Unclestocknotes retweeted
$ONDS getting in front of US Special Forces 🇺🇸
Insignito showcased its DUMBO acoustic drone detection system at SOCOM REPTILE 2026 in Mississippi.
Direct feedback from Navy SEALs, Army Rangers and discussions on maritime deployment and system integration.
🗣️ „…and strengthening our presence in the U.S. defense market.“
Unclestocknotes retweeted
Think it’s that time again folks..
Ask me any question about $ONDS and I’ll answer it to the best of my abilities🤝🏻
No such thing as a stupid question either, genuinely just want to help people where I can👊🏻
Unclestocknotes retweeted
【失竊案 😱】巴西聖保羅瓜魯柳斯國際機場(São Paulo/Guarulhos International Airport)發生嚴重高科技貨品失竊疑案。當地知名硬體代理經銷商 Oderço 證實,一批裝載於兩個大型卡板上的 ZOTAC GeForce RTX 5090 旗艦級顯示卡,在運抵機場後不翼而飛。經銷商估算,該批失竊貨品的採購成本高達 250 萬巴西雷亞爾(折合約 47 萬美元 / 港幣約 365 萬元 / 新台幣約 1,500 萬元),目前官方仍未掌握具體貨物去向。
據《VideoCardZ》報導,貨主為巴西經銷商 Oderço,其負責人 Oderço Matos 於 10 月 16 日正式發布影片聲明對外公開事件。Matos 指出,該公司營運至今已有 38 年歷史,並自 2008 年起長期從事電腦硬體進口業務,數十年來從未發生過進口貨品失竊事件。
videocardz.com/newz/two-pall…
Matos 表示,貨物抵達機場後隨即失去蹤影,公司雖多次向機場營運單位追查相關貨物的物流動向與交接細節,卻始終未獲任何合理解釋與正面回應。他更質疑,主管當局無法提供記錄貨物移動過程的閉路電視(CCTV)監控畫面,在求助無門且欠缺透明度的情況下,最終決定將整件事公諸於世,並直言:「兩個大型卡板的貨物,絕對不可能在國際機場內自行憑空消失。」
目前,瓜魯柳斯國際機場營運方尚未對巴西當地媒體的查詢作出回應,巴西聯邦警察(Federal Police)與聯邦稅務局(Federal Revenue Service)亦暫未正式發布公開調查通報。經銷商方面目前尚未公開貨物的確切到港日期與詳細貨運追蹤提單號碼,官方亦未將此案正式定性為內部盜竊抑或物流走私。
Unclestocknotes retweeted
$ONDS
👀 We have movement on Regna AI..
“The sovereign AI layer for defense”
“Brainchild of @PalantirTech & @OndasHoldings”
Location: Everywhere.
@regna_ai 👊🏻🔥
Unclestocknotes retweeted
#SemiVision has compiled an overview of key equipment suppliers for panel-level advanced packaging.
This supplier mapping is primarily based on our recent channel checks in Taiwan, covering major process steps from molding and die attach to RDL, CMP, debonding, thermal treatment, and inspection.
Advanced semiconductor packaging is creating significant business opportunities, and many new equipment suppliers are entering the market.
Please note that China-based suppliers have been excluded from this version, and the list is not intended to be exhaustive.
* Feel free to share, repost, or use this chart with credit to SemiVision. If I have missed any important suppliers, please let me know and help us improve the mapping.
Thank
Unclestocknotes retweeted
Any pursuit of superintelligence has to be grounded in the core principle that if the AI we build is not helping humanity and under human control, it's not worth pursuing.
We also need to accelerate and spread the benefits of AI, such that they are diffused broadly across countries, communities, and companies. This requires a frontier ecosystem in which both closed and open-source models can thrive.
And for firms, it’s imperative that they retain full control over their unique and tacit knowledge. Every organization should be able to build its own continuous learning loop/hill climbing machine, without becoming dependent on any one model provider, and have the ability to embed its own knowledge into models and weights they control.
So, in this context, we welcome the research, focus, and deliberate pacing needed to get alignment right as the design goal. We also welcome ideas like "embedded evaluators" and the broader efforts to develop the mechanisms to make this more than just talk.
The key is that this cannot be controlled by a handful of entities, but must have broad representation across the ecosystem, countries, and fields, including academia.
This is the approach we are taking: broad access and choice at every layer of the AI stack; enterprise control of learning loops and models; and the “Code of Conduct” that underlies our own first party MAI models that we’ll publish tomorrow for public consultation.
Korean media reported that Micron plans to add 60,000 wafers per month of HBM capacity by the end of this year, raising its total capacity to around 100,000 wafers per month.
