Broadcom 最新一季 AI 半導體營收達 167 億美元,年增 221%、季增 54%,下一季預估進一步升到 217 億美元,年增 236%。 $AVGO 預期 2026 年 AI 半導體營收約 580 億美元,2027 年翻倍到 1,150 億美元,2028 年再翻倍到 2,300 億美元。Broadcom 表示,目前實際需求甚至高於這個數字。 成長動能主要來自六家 frontier AI 客戶加速部署客製化 XPU,其中四家已經非常明確。Google 已大量出貨 Ironwood TPU v7,並開始量產 TPU v8i,Broadcom 預計未來幾年每年對 Google 的 TPU 供應都會達數百億美元規模。 Anthropic 今年部署約 1GW Ironwood,2027 年預計再部署 5GW TPU v8i,2028 年還有額外 10GW 需求,管理層預期 Anthropic 可能從 2027 年起成為…

Channel
IEObserve 國際經濟觀察
@ieobserve
On this record: Topic · Growth · Engagement · Posts · Citations · Telegram's recommendations · Cite this entry
14,877subscribers
-32 since we began measuring on 6 August 2026
Risers and fallers across the register · movement among entries of 10,000–31,623.
Register entry
| Telegram ID | -1001394251478 |
|---|---|
| Type | Channel |
| Username | @ieobserve |
| Created | Between 1 June 2018 and 31 July 2021 — estimated from Telegram’s id allocation, not measured. How this range is calculated. |
| First recorded | 6 August 2026 |
| Last confirmed live | 18 September 2026 |
| Measurements held | 34 |
| Confirmed unchanged | 1 time, most recently 18 September 2026 |
| On Telegram | t.me/ieobserve |
Topic
Technology — a classification, not a measurement. An on-box language model (Qwen3.6-35B-A3B-FP8, prompt version 1) read this channel’s own recent posts on 10 September 2026 and assigned it the closest of 31 fixed categories, at 80% confidence. This is a model’s judgement about what the channel is likely to be about, not a fact this register measured the way a subscriber count or a view count is measured — it can be revised on a later pass, and it carries no weight anywhere else on this page. How this classification works, and why it has no browse page of its own yet.
Growth
| Measured (UTC) | Subscribers | Change |
|---|---|---|
| 18 Sept 2026, 06:21 | 14,877 | -5 |
| 16 Sept 2026, 03:18 | 14,882 | -9 |
| 14 Sept 2026, 08:57 | 14,891 | -3 |
| 12 Sept 2026, 15:38 | 14,894 | -6 |
| 10 Sept 2026, 15:36 | 14,900 | -1 |
| 7 Sept 2026, 10:56 | 14,901 | -6 |
| 4 Sept 2026, 13:41 | 14,907 | +5 |
| 3 Sept 2026, 01:37 | 14,902 | +2 |
| 1 Sept 2026, 23:26 | 14,900 | -2 |
| 31 Aug 2026, 21:06 | 14,902 | +2 |
| 30 Aug 2026, 20:25 | 14,900 | +5 |
| 29 Aug 2026, 17:03 | 14,895 | -2 |
| 28 Aug 2026, 20:03 | 14,897 | +3 |
| 27 Aug 2026, 21:57 | 14,894 | -3 |
| 26 Aug 2026, 23:26 | 14,897 | -2 |
| 26 Aug 2026, 00:53 | 14,899 | -1 |
| 23 Aug 2026, 08:47 | 14,900 | -4 |
| 21 Aug 2026, 17:45 | 14,904 | -2 |
| 20 Aug 2026, 18:23 | 14,906 | -1 |
| 19 Aug 2026, 19:55 | 14,907 | first reading |
Engagement
24 posts held, back to 23 April 2026 — the reader has not yet reached the start of this channel’s public history, so older posts may sit further back, unread. Read across 53 pages of Telegram’s post history, 20 posts per page.
- ERR · 30 days
- 20.9%
- avg views ÷ 14,877 subscribers
- Avg views / post
- 3,110
- 3 posts measured
- Reaction rate
- —
- this channel exposes no reaction counts
- Posts in window
- 3
- of 24 held
ERR is average views per post over the last 30 days divided by subscribers, the definition TGStat uses, so this figure is comparable with the one you will see elsewhere. It falls structurally as a channel grows: a high ERR on a small channel and a low one on a large channel describe reach mathematics, not quality. We publish the figure and the sample it came from and pass no verdict on it.
