跟 Erik Voorhees 相关的钱包,6 天买了 84,837 枚 $ETH ,均价 $1,931 ,一共 $1.54 亿。今天又加了 $2,367 万,账上还剩 $1,296 万稳定币。人名、成本、余粮都在。
The Erik Voorhees-linked wallets bought another 13,134 $ETH ($23.67M)!
Over the past 6 days, these wallets have bought a total of 84,837 $ETH ($154M) at an average price of $1,931.
They still have $12.96M in stablecoins left to buy more.
teapawt.org/lookonchain/status/212…
20
7 月就在系统里的 bug,11 月一次日常流量改道,超过 80% 的 Outlook 服务器直接无响应。找出问题花了大约 30 小时。 $MSFT
$MSFT - MICROSOFT REVEALS MONTHS-OLD BUG BEHIND GLOBAL OUTAGE
Microsoft says a software bug undetected since July triggered last week's worldwide disruption affecting Outlook, Teams, Copilot and other services.
The failure began on November 11, when routine traffic rerouting overloaded servers and caused cascading system failures.
At the outage's peak, more than 80% of Outlook servers were unresponsive.
Microsoft took approximately 30 hours to identify the broader problem and begin deploying a fix.
The incident raises fresh concerns over cloud infrastructure resilience and enterprise dependence on Microsoft.
22
周一 $SPY 跌约 1.9%,等权的 $RSP 只跌约 1.0%。市值加权指数跟成分股平均的 26 周相关性掉到约 0.4,至少 1990 年以来最低。这不是普跌,是少数大票在拖指数。 $SPY $QQQ $RSP
The S&P 500 has stopped tracking the average stock in it more than at any point since at least 1990.
The 26-week correlation between the S&P 500 and its equal-weight version has fallen to about 0.4, from close to 1.0 for most of the past 35 years, per data from Ophir and the NYSE.
“It’s not a stock market; it’s a market of a few stocks.”
29
欢迎帖没写价格。这笔收购全股票大约 $82 亿,李飞飞出任 AMD 执行副总裁兼首席科学家,直接向苏姿丰汇报。空间智能团队,整建制进了芯片公司。 $AMD
Welcome @drfeifei and @theworldlabs team to @AMD! Super excited for all we will accomplish together. 🚀
32
iOS 27.3 给合上的 iPhone Duo 加了前置摄像头解锁,折上也可以锁机。不是外屏完整 Face ID,是软件先补的一刀,安全限制写在说明里。硬件那档还没到,系统先让你打开手机。 $AAPL
Apple Adds Facial Unlock Option on iPhone Duo in iOS 27.3 macrumors.com/2026/11/16/iph…
56
SemiAnalysis 把全球 AI 加速器开支预期最多砍了 41%。2030 年从 $1.531 万亿下调到 $9230 亿,单年少掉 $6080 亿;2028 年从 $1.054 万亿砍到 $7580 亿。长期账先被改写了,今晚 $NVDA 再交近况。 $NVDA
Semiconductor research firm SemiAnalysis has sharply LOWERED its global AI accelerator capital expenditure forecasts, with cuts of up to 41%:
2027: $770B → $627B (-19%)
2028: $1.054T → $758B (-28%)
2029: $1.352T → $796B (-41%)
2030: $1.531T → $923B (-40%)
That's a $608 billion reduction in its 2030 forecast alone.
SemiAnalysis also cut its cumulative 2026-2030 accelerator silicon shipment area forecast by 17%, with Nvidia accounting for 66% of that reduction.
Its 2028 HBM shipment forecast was lowered from 9.0 to 6.4 exabytes, a 30% reduction, while projected 2030 AI accelerator power demand fell from 122.8 GW to 80.7 GW, down 34%.
The firm cited slowing revenue growth at frontier AI labs, tighter capital markets, and lower assumptions for future chip and memory pricing.
Importantly, even after these cuts, SemiAnalysis still expects annual AI accelerator spending to more than double from $392 billion in 2026 to $923 billion by 2030.