This marks an aggressive expansion that nearly doubles last year's capacity level of 40,000 to 50,000 wafers per month within a short span of time.
With SK Hynix's HBM capacity expected to reach 200,000 wafers per month and Samsung's 250,000 wafers per month by the end of this year, Micron's expansion is the most aggressive in scale among the three companies.
The industry expects Micron's HBM4 12-High product mix to rise from 20 to 30 percent of its total HBM output in early this year to as much as 50 percent by the end of the year.
view.asiae.co.kr/article/202…
Unclestocknotes retweeted
從 1.6T 到量子網路解密半導體巨頭的 PIC 光子晶片終極戰役
Call Memo
20260901
重點摘要
1. 1.6T進入量產,200G/lane成主流。
2. CPO加速導入,主動對位需求延續至2028年。
3. PIC放量,InP、PD與對位設備恐成瓶頸。
論壇重點
1. 1.6T光互連可透過8條200Gbps通道或增加100Gbps通道數達成,主流仍採PAM4。現階段可插拔光模組、LPO、NPO及CPO並行,差異主要在DSP配置、光引擎與ASIC距離、功耗、成本及可維修性。
2. Broadcom Tomahawk 6已出貨,單晶片交換容量102.4Tbps,支援100G/200G SerDes及CPO;其TH6-Davisson CPO方案由16個6.4T光引擎組成,每通道200Gbps,並已著手開發每通道400Gbps的第四代CPO。
3. NVIDIA Vera Rubin平台導入Spectrum-X Ethernet Photonics,將CPO直接整合於交換器ASIC封裝,200Gbps SerDes方案已進入量產,主要訴求為降低功耗、提升可靠度及支援百萬顆GPU規模的AI叢集。
4. 聯電與SILITH已由新加坡12吋廠交付首批量產矽光子晶圓,1.6T平台由開發至量產準備歷時18個月,累計出貨逾800萬顆100G/200G每通道PIC;後續將開發400G每通道純矽光子及TFLN方案。
5. 2026至2028年PIC與FAU整合仍以主動對位為主,需奈米級定位、點膠固化及光學測試;被動對位可簡化組裝與重複測試,但量產良率、標準化及產能仍待建立。
6. 1.6T供應鏈涵蓋CW Laser、EML、VCSEL、PIC、DSP、PD、TIA、FAU、光纖及光模組;Coherent、Lumentum與Broadcom位於光源及光晶片環節,康寧則擴充光纖、光纜與高密度光連接產能。
7. 預期可插拔光模組仍可延續至2028至2029年;隨交換器頻寬、熱密度與傳輸距離要求提高,CPO將先由交換器端擴大,後續再延伸至XPU及HPC互連。
8. 1.6T後續將朝3.2T及6.4T發展,單通道速率由200G提升至400G。當可插拔模組面臨功耗與散熱限制,光引擎將更靠近ASIC,並逐步由2.5D封裝走向3D整合。
9. LPO移除DSP可降低光模組功耗與成本,但對SerDes訊號品質、溫度補償、線路距離及模組一致性要求更高;短距離互連將呈現LPO、NPO及CPO多種架構並存。
10. PIC放量後,短期瓶頸將轉向主動對位及光測設備、InP磊晶品質與外延產能、PD良率、TIA供應,以及EIC/PIC之間的微凸塊與3D整合製程。
11. 被動對位若能透過Glass Bridge、微透鏡及高密度光纖連接降低組裝與測試成本,可能自2029年開始滲透;惟估計真正形成大規模產能或須至2030年前後,此為產業情境推估。
12. Micro LED與Micro-VCSEL可能用於更短距離、高密度光互連,但光源波段、耦合方式及通訊標準尚未定案,現有顯示用Micro LED產能不必然可直接轉為資料通訊用途。
13. 光量子運算須依量子位元類型區分;傳統LPO難以直接適用於極低溫量子系統,未來可能需整合單光子光源、PIC及超導奈米線單光子偵測器(SNSPD)的專用量子光互連架構。
※Q&A
14. PIC開始放量後,最可能先出現供不應求的環節為何?
2026至2028年仍以主動對位為主,FAU貼合前需精密對位、點膠固化及光測,PIC亦可能需要雙面晶圓測試,因此奈米級主動對位、光耦合及光測設備最可能先形成缺口。