ER is defined industry-wide as (forwards + reactions + comments) ÷ views — note the denominator is views, not subscribers. Telegram’s public web preview carries views and reactions but not forward or comment counts, so the reaction rate above is the reactions term only and is therefore a floor: the true ER for this channel is higher by an amount we have not measured and will not estimate.
| Window | Rolling 30 days · latest post in window 2 September 2026 |
|---|---|
| Posts held | 24 (23 April 2026 – 2 September 2026) |
| Views total | 9,340 |
| Reactions total | — |
| Forwards / comments | not exposed by the public surface — not measured, not estimated |
| Readings taken | 3 Sept 2026, 11:04 UTC |
Views are a single reading per post, taken at the time above. A post published in the last day or two is still accumulating views, which pulls the 30-day average down slightly. That is a property of the standard definition rather than a fault in it, so we keep the definition rather than “correcting” the number into something nobody can reproduce.
Precision. Telegram publishes view counts on its public widget in short form — 8.12K, 3.7M — so any reading at or above 1,000 reaches us rounded to three significant figures, and only counts below 1,000 are exact. Averages and rates derived from them are shown to the same precision rather than to the unit: a figure like 3,701,250 would assert digits nobody measured.
Reaction counts are published per emoji and rounded the same way, so a total below 1,000 is exact and a larger one is a sum that may carry a rounded component from each emoji above 1,000. Because it is a sum, it does not look rounded — read a large reaction total as three significant figures per contributing emoji rather than as the figure it prints.
Recent posts
Marvell 對未來兩年的成長預期再度明顯上修,而且 AI 成長動能已經從光通訊 DSP,擴大到交換器、scale-up optics、CXL 與客製化晶片。 $MRVL #未來成長主要來自三條線 #AI網路與光互連。 800G 需求仍強,1.6T 正快速放量,51.2T Ethernet switch 今年營收預估超過倍增;更大的增量開始來自 scale-up。隨著 AI cluster 愈來愈大,銅線在距離、頻寬與功耗上逐漸碰到限制,NPO、CPO 等 scale-up optics 開始提前導入。管理層甚至表示,2027 財測多上修的 15 億美元中,connectivity,尤其 scale-up optics,是最重要的增量來源之一。 #客製化 Marvell 預估 2027 custom revenue 將超過倍增,2028 再明顯加速,成長不只來自 XPU,也來自 NIC、storage controlle…
以Nvidia這種體量,每季營收增速竟然還在加速是十分可怕。這季的電話財報會議很精彩,回應了大部分市場對Nvidia的質疑,包含循環融資、客戶ASIC競爭以及hyperscaler 資本支出到達極限的問題。 這季 $NVDA 營收962億美元,年增106%,Data Center年增117%,而且已經連續四季年增率加速。更誇張的是,NVIDIA罕見提前預告2027年營收仍可成長約70%,而且黃仁勳直接說,這不是需求上限,而是供應上限,客戶需求其實接近翻倍。 他對AI市場的判斷:AI已經開始做真正有生產力的工作,Token已經可以賺錢,「算力會直接轉化成營收」。下一波則是Agentic AI,一個Agent執行任務需要的算力可能是人類直接使用AI的15~100倍,推論需求還會再被放大。 市場擔心Hyperscaler資本支出見頂,但NVIDIA現在近一半Data Center營收已來自Neocloud、企業、主權AI與AI新…
AI 光通訊現在最大的問題,是產能跟不上需求。 兩大光通巨頭最新財報都在加速。Lumentum 最新一季營收 10.1 億美元、年增 109%,下一季指引中間值直接拉到 12.5 億美元、年增超過 130%,毛利率也升到 50.4%。 Coherent 最新一季營收 20.5 億美元、年增 34%,其中資料中心業務更年增 66%;公司預期 FY2027 結束前,單季營收將突破 30 億美元。 更重要的是,這波成長還在往上游傳。800G 繼續增加,1.6T 已經快速放量,EML、CW Laser、Pump Laser、高功率雷射到 InP 基板都開始吃緊。Coherent 表示主要瓶頸就是 InP;Lumentum 的高功率雷射需求增加速度快過擴產。 其他光通公司的財報也印證同一件事。AOI 表示 800G、1.6T 需求至少到 2027 年中都高於產能;AXT 的 InP 營收創歷史新高,backlog 已超過 1 億美…
