Unclestocknotes

U.S. stock market influencer | Financial background | Global traveler | Sharing insights on markets, trade, and economics | Swim, basketball, golf enthusiast.

[Unclestocknotes @Unclestocknotes]

免責聲明更新：
大叔美股筆記歡迎各位投資同好共襄盛舉，為維護個人權益，在瀏覽各個投資筆記前，請先閱覽以下聲明，閱覽者將視為同意本聲明條款：

1. 大叔美股筆記對於所發表之投資筆記、分析數據及參考資料內容，其立論依據皆係合法引用自網路上公開資訊以及合法管道取得之資料，不涉及任何抄襲等侵害智慧財產權之行為;對於閱覽者將大叔美股筆記所發表之投資筆記，在未經同意下，從事行銷推廣、背書代言、營利等商業相關活動，無論是否獲利，大叔美股筆記將採取法律行動追究其責。
2. 大叔美股筆記所發表之投資筆記、分析數據及參考資料內容，不得作為任何交易之依據，閱覽者依筆記或數據資料內容進行任何投資行為所產生之風險及盈虧，未經自行了解判斷，而導致交易損失風險，應自行負擔，大叔美股筆記不對使用者之投資決策負任何責任。
3. 大叔美股筆記所發表之投資筆記對於所提及之其他個人、公司或組織，大叔美股筆記均不對其交易安全、商品及服務內容負任何擔保責任，其間之所有交易、擔保、賠償責任及售後服務皆係存在於閱覽者與商品或勞務提供者之間，閱覽者如因商品或服務提供者發生任何變故或糾紛，大叔美股筆記不負任何法律責任。
4. 大叔美股筆記中，涉及到其他個人、公司或組織之網站連結，對被連結之網站或其所提供之任何產品、服務或資訊，大叔美股筆記不擔保其正確性、真實性、即時性、及完整性。
5. 閱覽者同意由大叔美股筆記保留修改本聲明之權利，如有變更，將公佈於大叔美股筆記社群平台上，不另行通知。

Disclaimer Updates:
Uncle Stock Notes welcomes all investment enthusiasts to participate. To safeguard your rights and interests, please read the following disclaimer carefully before reviewing any investment notes. By accessing or reading the content, you are deemed to have agreed to the terms set forth herein:

1. All investment notes, analytical data, and reference materials published by Uncle Stock Notes are based on legally sourced information, including publicly available online data and materials obtained through lawful channels. No infringement of intellectual property rights, including plagiarism, is involved. Any unauthorized use of Uncle Stock Notes’ content for marketing, promotion, endorsement, commercial gain, or related activities—whether for profit or not—without prior consent is strictly prohibited. Uncle Stock Notes reserves the right to take legal action against any such violations.

2. The investment notes, analytical data, and reference materials provided by Uncle Stock Notes are for informational purposes only and shall not be construed as a basis for any trading decisions. Any investment actions taken by readers based on such content are undertaken at their own risk. Readers are solely responsible for any profits or losses incurred, especially where decisions are made without independent evaluation and judgment. Uncle Stock Notes assumes no responsibility or liability for any investment decisions made by users.

3. With respect to any individuals, companies, or organizations mentioned in the investment notes, Uncle Stock Notes makes no representations or warranties regarding the security of transactions or the quality of any products or services provided. All transactions, guarantees, compensation, and after-sales services are solely between the reader and the respective provider. Uncle Stock Notes shall bear no legal liability for any disputes, losses, or issues arising between readers and such third parties.

4. Any external links to websites of individuals, companies, or organizations provided within Uncle Stock Notes are for reference only. Uncle Stock Notes makes no guarantees regarding the accuracy, authenticity, timeliness, or completeness of the content, products, or services offered on such linked websites.

5. Uncle Stock Notes reserves the right to amend this disclaimer at any time. Any changes will be published on its official platforms without prior notice.

[Unclestocknotes @Unclestocknotes]

@RYANHINGSHING @LIWEI_TWCapital @pcbanalysis @dnystedt @Aaronwei3n 你做到值得肯定👍

[Unclestocknotes @Unclestocknotes]

@photonjaeger @LIWEI_TWCapital @pcbanalysis @dnystedt @RYANHINGSHING 非常贊成，他們都很棒！從他們身上看到不同的觀點

[Unclestocknotes @Unclestocknotes]

SpaceX has taken another step toward its own mobile network. A few numbers need context.

