up in market share and technological gap.
#长鑫科技上市，全球存储竞争添变量 【图文观察｜央行周温度】北京时间13:15，金十文章关注：美联储决策从“主席引导”转向“委员博弈”，达成共识难度拉满。
背景摘要：市场不再相信“口头抗通胀”，美债正在逼美联储亮牌。分析师存在两种推演：一是委员会集体推动加息，沃什顺势附和；二是沃什暂时凭借自身影响力按兵不动，但未来仍需加息。
跨资产快照：现货黄金 4,094.57（+1.02%）；欧元/美元 1.1406（+0.33%）；美元/日元 163.56（-0.18%）。黄金、欧元和日元能同时反映利率预期、美元强弱与避险需求，适合作为加密风险偏好的外部温度计。
验证点：若黄金走强而美元同步转强，风险资产更可能承压；若美元回落且美股修复，BTC/ETH更容易跟随风险偏好回暖。
风险提示：央行口径、PCE/CPI或就业数据若超预期，上述跨资产观察需要重新评估。仅作市场观察，不构成投资建议。#Gate.io版临时工
Gate官方持续声称对接我们ALD社区的Robin是冒充人员、骗子，这里有几个无法回避的核心疑问，请正面答复：
1. 如果Robin仅仅是外部骗子、并非Gate工作人员，一名不受官方授权的冒充者，凭什么拥有权限完成Gate Alpha完整上币流程，成功将ALD代币上线平台？
Gate上币具备内部多层审批机制，绝非外部人员可以私自操作。倘若外人随便冒充员工就能完成代币上线，是否证明Gate内部权限管理彻底失控，任何人都能冒充工作人员主导项目上币？
2. 我们按照对接人要求，足额支付上币对应的USDT与ALD。若Robin属于个人欺诈，为何骗子指引我们转账的资金最终流入Gate体系，并且代币如期上线？
普通人实施诈骗，目标是私自侵占资金；而本次资金交割完成后代币成功上架平台，完全不符合普通骗子的作案逻辑。
3. Gate不能简单用“对接人是骗子”单方面撕毁双方达成的上币约定。
代币成功上线Gate Alpha是客观既定事实，交易行为、履约结果真实发生。不能享受项目方缴纳费用带来的收益，同时以“人员冒充”为由拒绝履行全部协议义务。
4. 希望Gate公开本次ALD上线Gate Alpha完整审批链路、内部经手工作人员。
如果Robin无任何官方授权，请解释：一名外部冒充者，是如何绕过全部内部风控、审批，打通上币全流程的？ 这是否意味着Gate Alpha上币渠道存在重大漏洞，所有项目方都面临被虚假人员诱导的风险？