Post

2014年，选举结束后市场还在跌，又磨了70天才真正见底。 2018年，选举后40多天，BTC砸出周期低点。 2022年，几乎是选举那几天，底部同时出现。 Killa的推演是：低点就在选举前或选举附近，8万上方撑住，然后往上走。听起来很顺。但Killa自己给了第二个情景，而且他说得更谨慎：先冲高，甚至摸到94-95k，然后选举后去风险下跌。过去4次中期选举，有3次是选举后跌的。问题来了：这一次，市场在提前交易哪个剧本？ BTC没冲上85k，直接跌破82k。触发点不是美联储，不是ETF流出，是一条新闻：美国和以色列军方已制定计划，在中期选举结束后对伊朗再次发动打击。市场在提前计价一个选举之后才会发生的事。特朗普10月8日亲口说过，不会在中期选举前攻击伊朗。福克斯新闻的判断更直白：他不太可能在选举前恢复打击。翻译一下：选举前，战争被按了暂停键。选举后，暂停键松开。伊朗在故意拖延谈判，想在选举前削弱特朗普的政治地位。特朗普自己在社交媒体上说伊朗“竭尽全力在选举前抬高油价”。所有玩家都在等11月3日。而市场最讨厌的，就是“等”。所以“8万上方的周线底”，有一个致命的前提。 Killa的提前见底情景，建立在“低点形成→反弹→持有看涨”的链条上。但低点形成的前提，是卖压在那之前释放完。现在的卖压是什么？不是加密内部杠杆。BTC从85k跌到81k，伴随的是纳斯达克六月以来最大单日跌幅，AI资本支出质疑、半导体暴跌、油价五个月新高。加密在跟宏观的跌，而不是自己在跌。而宏观最大的那颗雷，被精准地安排在选举之后。 8万附近有反弹吗？大概率有。 80,000是主要整固区，多次提供买盘，双底形态的技术性反弹完全可能发生。但那个反弹的性质，决定了你该拿还是该走。如果选举前BTC在80k-82k磨底、缩量、然后往85k修复——这是“选举前维稳”的战术性修复，不是趋势反转。真正的定价权，在11月3日之后才交还给市场。美伊冲突的重启、12月加息预期（CME显示12月累计加息25bp概率62.2%，维持不变仅9.9%）、油价对通胀的二轮传导——所有这些，都在选举后的日历上排队。 Killa自己也提醒了：不能简单照搬“选举后必跌”，但如果冲高后开始走弱，就该去风险。换句话说：8万上方的底，可能是真的。但它大概率是一个“选举底”，不是“周期底”。别在选举前赌方向。 8万上方的反弹可以吃，但别把它当成“牛市重启”去加杠杆。Killa的量化背景让他可以灵活反手，你未必有那个手速。把85k当成试金石。如果BTC在选举前反复冲85k不过、然后回落，那是“经典去风险”剧本在提前彩排。反过来，如果80k守住、低点逐步抬高，再谈持有看涨。选举后的第一周，比选举前更重要。历史告诉你的底部，往往需要一次“事后确认”。2022年是选举期间见底，但真正让人拿住仓位的，是之后的走势，不是选举那一天的价格。历史不会简单重复，但会押韵。中期选举的“韵脚”是：底部往往在选举前出现，但真正的风险，在选举后释放。 $BTC $ETH $SOL

On October 22, Meta released its Q3 earnings report: revenue of $64.6 billion, a 26% year-over-year increase, beating expectations by $1.2 billion. Adjusted EPS was $7.40, crushing the expected $6.80. Then the stock price plunged 6.98%. Why? Because in the same quarter, Meta's free cash flow was -$2.12 billion. In the same period last year, this figure was a positive $10.625 billion. This is the first time since Meta's 2012 IPO that free cash flow turned negative. In one quarter, from "earning tens of billions in free cash flow annually" to "negative $2.1 billion." All it took was one capital expenditure increase. Meta raised its future spending commitment from $700 billion in July to $825 billion in one go. Q3 capital expenditure alone was $34.9 billion, directly exceeding operating cash flow of $32.8 billion. What if there’s not enough money? Issue debt. And this is what the bond market is truly pricing. Meta’s five-year CDS—simply put, the price of "buying default insurance on Meta’s debt"—soared to 119.3 basis points, the highest since the contract became actively traded. Meta’s bonds weakened noticeably relative to U.S. Treasuries, with the risk premium on bonds maturing in 2036 widening by 6 basis points. Stock investors are discussing the AI narrative. Bond investors are buying insurance. On the same day, CDS spreads for Microsoft, Oracle, and Broadcom also heated up. Oracle’s CDS spread was only 144 basis points at the start of the year but surged to 322 basis points in October. SpaceX’s CDS rose to a record 226 basis points. Trading volume tells the story even more clearly: in September, the weekly average trading volume of CDS linked to bonds of Oracle, SpaceX, Alphabet, Amazon, Meta, Broadcom, and Nvidia exceeded $9 billion. In July, this figure was less than $2 billion. The chief strategist at Société Générale said a blunt truth: "CDS have become