The FCC accepted its application to acquire Grain’s 800 MHz spectrum portfolio: 546 paired licenses covering nearly the entire US population. That describes the licenses’ geographic reach, not an already-built Starlink Mobile network. Acceptance also does not mean regulatory approval.

This deal adds up to roughly 14 MHz. The “89 MHz” figure refers to SpaceX’s combined attributable spectrum holdings in relevant markets after closing—not 89 MHz of newly acquired 800 MHz spectrum. Low-band airwaves help with geographic and indoor coverage; capacity and speeds still depend on bandwidth, site density and network deployment.

Wireless service should keep evolving. But spectrum is one step in a costly, multiyear build. Watch $AMT, $CCI and $SBAC for actual signed leases; watch incumbent carriers for pricing and churn pressure. FCC approval, binding contracts and commercial milestones will tell us more than the excitement around the announcement. Then decide what that future is worth.

[Unclestocknotes @Unclestocknotes]

SpaceX朝自建行動網路又走一步，但先把幾個數字看清楚。
1. FCC已受理SpaceX收購Grain旗下800 MHz頻譜的申請，涉及546張成對執照，涵蓋近全美人口。但這是頻譜執照的地理範圍，還不是Starlink Mobile已建好的服務覆蓋；受理申請，也不等於核准交易。
2. 這筆交易取得的是最高約14 MHz頻寬。「89 MHz」指交易完成後，相關地區合計可歸屬SpaceX的頻譜，不能全算成這次買進的800 MHz。低頻有利於廣域與室內覆蓋，但承載多���用戶、速度能有多快，還要靠頻寬、基地台密度與網路建設。

我樂觀看待新競爭者，電信服務也該繼續進化。但頻譜到手後，還有建網、資金與執行的長路。前提是SpaceX真的簽下租約；三大電信的壓力，則要看未來定價與客戶流失。先看FCC核准、正式合約與商用時程，再決定願意為這個未來付多少錢。

[Unclestocknotes @Unclestocknotes]

#NVDA 輝達跌下神壇？黃仁勳風暴後，$NVDA 是機會還是危機？
輝達從高點跌了三成，約三分之二的跌幅發生在他承認出軌之前。
對大叔而且，人的問��是折價，公司的問題是重估，需求的問題才是危機。
所以我們現在走到哪一層？
大叔的新文章：我把估值、分批、避險與離場條件，加上 11/17 財報夜的五個檢查數字都分享給各位。

https://t.co/zbJq96RftV

[Unclestocknotes @Unclestocknotes]

@JohnFrontier_Z 沒錯，我跟你看法一樣，但目前為止甚麼都沒發生，就算發生應該是針對那些還在審核，但那些在德州州長就在8月喊停了。

[Unclestocknotes @Unclestocknotes]

民主黨翻轉德州、喬治亞、俄亥俄、愛荷華與內華達五州州長席次，依這份統計，合計涵蓋約37%的美國規劃中資料中心容量，德州占22.8%。德州、喬治亞與俄亥俄候任州長提出暫停大型資料中心的主張，地方公投也普遍支持限制。
Vistra週五交出的財報，第三季調整後EBITDA為23.78億美元，高於去年15.81億美元；EPS 3.01美元，優於預期2.71美元。公司上調2026年EBITDA指引中位數，區間為71–75億美元，並給出2027年77–85億美元展望，追加15億美元庫藏股，股價反彈約10%。2027年與2028年發電部位分別已有98%、78%避險，支撐近期能見度；但避險比例不等於資料中心長約覆蓋率，未簽約的未來需求仍有變數。
新建容量若受阻，既有、已通電且位置合適的容量可能更稀缺，因此 $VST、$CEG 值得與高槓桿開發商分開評估。$MSFT、$AMZN、$GOOGL 的既有電力與據點也有相對優勢，但建設放慢同樣可能延後雲端收入，不能直接推導為投資報酬改善。$FSLR、$NEE 則受政策預期支持，能否轉成訂單還要看具體措施。德州此前已有電網接入暫停安排，新增政治風險應與既有限制分開計算。