15. PIC通道數增加後,接收端可能出現哪些瓶頸?
通道數由8條往16條以上增加,PD與TIA需求同步放大;瓶頸主要在InP材料及外延產能、磊晶品質、TIA供應,以及EIC與PIC對貼所需的微凸塊與3D整合能力。
16. PD短期限制主要是產能、良率或客戶認證?
材料與製程良率是主要限制。不同材料間約4%的晶格失配及熱膨脹差異會增加缺陷,PD製程良率難以穩定突破50%,因此擴產前仍須先改善磊晶與異質整合良率。
17. Glass Bridge與被動對位何時可能明顯取代主動對位?
2026至2028年仍由主動對位主導,被動對位可能自2029年開始滲透;考量美國新廠、FAU、光纖與特殊玻璃產能建置至少需兩年,真正明顯影響主動對位需求可能落在2030年前後。
18. Micro LED光互連與台灣面板廠有何關聯?
若採外置光源,具巨量轉移能力的面板廠可能參與供應;若由晶圓代工廠整合內置光源,價值將偏向半導體製程。關鍵仍是光通訊標準未定,既有顯示產線能否符合波長、可靠度及耦合要求尚待確認。
19. 聯電的矽光子與TFLN布局進度如何?
聯電與SILITH的1.6T矽光子平台已進入量產,並共同開發400G每通道純矽光子;TFLN則與生態系夥伴持續開發。預期2Q27可能出現新進展,但合作對象與時程未獲官方完整揭露。
Unclestocknotes retweeted
Tesla has just shared the 4K version of the Cybercab's unboxed manufacturing process from last week's launch event.
The line is designed so that one Cybercab can roll off the production line every 10 seconds, with a long term goal of ~5 seconds.
Unclestocknotes retweeted
Geez… I can’t imagine how Nike could possibly get removed from the S&P 500 and its stock is down roughly 78% from its 2021 peak.🥴🫠
Unclestocknotes retweeted
Sachsen-Anhalt hat gesprochen: Die AfD ist mit 44,2% mehr als doppelt so stark wie die CDU und hat einen Regierungsauftrag erhalten. Gratulation an Ulrich Siegmund und die AfD Sachsen-Anhalt!
When TSA says Part 108 rule is forth coming.
Following publication of the final rule, TSA plans to host a series of informational meetings with industry and external stakeholders to inform TSA's development of options that UAS BVLOS operators could employ to meet TSA's security program requirements.
Requests to participate in the technical roundtables must be received by the individual listed in FOR FURTHER INFORMATION CONTACT section. October 19, 2026
TSA inserted itself into Part 108 with closed-door BVLOS security roundtables, published in the Federal Register today. The notice solicits representatives of BVLOS operators and industry associations to help draft model language for TSA-approved security programmes under the forthcoming rule. Attendance is capped at two people per organization, every participant must be cleared for Sensitive Security Information and sign an NDA, and requests are due 19 October 2026. The sessions are explicitly closed to the public.
Unclestocknotes retweeted
$AVAV AeroVironment secured a $464.8M U.S. Army contract to deliver dozens of LOCUST X3 high-energy laser systems over the next few years.
It marks the first U.S. production contract for directed-energy systems, moving the technology from prototype into scaled deployment.
Unclestocknotes retweeted
日月光投控先進封裝業務報捷,受惠Google張量處理器(TPU)量能放大,大舉追加先進封測訂單,英特爾加快推進EMIB先進封裝平台,同步擴大對日月光投控委外釋單,伴隨先前調高封測報價效益顯現,三大利多匯聚,日月光投控營運迎來爆發期。
#ASE
Unclestocknotes retweeted
Volume increase + Mix shift worth noting...
Autonomous interceptors are increasingly important.
Unclestocknotes retweeted
最近出了兩篇深度non consensus 報告,短期讓我重回Substack 前10 Rising in technology,一篇講台積電明年資本開支可能上看1000億美元(台積電很保守,應該會慢慢上修到1000億,不會一次到位),一篇講 Nvidia 在3-5年內遲早下海自己一條龍做 AI token foundry,因為以Nvidia 的1GW建廠成本來看,應該會比谷歌,亞馬遜還低一半,但買 Nvidia AI GPU/機櫃方案的客戶,就要多付3倍的成本才能建1GW算力,舉例而言,150億美元是 Nvidia 賣鏟子成本,600億美元是 Nvidia 鏟子及其他輔助工具總售價,人家谷歌,亞馬遜用400-450億美元鏟子及輔助工具去淘金,Nvidia 客戶要用600億美元鏟子及輔助工具去淘金,這些新興外溢雲客戶遲早要被市場淘汰,還不如 Nvidia 自己拿自己做的200億美元鏟子加輔助工具直接下海淘金,還比谷歌/亞馬遜便宜一半。你們覺得呢?
substack.com/@andrewhclu/not…
substack.com/@andrewhclu/not…






































