Post

峰哥的交易日记
峰哥的交易日记
Show original
AI可能正在成为美联储继续加息的理由,而不是降息的理由。 美国10月CPI环比上涨0.4%,高于市场预期的0.3%;核心CPI环比上涨0.3%,已经连续第三个月高于预期。 纽约联储主席威廉姆斯的态度很明确:通胀仍然过高,美联储还有工作要做。 更值得警惕的是,鹰派声音不止一个。 古尔斯比提醒,AI基础设施建设可能正在加剧通胀压力;克利夫兰联储主席哈玛克认为,货币政策需要转向限制性立场;美联储理事沃勒甚至没有排除12月加息50个基点的可能。 这意味着什么? 市场原本担心的是通胀降不下来,现在还要担心一个新问题:AI投资本身,可能让通胀更难降下来。 为什么? 因为AI不是只存在于屏幕里的一个概念。 微软、谷歌、亚马逊、Meta这些巨头要建设AI数据中心,需要什么? 芯片、服务器、电力、土地、冷却设备,还有大量工程师和技术人员。 这些东西不是按一下按钮就能无限增加的。 数据中心可以快速宣布开工,但电网扩容需要时间,土地和设备供应需要时间,专业人才也不是一夜之间就能培养出来。 结果就是:钱先砸进去了,资源先被抢走了,成本先涨起来了。 而AI真正带来的生产率提升,可能还需要几年才能全面兑现。 这就是当前最尴尬的矛盾: AI的长期价值可能是降低成本,但AI建设的短期过程,却可能不断推高成本。 美联储看的是眼前的通胀,不是五年后的生产力。 它不能因为AI未来可能让经济更高效,就对今天持续高企的物价视而不见。 当然,这不意味着AI投资越多,美联储就必然加息越多。美联储官员内部仍有分歧,AI带来的供给扩张最终也可能缓解价格压力。 但至少现在,市场必须接受一个现实:AI繁荣不等于货币宽松。 再把这个逻辑放到加密市场。 过去几年,很多人习惯了这样的交易逻辑: 英伟达涨,科技股涨,风险偏好回升,比特币跟着涨。 但现在,这条链条可能正在出现裂缝。 如果AI资本开支持续推高电力、设备和融资成本,通胀迟迟不回落,美联储就可能维持更高利率。 利率越高,美元资产的吸引力越强;融资成本越高,市场愿意承担的风险就越低;资金越谨慎,比特币、以太坊这类高波动资产就越容易承压。 最新行情已经给出警告:比特币周五跌破7万美元,市场同时面对更鹰派的利率预期和科技股波动。 所以,这一轮真正值得警惕的,不只是CPI高了0.1%,而是通胀背后的结构正在变化。 以前是经济太热,美联储踩刹车;现在可能是AI建设太热,美联储也得踩刹车。 这对加密市场很残酷。 因为AI越繁荣,未必意味着市场流动性越充裕;科技巨头越舍得花钱,也未必意味着资金会流向比特币。 AI可以讲五年的故事,但市场只愿意为今天的流动性买单。 如果你还在用“科技股涨,比特币就该涨”的旧逻辑交易,可能会发现:AI正在改变的不只是产业,还有美联储对利率的判断。 $BTC $ETH $ZEC
峰哥的交易日记
峰哥的交易日记
AMD at $570, are you chasing it? Jensen Huang just caused some chaos, and funds rushed into AMD like crazy, surging 30% in a single week—but if you chase it today, it gave back gains pre-market. Is this wave really a relay race for AI chips, or just a pump by speculators riding NVDA's coattails to cut those chasing highs? First, the surface: explosive good news, but price hesitation begins. Up over 30% in the past week, pulling back hard from a low just above 400 to 570, reclaiming most of the ground lost since the October 658 high. But after hitting 570 today, it started giving back gains pre-market and in perpetual futures—a classic "profit-taking after strength" scenario. The candlesticks tell you: 50-day and 200-day moving averages are fully crossed above, MACD histogram turned positive, RSI at 57-58 neutral to slightly strong—the structure is intact, but short-term it’s definitely overheated. First thing: AMD didn’t get stronger; NVDA got into trouble. This rally owes thanks to Jensen Huang. Whenever there’s a stir in NVDA’s boardroom, funds immediately look for "alternatives," lifting AMD and Broadcom. AMD is being forcefully packaged as the "Plan B for AI chips." But be clear—AMD is still the chaser, not the leader. That said, chasers have their perks: NVDA dominates and takes the lion’s share, AMD just needs to sip the soup. At this $570 price, that means tens of points of upside potential. What’s rotation? When the leader sneezes, the second-in-command takes off. Second thing: The earnings report is solid, not just hype. Q3 earnings on November 3: Revenue $13.4B, +45% YoY Data center business $8.5B, +97% YoY—doubled Q4 guidance $17B, +66% YoY, still accelerating This isn’t "expectation," it’s already booked revenue. EPYC server CPUs are grabbing Intel’s share, Instinct MI400 series GPUs are taking leftover share from NVDA, Helios rack solutions are scaling up. BofA raised the target price to 700, consensus target 658. AMD’s story now: I have CPU market share, GPU cost-performance, and a full rack solution. This is an "end-to-end AI platform," not just selling a single chip. Third thing: Macro is dragging, the biggest risk. 10-year US Treasury yield at 5.12%—this number is a knife for high-valuation growth stocks. AMD’s beta is about 2.5; when macro tightens, it falls twice as hard as the market. Questions remain on the sustainability of AI capital spending, data center power approvals, and policy uncertainty post-midterm elections—these risks are still lurking. Fundamentals are the engine; macro is the brake. Both are pressed simultaneously now. Bull vs. bear, you decide: On the bullish side: Q3 data center +97% YoY, Q4 guidance accelerating NVDA turmoil keeps rotation logic alive Technicals above all moving averages, mid-term bullish structure Analysts collectively raising estimates, consensus Strong Buy Volume surge this week, high fund participation On the bearish side: 30% weekly surge, short-term seriously overheated Pre-market at 570 already giving back 10Y yield 5.12%, pressure on high valuations Beta 2.5, vulnerable to market shocks If NVDA rebounds quickly, rotation logic cools off Key levels: Resistance above: 590-600 (psychological + previous highs) → 620-640 → 658 (ultimate previous high) Support below: 550-560 (recent lows/gaps/moving averages) → 530 (strong support) → 500-520 Trading strategy: Short-term traders: Wait for a pullback to 550-560 to buy in batches, stop loss below 535-540. First target 590-600 to sell half, if it breaks and holds above 590-600, add more, stop loss trailing below 570. Chasers can’t hold through a single pullback. Swing traders: 550-560 is the main battlefield, build positions in batches, target 620-640, extreme target 658 previous high. Keep position size within half of your risk tolerance, avoid heavy overnight holdings. Leverage traders: Keep perpetual leverage within 5-10x, funding rates near neutral but monitor position changes. AMD’s volatility isn’t for 20x leverage. AMD now looks like Nvidia at the start of 2023— Everyone thought "AI is a bubble," but earnings kept proving them wrong, and the stock kept flipping higher. But those who chased NVDA at the top in 2023 were stuck for half a year before breaking even. The $570 level isn’t a no-buy, but it’s not a buy-with-eyes-closed either. $BTC $NVDA $AMD

Disclaimer: OKX Orbit content is provided for informational purposes only. Learn more

Replies

No comments yet. Be the first to reply!