同樣叫「AI電力股」，已發電、已避險的資產，與借錢等許可的工地，承受的壓力差很多。接下來看候任政府如何界定新案、已核准案與施工中專案，也要看居民電費保護規則如何分攤電網成本。Vistra的財報提供了基本面支撐，但往後能賺多少，仍要逐筆看合約、避險到期與新增負載，不能只靠「電力稀缺」四個字。

[Unclestocknotes @Unclestocknotes]

@jimmyhuli 笑死

[Unclestocknotes @Unclestocknotes]

@chenyounan5 真的

[Unclestocknotes @Unclestocknotes]

我肚子好痛，這句話深得我心。
先讓我緩一下，這後勁太強。

[Unclestocknotes @Unclestocknotes]

@EeveeEeveeEevee 笑死

[Unclestocknotes @Unclestocknotes]

Google expanded its eighth-generation TPU lineup with two systems. TPU 8tx pairs with fifth-generation AMD EPYC hosts for hardware-level confidential computing, protecting data and model weights during processing. Google claims latency can be halved versus TPU 8t. TPU 8i Max combines two 8i dies in one package, doubling compute, SRAM and HBM bandwidth within the existing rack footprint. These are vendor specifications and performance claims; results will depend on workloads.

The investment signal is increasing specialization across confidential inference, long context and agent workloads. Custom accelerators can win deployments by reducing the cost of useful inference, provided software compatibility, migration costs and reliable supply support adoption. The announcement supports an evolving spending mix without establishing that total AI capex must rise. Doubling HBM bandwidth also does not automatically mean doubling memory capacity or procurement spending.

$AMD’s EPYC attachment is confirmed by Google. $AVGO merits tracking through its TPU design partnership, alongside $TSM, advanced packaging and HBM suppliers. However, this announcement does not provide a complete supplier list, allocations or order values. Benefits for SK hynix, Samsung and $MU depend on qualification, memory configurations and shipment volumes. $CLS and Quanta should likewise remain potential supply-chain exposures rather than confirmed suppliers to these systems.
Attractive specifications still need availability dates, cloud pricing and customer benchmarks before investors can assess the economics. Adoption depends on how much customers save after accounting for migration and development work. Memory and packaging demand require a system-level assessment, while conventional DRAM forecasts cannot be treated as HBM contract-price guidance. Keep following deployments, margins and cash conversion.

[Unclestocknotes @Unclestocknotes]

Google擴充第八代TPU產品線，推出兩款新系統：TPU 8tx搭配AMD第五代EPYC主機處理器，主打硬體層級的機密運算，保護推論過程中的資料與模型權重；Google宣稱，相較TPU 8t，延遲最高可減半。TPU 8i Max則把兩顆8i晶粒放進同一封裝，讓運算、SRAM與HBM頻寬加倍，維持既有機櫃占地。這些是Google公布的規格與效能主張，實際表現仍要看工作負載。

投資上值得注意的是，AI加速器正針對機密資料、長上下文與代理工作負載做更細的分工。客製晶片若能降低每次有效推論的成本，就有機會爭取更多部署，但還需要軟體相容性、遷移成本與穩定供貨配合。這次發表支持AI支出結構持續改變的方向，尚不足以證明整體資本支出一定增加，也不能��HBM頻寬加倍直接當成記憶體容量或採購金額翻倍。

$AMD 的EPYC搭配已由Google確認。$AVGO 可作為TPU設計合作的追蹤方向；$TSM 與先進封裝、HBM供應鏈也值得留意，但這次公告沒有完整揭露供應商、份額與訂單金額。SK hynix、Samsung、$MU 的實際受益，取決於認證、搭載規格與出貨量；$CLS、廣達等機櫃業者，也不能直接視為這兩款系統的已確認供應商。
新規格很吸引人，接下來要等正式供應時間、雲端價格與客戶實測，才能算清楚經濟效益。TPU能否擴大採用，關鍵在於客戶搬過去之後，成本能省多少、開發負擔增加多少。同時HBM與封裝需求要看整體系統配置；一般DRAM報價預測，也不能直接套成HBM合約漲幅。最後仍要回到部署量、毛利與現金流。

[Unclestocknotes @Unclestocknotes]

The SOX plunged 12.25% to 9,011.14 on Nov. 5, its largest daily decline since March 16, 2020. Micron and Lam Research lost roughly 17%, versus Nvidia’s 9.3% decline, putting memory and equipment under greater pressure. Friday’s roughly 4.2% semiconductor rebound offered relief, but the Nasdaq still lost 3.4% for its fourth consecutive weekly decline. Korean and Japanese equities each fell about 6% Friday. More trading sessions are needed to establish whether selling pressure is subsiding.