27
市场还在定价 2028 年资本开支见顶,郭明錤给出的 TPU 出货却是 2027-2029 年 8-9M、14-15M、16-18M,一年比一年多。这两年单 OpenAI 大约 1200 万颗、Anthropic 大约 1400 万颗。 $GOOGL $AVGO $NVDA
Google TPU 重點更新:OpenAI 與 Anthropic 新訂單推升出貨能見度至 2029 年、v10 產品線大幅調整 (改為 Quillfish)、聯發科與 Broadcom 為供應商中之領先關注焦點
我最新的產業調查與研究結論:
1. 訂單能見度與出貨規劃:
受惠於 OpenAI 與 Anthropic 強勁新訂單,Google TPU 出貨能見度已達 2029 年,優於 Nvidia GPU 與其他 AI ASIC(多數為 4Q27–2Q28)。
(1) 自 2028 年起,v9 將成 Google 自建算力的主力方案,v8 系列(8t/8i 或其設計變種)訂單可視為來自 OpenAI 與 Anthropic。
(2) 2028–2029 年合計需求數量:OpenAI 約 12M(聯發科設計的 8t),Anthropic 約 14M(Broadcom 設計的 8i),兩者需求均呈逐年成長趨勢。OpenAI 採買斷,Anthropic 則是租賃與買斷都有。
(3) 考量供應鏈限制,2027–2029 年較合理的 TPU 出貨估算分別為 8–9M、14–15M、16–18M。市場普遍預期 AI 資本支出於 2028 年見頂,但此規劃顯示 TPU 在 2029 年仍具出貨成長動能,關鍵在於 AI 服務龍頭 OpenAI 與 Anthropic 的直接訂單。
2. v10 產品線與供應鏈調整:
v10 由原先的 Icefish 與 Merope/Bladerunner 改為 Quillfish,預計 2029 年量產,設計服務訂單由聯發科取得。
(1) Icefish 改為 v11,預計 2030 年量產,設計服務訂單同由聯發科取得。Merope 能見度較低,供應商未定,也可能取消。
(2) 聯發科將是 v9(Humufish & Triggerfish)與 v10(Quillfish)的獨家供應商,2027–2029 年占整體 TPU 供應比重分別約 45–50%、50–55%、55–60%。TPU 對其營收與利潤的貢獻預計逐年增加。
(3) 自 v9 起新款 TPU 均採 Intel EMIB-T 先進封裝。考量初期量產能力,預期至 2029 年仍以 v8 系列為 TPU 出貨主力。v8 預期將是聯發科與 Broadcom 未來數年的 TPU 事業獲利成長的主要來源,因長期累積的學習曲線效益明顯,加上雙方負責關鍵零組件與材料採購。
3/ 市場關注度與供應商展望:
近期事件有助提升並維持 Google TPU 的市場關注度,關鍵設計服務商聯發科與 Broadcom 可望成為供應鏈首要焦點。
(1) Nvidia CEO 接班的不確定性,有利 Google 與客戶洽談 2028 年後的長期規劃。OpenAI 與 Anthropic 均為 TPU 關鍵客戶,其 IPO 也有助提升並維持 TPU 關注度,而聯發科與 Broadcom 是市場最容易聯想到的供應商。
(2) 聯發科無疑是未來數年 TPU 的最大贏家,身為 v9 後的獨家設計供應商,加上供應比重與 TPU 出貨成長同時逐年提升,均有利 TPU 事業營收規模擴大。高毛利率的 8t 系列出貨能見度達 2029 年,為關鍵利潤來源。
(3) 因 Broadcom 主導 8i 設計,故毛利率更高於 8t,有利該公司利潤成長。不過,需注意的是,Broadcom 未來數年的營運重點為非 TPU 的 XPU 與 AI 網路晶片訂單,預計最快 2027 年,非 TPU 營收將超越 TPU。
81
Qwen4-Max 上线。比上一代最多少用大约一半 token,内部评估单次任务成本降超 60%。Work360 平均每次从 $4.43 掉到 $1.67。输入 $1.5、输出 $4.5 每百万 token。用量和单价一起砍。 $BABA
📢 Meet Qwen4-Max, your new AI teammate for work, creativity, and discovery!
Qwen4-Max is built on a new MoE architecture with native vision and a 1M-token context window.
Highlights:
- Outperforms Qwen3.8-Max with up to ~50% fewer tokens, while average cost per task drops over 60% in our internal evaluation.
- Advances real-world productivity through greater proficiency with professional tools, stronger long-horizon execution, and higher-quality deliverables.
