Post

血洗前夜？油价月涨20%撞上PCE转负，美联储正在失去“降息借口”！ 7月31日盘中，WTI原油期货一度跌超3%至80.12美元，布伦特跌近3%报84.4美元。但全月算下来——WTI累计涨了约20%，布伦特一度突破90美元。整个7月，美伊在约旦互炸。7月中旬伊朗袭击美军基地，2名美军死亡；7月29日伊朗弹道导弹再袭约旦，美军报复空袭。油价跟着导弹一起飞。一个月内，缓和与升级切换了两次。就在油价飞天的同一天，美国公布了6月PCE数据——环比下降0.1%，四年来首次月度转负。年化PCE 3.7%，核心PCE 3.3%。一边是油价月涨20%，一边是通胀指标四年来首次转负。这叫什么？叫滞胀幽灵。拆开看，这条逻辑链有多致命：第一环：油价是通胀的发动机。布伦特从71美元冲到90美元上方，能源成本沿产业链逐级传导。霍尔木兹海峡通航量降至零，这不是短期脉冲，是供应端的结构性冲击。第二环：通胀预期逆转，直接改写利率路径。7月初市场认为9月加息的概率只有18%，到7月20日已飙到61.4%。CME FedWatch最新数据显示，9月加息概率仍高达65.2%。第三环：7月30日美联储刚宣布维持利率在3.5%-3.75%不变——但内部已经炸了。3名票委投出加息反对票，2016年以来首次。第四环：加息预期推高美元实际利率，而实际利率是风险资产定价的锚。利率上升=未来现金流折现率提高=估值倍数收缩。这套组合拳打下来——美股、AI股、加密货币，谁也跑不掉。说人话：市场过去半年一直在交易一个故事—— “通胀回落→美联储降息→流动性释放→BTC起飞” 。油价站上90美元，这个故事的基础没了。市场现在交易的是另一个故事—— “油价涨→通胀反弹→加息→流动性收紧→风险资产承压” 。你手里拿的BTC，定价逻辑已经变了。最近BTC在6.4万美元附近震荡。美股却在深V反弹，道指涨1.19%，纳指大涨2.78%。为什么？美股在交易“GDP增速放缓→美联储不敢加息”的软着陆叙事。但加密市场在交易什么？流动性收紧的预期。同一个宏观数据，两个市场，两种解读。谁对谁错？看油价。 WTI如果守不住81美元（7月31日盘中已跌破80），市场在交易“需求崩塌”的衰退逻辑。站不稳81，大盘有系统性回调风险。站稳了，才是利空出尽。说句难听的：不要问打仗对币圈有什么影响。你只需记住—— 每次中东起飞，都是大户对冲做空大盘的信号枪。 $BTC $BZ $CL #美伊报复循环加速，油价月内累涨20%

The market is repricing "burning money"; 220 billion is not a cost, it's a barrier Amazon released its earnings yesterday. Revenue was 200.6 billion, beating expectations. AWS revenue was 42.2 billion, up 37% year-over-year, the fastest growth in 18 quarters. Operating profit was 27.5 billion, up 43% year-over-year. So what? Full-year capital expenditure was raised from 200 billion to 220 billion. Free cash flow turned negative, with a net outflow of 7.6 billion. Q3 revenue guidance is 197 billion to 202 billion, below market expectations. According to the previous day's script—Meta's guidance missed expectations and its stock dropped—Amazon should have fallen, right? Instead, after-hours trading saw a surge of over 9%. The same "thunder" produced completely different results. Why? Because the market finally distinguishes between two types of money burning: Meta's money burning—Metaverse, large models, pie-in-the-sky. Returns are unclear, and no one knows when profits will come back. Amazon's money burning—AWS data centers, AI computing power. Burning money while earning, the more they burn, the more they earn. Look at this set of data: AWS revenue growth is 37%. The market previously expected only 31%. AWS operating profit is 16.6 billion, up 64% year-over-year. Profit margin rose from 32.9% to 39.4%. Profit growth of 64% outpaces revenue growth of 37%. In plain language: AI computing power is not only selling more but also selling at a premium. Enterprises are willing to pay a premium for AWS computing power because of shortages. Amazon CEO Jassy's exact words: "Even with full-year capital expenditure at 220 billion, computing power supply in 2026 still cannot fully meet customer demand. There will also be a gap in 2027. By 2028, a large number of orders have already been locked in advance." AWS has 496 billion in orders on hand. AI business annualized revenue exceeds 25 billion, with triple-digit year-over-year growth. What is a barrier? This is a barrier. How does traditional accounting view 220 billion? Depreciation burden. Cash flow pressure. Decline in shareholder returns. But how does the market view it now? This is a defensive moat. You can build data centers, others can too. But you build first, lock in customers first, and secure 496 billion in contracts first. By the time others finish building, customers have already been running AI training in your ecosystem for two years—the migration cost is prohibitively high. The 220 billion is not buying servers; it's buying the right to levy the "computing power tax" for the next decade. What does this mean for Crypto? It means the "fat protocol" logic still holds in the AI era. Amazon is the largest "L1"—it earns native tokens (dollars) by selling computing power. AWS's "tokenomics" is extremely simple: you pay, I give you computing power. The greater the demand, the stronger the pricing power. Compare this to many Crypto AI projects—burning VC money on inference, collecting no fees, unable to retain users, with zero moat. Wall Street's current pricing logic is straightforward: Only "computing power L1s" that generate positive cash flow deserve a premium. AWS proves that AI computing power demand is real, sustainable, and highly premium. Those Crypto projects riding the AI concept but collecting zero fees—when the tide goes out, it's obvious who's swimming naked. While you are fixated on K-line charts debating whether to bottom-fish, Amazon spent 220 billion buying the future. While you study the next 100x coin, AWS earned 16.6 billion in one quarter. Wall Street used to fear burning money; now it fears you not burning money. Because not burning money means you are already out of the AI game. Burning money is not scary; burning money and not earning it back is scary. Amazon burned money and earned it back. It's that simple. $MSFT $META $AMZN #财报观察员：亚马逊指引不及预期，股价却反涨9%