Nvidia also faces additional governance uncertainty. On Nov. 8, its board announced an investigation overseen by independent directors and outside counsel, with Jensen Huang’s cooperation. The board said it currently had no reason to believe national security or company secrets were compromised and reaffirmed confidence in management. Reports involving foreign intelligence remain allegations rather than investigation findings. The roughly 5% premarket decline on Nov. 9 should also remain distinct from closing performance.

The sector’s weakness could reflect deployment delays, financing costs and crowded-position unwinding together; price moves alone cannot establish a demand downturn. $MSFT, $AMZN and $GOOGL have relative financial resilience, although delayed construction can defer cloud revenue. $AVGO offers custom-ASIC exposure but remains sensitive to the same major customers’ budgets. $TSM serves both GPU and ASIC architectures, reducing exposure to individual vendors’ market-share shifts. Changes in total spending and wafer-order timing still matter.

[Unclestocknotes @Unclestocknotes]

11月5日，費半重挫12.25%至9,011.14，創2020年3月16日以來最大單日跌幅。美光與科林研發跌約17%，Nvidia跌9.3%，記憶體與設備股承受更大壓力。週五費半反彈約4.2%，但Nasdaq全週仍跌3.4%，連續第四週收黑；韓國與日本股市週五也各跌約6%。反彈提供喘息，還需要更多交易日確認賣壓是否消化。

Nvidia又增加了公司治理的不確定性。董事會11月8日宣布，由獨立董事與外部律師進行調查，黃仁勳支持並配合。董事會同時表示，目前沒有理由相信國家安全或公司機密受到侵害，並重申對管理團隊的信心。涉及外國情報活動的報導仍不能當成調查結論；11月9日盤前約跌5%，也應與正式收盤表現分開看。

這波走勢可能同時反映AI建設時程、融資成本與擁擠部位減碼，單靠跌幅無法判定需求已經轉弱。$MSFT、$AMZN、$GOOGL 的資金與客戶基礎提供相對韌性，但建設延誤也會影響雲端收入。$AVGO 的客製ASIC布局提供不同架構的曝險，仍受同一批大型客戶預算影響。$TSM 橫跨GPU與ASIC供應鏈，能降低單一晶片廠商市占變化的風險；總投資規模與投片時程改變，照樣會影響需求。

[Unclestocknotes @Unclestocknotes]

October payrolls rose 165,000, well above roughly 65,000 expected. September was revised from 29,000 to 60,000, with combined August–September revisions adding 84,000. Unemployment fell from 4.2% to 4.1%, while average hourly earnings increased 0.4% monthly and 3.1% annually. The two-year Treasury yield rose about six basis points to 4.52%, and the ten-year finished near 4.86%. CME pricing put the probability of a December quarter-point hike around 77%, leaving further tightening as the favored outcome.

Tech now faces pressure from both AI infrastructure economics and monetary policy. Investors are reassessing financing and deployment timelines, while resilient employment gives the Fed less reason to move toward easing. Annual wage growth of 3.1% does not suggest a wage spiral, although the 0.4% monthly increase deserves attention. With oil around $90–100, inflation remains a constraint. October CPI is the next test: consensus calls for 0.3% monthly headline inflation and 0.2% core. A softer reading could ease hike expectations and valuation pressure.

Stronger hiring weakens the immediate case for a Fed rescue. Yet the 339,000 increase in involuntary part-time employment reminds us to examine job quality alongside headline growth. Watch Nov. 10 CPI, Nov. 13 PPI and the ten- and thirty-year Treasury auctions. For tech to stabilize, companies need to deliver growth while inflation and financing costs stop adding pressure.