a key barometer for investors on the financing and broader AI trade sentiment of hyperscale cloud providers. The market is increasingly focused on two things: CDS and free cash flow." Not valuation. Cash flow. The AI narrative is entering its second phase. The first phase was "Whoever buys the cards rises." Nvidia sells chips, cloud providers build computing power, and the stock market applauds capital expenditure. Burning money equals growth; investment equals confidence. The second phase is "Whoever borrows money to buy cards gets repriced." When capital expenditure starts to exceed operating cash flow, when future commitments jump from $700 billion to $825 billion, when issuing debt becomes a necessary means to keep burning cash—the bond market stops applauding. Since 2026, AI-related companies have issued an estimated $466 billion in debt. In the same period in 2025, it was only $101 billion. The pace of borrowing is more than four times last year’s. The CDS market is more honest than the stock market. Stocks can be propped up by narratives; bonds only recognize cash flow. When CDS hit record highs, the message is clear: bond investors are already pricing in the risks of AI infrastructure debt expansion. On October 23, Bitcoin fell below $81,000, dropping more than 4% in 24 hours. Crypto-related stocks fell in tandem: MSTR down 5%, BMNR down over 6%, COIN down 4.49%. On the surface, this was a risk-off reaction triggered by Middle East tensions and oil prices breaking $100. But the deeper logic is here: AI concept stocks and crypto concept stocks share a vulnerability—they both heavily rely on incremental financing to maintain valuations. AI infrastructure relies on issuing debt. The crypto market relies on ETF inflows, leverage, and new money. Both are "financing-driven" assets. When financing costs rise, when credit spreads widen, when the bond market starts saying "no"—the first assets to be sold are always those with the best liquidity and highest leverage. Bitcoin trades 24/7. It has no close, so it feels the chill first. Meta’s revenue is fine. Advertising is still growing, and Q4 guidance is still being raised. The long-term logic of AI is not bankrupt. But the "burn money at all costs" phase is being stopped by the bond market. The $825 billion spending commitment means Meta will have to keep borrowing from the market for years to come. And the CDS record highs mean borrowing costs are rising. This is a self-reinforcing cycle: the more money burned, the weaker the credit; the weaker the credit, the more expensive the financing; the more expensive the financing, the harder it is to burn money. Zuckerberg says these investments are necessary. Maybe he’s right. But the bond market doesn’t debate "necessity." It only gives you a price. The most fragile link in the AI bull market isn’t on the valuation sheet. It’s on the cash flow statement. It’s in the CDS quotes. It’s in that number that went from "positive $10.6 billion" to "negative $2.1 billion." Stock investors are still asking "How long can AI rise?" Bond investors are already asking "Who will pay back the money?" $BTC $META $MSFT