SpaceX IPO後第一次季報,營收年增92%至78億,營業虧損大幅收窄,主要受惠於Starlink 作為現金奶牛輸血以及AI算力對外租賃變現。 老馬說太空算力要和Nvidia合作,Musk 在電話會議中表示,SpaceX正在開發名為 Starmind的AI衛星,其核心將是一套經過最佳化的NVIDIA Vera Rubin NVL72 AI運算系統。這不是只存在於概念階段的遠期計畫,SpaceX預計在2027年開始發射。 SpaceX並不打算直接照搬傳統資料中心機櫃,而是會「徹底簡化」一般NVL72機架設計。他認為這套重新設計的系統將具備兩項優勢: 第一,成本會更低。 第二,運算效率或整體系統效能會更高。 SpaceX甚至打算將相同設計同時部署在地面與軌道。Elon的原話是,既然已經為太空環境重新設計出一套更簡化的NVL72系統,就沒有理由不把它也放在地面資料中心使用。 更詳細的SpaceX財報與電話會議問答和展望…
Palantir Q2 營收年增加速至93%到19.35億優於預期的18億,營業利潤率也拉高到62%,EPS $0.41 優於預期的 $0.35。 Rule of 40 達到驚人的155%。 美國商業營收年增149%,AIP進入大型合約階段。美國政府業務年增90%,Maven平台持續擴張。 $PLTR 將全年營收預期從先前的 76.5-76.6 億上調至 81.5-81.6 億美元。 CEO Alex Karp 說強勁成長至少會持續18個月。Palantir Q2整體合約總價值為33.7億美元,年增49%。總剩餘合約價值達131億美元,年增83%、季增11%。剩餘履約義務為49億美元,年增103% 這份財報反映AIP已經開始從企業AI實驗工具,轉變成企業營運系統。企業不再只購買模型或Token,而是開始為資料整合、工作流程、自動化決策、模型治理與資訊安全付費。Palantir正好站在這個AI價值鏈重新分配的中心。 更詳細…
Amazon也上修了2026資本支出從2000億上修到2200億,主要也是記憶體漲價因素。Q2的自由現金流也轉負。AWS的營收成長加速至年增37%,營業利潤年增64%。 #AWS 的AI業務年化營收已超過250億美元,維持三位數成長;自研晶片業務的年化營收也超過250億美元,同樣維持三位數成長。CEO Andy Jassy特別強調,AI並沒有排擠AWS原本的雲端業務,反而正在拉動傳統運算、儲存與資料庫需求。 #資本支出 即使投入2,200億美元,Amazon仍預期2026年無法滿足所有客戶需求,2027年也可能維持供不應求。2027年新增的大部分容量已被預訂,甚至2028年也已有相當規模的客戶承諾。CEO說亞馬遜的支出狂潮短期內不太可能放緩。 Amazon仍維持原先目標:到2027年底,AWS可用電力容量將達到2025年底的兩倍。 #Amazon預估Q3 營收介於1,970億至2,020億美元,年增9%至12%;營業利益…
Tesla同樣也是自由現金流轉負,馬斯克比較猛的是他很多投大錢的未來project離大規模變現其實頗遠,很多大餅也還沒證明市場的可行性 但老馬沒錢的時候都敢豪賭了,現在都當首富了當然是要玩更大。從人形機器人到Terafab野心都是極大,馬斯克時程雖遲但到的承諾其實很多人也都相信 至少美光願意給面子用比較低的價格給他保證供貨 詳細財報電話會議說什麼可以看這篇 https://www.patreon.com/ieo/posts/164594526
Google 的資本支出真的換錢出來了,再說AI是泡沫的人真的是瞎了眼。這次EPS暴增具有迷惑性因為認列了SpaceX的投資收益。Google的EPS也會受到SpaceX的暴漲暴跌影響。 GCP年增又加速到82%,營業利潤則年增211%。Cloud 未履行合約金額增加超過 500 億美元,達到 5,140 億美元。公司預計,超過一半將在未來 24 個月內認列為營收。這個 backlog 並非全數來自 TPU 硬體合約。公司表示,大部分仍是由廣泛客戶組成的典型 GCP 長期合約。不過,TPU 系統銷售也已納入 backlog,並將在設備交付時認列收入。 與此同時又再大打市場的臉,體量非常龐大的搜尋業務非但沒有因為AI掛掉,還繼續有17%的年增,因為有了更多以前不會搜的長尾搜尋,同時AI也幫助可變現搜尋的市場,同時改善 AI 搜尋的使用量、廣告價值與推論成本。 當然,因為資本支出龐大,現金流的壓力也不容小覷。Google預估2…
Netflix Q2營收125.6億美元,年增13.4%,營業利益率33.4%,但營收成長從2025年16%至17%r降至13%上下,UCAN(美國與加拿大)成長更降至10%,Q3指引隱含進一步減速至11.7%,Q3營收128.6億與EPS 0.82美元的指引雙雙略低於分析師預估,股價盤後一度重挫9%。 GenAI已在約300部作品的製作流程中落地,管理層首次給出具體的成本效益數據。Sarandos給出的關鍵案例是紀錄片《The American Experiment》,其中17分鐘的AI強化畫面以傳統方法一半的成本、兩倍的速度完成。許多這類鏡頭在沒有GenAI的情況下,劇組根本負擔不起,只能捨棄。 Netflix的AI布局包括收購的Interpositive、旗下的Eyeline工作室與動畫實驗室,三者協同推進從概念、預視覺化到後製交付的全流程。對於成本節省是否會壓縮200億美元的現金內容預算,Sarandos的答案是否定…
台積電Q2營收年增34%至402億,毛利率66.7%,淨利年增77.4%。Q3營收指引年增37%,全年美元營收成長率從「超過30%」上修至「略高於40%」,資本支出年內第三度上調至600億至640億美元,並宣布亞利桑那追加1000億美元投資 #營運結構 HPC占比衝上66%,智慧型手機降至22%,AI已是絕對主引擎。2奈米甫量產即貢獻3%,3奈米占30%,先進製程合計占77%。2026年股利提高至每股24元,年增33% 。 #毛利率的代價 Q3毛利率指引65%至67%,2奈米陡峭爬坡將在下半年稀釋3至4個百分點,海外廠再稀釋2%至3%。這是以短期毛利率換取長期節點主導權的階段。 魏哲家對定價哲學的表態:「我們是合作夥伴,我的客戶必須成功,我不會突然把價格調高4倍、5倍。」並笑稱羨慕記憶體同業86%的毛利率,「68%的話我就很滿意了」。 #AI是一個新產業 魏哲家:「從今天起一路到2029年、2030年,需求都非常強勁。我相…