[Unclestocknotes @Unclestocknotes]

美國10月非農新增16.5萬人，明顯高於市場預期約6.5萬��9月由2.9萬上修至6萬，8、9月合計上修8.4萬。失業率從4.2%降至4.1%，平均時薪月增0.4%、年增3.1%。數據公布後，2年期美債殖利率上升約6個基點至4.52%，10年期收在約4.86%。CME顯示，12月升息25個基點的機率升至約77%，市場仍偏向聯準會繼續緊縮。

AI基建的融資與建設時程正在被重新評估，就業韌性又讓聯準會少了急著轉向寬鬆的理由。薪資年增3.1%本身尚未顯示失控，但單月0.4%仍值得留意，加上油價維持約90–100美元，通膨壓力還沒退場。接下來的關鍵是10月CPI，市場預期整體月增0.3%、核心月增0.2%；若低於預期，才有機會替升息押注與科技股估值一起降溫。

就業比預期強，讓「經濟快撐不住、聯準會會來救」這個期待暫時少了支撐。不過，非自願兼職增加33.9萬人，也提醒我們就業人數與工作品質要分開看。科技股要站穩，除了公司交出成長，還需要通膨與資金成本別再往上添壓力。

[Unclestocknotes @Unclestocknotes]

Democrats flipped five governorships—Texas, Georgia, Ohio, Iowa and Nevada—covering roughly 37% of planned US data-center capacity under the cited estimate, with Texas alone at 22.8%. Incoming governors in Texas, Georgia and Ohio proposed pauses on large facilities, while local ballot results broadly favored restrictions.
Vistra’s Q3 adjusted EBITDA reached $2.378B versus $1.581B a year earlier, and EPS of $3.01 beat the $2.71 consensus. It raised the midpoint of its 2026 EBITDA outlook, with a $7.1–7.5B range, introduced 2027 guidance of $7.7–8.5B and added $1.5B to buybacks. Shares rebounded roughly 10%. Its generation book is 98% hedged for 2027 and 78% for 2028, supporting near-term visibility. Those hedges, however, are not equivalent to data-center supply contracts; unsigned future demand remains exposed.

Restrictions on new construction could increase the scarcity value of existing, energized capacity in suitable locations, supporting a different assessment of $VST and $CEG from leveraged developers. Existing sites and power arrangements also give $MSFT, $AMZN and $GOOGL relative advantages, although slower deployment can delay cloud revenue rather than automatically improve returns. $FSLR and $NEE benefit from policy expectations, with actual orders depending on implementation. Texas already had grid-connection restrictions, so incremental political risk should be assessed separately from existing constraints.

An operating, hedged power asset and a debt-funded project awaiting permits carry very different risks—even when both wear the “AI power” label. Watch how incoming administrations treat new, approved and already-under-construction projects, alongside rules allocating grid costs to large customers. Vistra’s results provide fundamental support, but future earnings still require a close look at contracts, hedge expirations and incremental load. Scarcity alone does not settle the valuation.

[Unclestocknotes @Unclestocknotes]

Grab Holdings $GRAB 2026 年第三季財報分析
交出上市以來最完整的成績單：
・營收年增 27%
・調整後 EBITDA 年增 73%，創新高
・金融服務首度轉盈，三個事業同時獲利
・全年指引今年第二次上修

但有三件事新聞稿沒有強調：
上修之後第四季其實是持平、不良率從 2.2% 回升到 2.6%、淨利只有 1,800 萬美元。
還有一件事發生在財報之後：Grab 要拿 7.5 億美元去蓋 AI 資料中心。
我用四關法走了一遍，過了兩關半。免費完整分析在這裡 👇
https://t.co/ZOSGqMt27U

[Unclestocknotes @Unclestocknotes]

Onto Innovation $ONTO 2026 年第三季財報分析
營收年增 88%，每股盈餘超過財測上緣，訂單排到 2028 年，股價��後再漲 9%。
過去，封裝只占一顆處理器價值的一成；現在超過三成。當一個封裝裡塞進好幾顆昂貴晶粒，任何一顆出錯就是整組報廢，於是「檢查」這件事忽然變得比「製造」更值錢。
Onto Innovation這一季，它的封裝業務單季成長 35%，台灣營收年增將近兩倍，客戶為了 2028 年的產能提前排隊下單。
https://t.co/eCKed2YITL

[Unclestocknotes @Unclestocknotes]

延遲，不是減少—但這對「經濟」和「股價」是兩回事 https://t.co/qrGdc70Ohf