- Handles real tasks, including 3D modeling, game development, manufacturing, finance, law, healthcare, and frontier scientific discovery.
- $1.5/M tokens for input, $4.5/M tokens for output, and $0.075/M tokens for implicit caching — all lower than Qwen3.8-Max, with implicit caching down by 70%.
The API is live on QwenCloud. You can also try Qwen4-Max on Qwen Studio, QwenWork and Qoder.
Start building today! 🚀
- Blog: qwen.ai/blog?id=qwen4-max
- API: qwencloud.com/models/qwen4-m…
- Token Plan: qwencloud.com/pricing/token-…
More to come! The open weights of smaller Qwen4 models will be released. Stay tuned. 👀
38
黑石旗下 QTS 想私募 $8 亿债,规模砍到大约 $3 亿,少了六成。买家不够,它又不肯加收益率。2021 年 $100 亿买下的这家机房商,投资级债也卖不满了。 $BX
First Junk, Now Investment Grade: Blackstone's Own Data Center Giant Can't Sell $800 Million Of Debt zerohedge.com/markets/first-…
13
$AAPL 只比高点低 1%,$NVDA 已经回撤 33%,$BTC 接近腰斩,$MSTR 到了 −79%。同一张图里,回撤完全不是一个数量级。
% Below All-Time High
Apple: -1%
S&P 500: -4%
Microsoft: -7%
Amazon: -7%
Meta: -20%
AMD: -20%
Palantir: -21%
Tesla: -22%
Google: -28%
Gold: -31%
Nvidia: -33%
SpaceX: -38%
Bitcoin: -49%
Netflix: -54%
Ethereum: -64%
Strategy: -79%
Dogecoin: -92%
TrumpCoin: -99%
MelaniaCoin: -99%
1
1
68
新款 Apple TV 内存直接翻倍,但应用和游戏还是用不了多出来的那一半。硬件先给了,接口没打开。 $AAPL
The new Apple TV has twice as much RAM as the previous model, but apps and games can't use it 9to5mac.com/2026/11/15/new-a… by @bzamayo
43
第一代定价 $1999 的 Duo,下一代想补四样:外屏 Face ID、Action 按钮、长焦镜头,再做薄一点、轻一点。Face ID 排在最前,但明年秋季这代不一定赶得上,有可能要留到第三代。 $AAPL
Apple Aiming to Upgrade the iPhone Duo in Four Ways macrumors.com/2026/11/15/app…
30
稳定币总供应一个月缩了 $170 亿,是 2022 年以来最陡的一次。把稳定币换回现金,就是把买盘从 $BTC 和加密市场里抽走。下跌这边,流动性先走了一步。
Stablecoin supply is shrinking fast.
It fell by $17B over the last month, the steepest drop since 2022.
As holders swap stables to fiat they drain buying power from bitcoin:native and crypto.
For a bull market to happen, capital needs to enter, not leave the ecosystem.
1
38
英伟达七个最大客户,贡献了绝大多数在内部署的 GPU。其中六个已经在自研芯片:Google、Microsoft、Meta、Amazon、Anthropic、OpenAI,只剩 SpaceX 还没自己做。买得最多的那批,开始不想只当客户。 $NVDA
Six of Nvidia’s seven biggest customers are building their own chips.
“NVIDIA’s seven biggest customers account for the vast majority of GPUs being deployed: Google, Microsoft, Meta, Amazon, SpaceX, Anthropic and OpenAI. All of them, except SpaceX, are making their own chips.”
“If NVIDIA didn’t prop up CoreWeave and the other neoclouds, those industries just would not exist, and the hyperscalers would have full control.”
“Jensen’s the only executive in the semiconductor industry convincing supply chain companies to invest like they’ve never done before. From our benchmarking, the TPU is just as good as NVIDIA. But Google doesn’t have that DNA of wheeling and dealing.”
2
58
Solana treasury company Forward Industries(@FWDind), which holds 8.73M $SOL ($600M), deposited 1.78M $SOL ($123M) into #CoinbasePrime just 1 hour ago.
Is Forward Industries planning to dump $SOL?
arkm.com/explorer/entity/for…
32



