昨天基本上是看阿祖的Meta表演,全部都是他的消息,先是再投資加拿大92億約1GW的算力,是Meta最大規模的海外數據中心。接著說目標2027年算力翻倍擴大至14GW,並加速自研ASIC晶片推出。 最後阿祖自己在彭博專訪中提到,Meta 需要盡可能多的運算能力,但在一個缺乏運作和開發人工智慧產品所需資源的市場中,他也在考慮,算力資源所獲得的報價非常高,在某些情況下,出租或考慮這類交易可能比自己內部使用更有意義。 「我認識的業內人士沒有誰覺得算力過剩」 基本上就是在說,Meta算力也沒有過剩,他想要建制越多越好,只是當算力資源實在很賺錢的時候,他不排除先拿出去賣。 「我認為SpaceX的模式很有意思,它只做短期交易,但價格很高。所以我們會收到各種各樣的報價,我們會評估它們,看看哪些是合適的。」 果然,抄襲王阿祖絕不會讓人失望
Showing the 12 most recent of 24 posts we hold for @ieobserve. View and reaction counts are the latest single reading for each post, not a live figure, and a recent post is still accumulating both. A view count marked ≈ was rounded by Telegram before we ever saw it — t.me prints views in full below 1,000 and to three significant figures above, so ≈1,200,000 means somewhere between 1,150,000 and 1,249,999. Unmarked counts are exact. Text is reproduced from the public post preview and truncated for length.
Forward network
Republished by
Channels on the register that have forwarded this channel's posts into their own feed.
Built only from forwarded posts we have actually read, on both sides. Coverage is early and deliberately incomplete: a missing link means we have not read the post that would prove it, never that the relationship does not exist. Counts are distinct forwarded posts observed, so they only ever go up as we read more.
Appears in Telegram’s recommendations for other channels
The reverse of the list above, and a different kind of signal. This does not require this channel to have ever been asked about directly — each row below is a channel we DID ask Telegram about, whose Telegram-generated list happened to include this one. A channel can appear here with an empty list above it, because being named by someone else’s query is independent of having been queried itself.
@Gooaye · 105,832
Telegram ranks this channel #6 of 80 here — alongside 79 others — read 30 August 2026
@harryt30tybjb · 32,708
Telegram ranks this channel #48 of 73 here — alongside 72 others — read 4 September 2026
@johnstock888 · 49,716
Telegram ranks this channel #48 of 70 here — alongside 69 others — read 30 August 2026
@blocktemponews · 39,179
Telegram ranks this channel #61 of 87 here — alongside 86 others — read 30 August 2026
This channel appears in 4 seed channels' Telegram-generated recommendation lists in total. Each is Telegram’s list for THAT channel, not this one — see how this is measured.
Cite this entry
A live page changes as we take new readings, so a citation should name the measurement it is based on, not just the URL. The line below cites the subscriber count as measured 18 September 2026 — this entry's latest reading, not the date you are reading this.
“IEObserve 國際經濟觀察” (@ieobserve), 14,877 subscribers as measured 18 September 2026. Telegram Register, tgregister.com/channel/ieobserve.
Full measurement history, CC BY 4.0. Every reading this register holds for this entry, not just the latest one, as a dated, downloadable record: CSV · JSON. Free to use with attribution to tgregister.com. Each file carries its own generation timestamp, which is the figure to cite for exactly when the data was retrieved.