供应链分析
产业链结构、上下游与瓶颈环节研究 · 共 1413 条 · 滚到底自动加载更多
-
推测Alchip或QCOM将跟随MRVL和联发科趋势，ASIC市场整体扩大。
@balllvest 鉴于 $MRVL 和联发科(Mediatek) 的趋势走向，我猜测 Alchip（披露：我持有该股）或 $QCOM 可能是下一个。不过我认为应用专用集成电路(ASIC) 的市场份额对所有这些公司来说都会扩大。
英文原文
@balllvest With how the trends are going with $MRVL and Mediatek. My guess is Alchip (disclosure, I own this) or $QCOM might be next. I think the ASIC pie gets bigger for all of them though.
-
联发科与Ayar合作CPO，SIVE为其激光器供应商但偏上游。
@Matrix_B0SS 我今天才知道，根据公开文件，联发科(Mediatek)是共封装光学(CPO)领域Ayar的合作伙伴（不仅仅是投资者）。当然，众所周知，$SIVE 是 Ayar 的主要激光器供应商。但 Sivers 目前可能处于产业链的上游环节。 https://t.co/gs8e1z2rf1
英文原文
@Matrix_B0SS I actually learned today that Mediatek is partners with Ayar for CPO (not just investors) from public filings. And of course it's known that $SIVE is the primary laser supplier to Ayar. But Sivers is probably too far upstream for now. https://t.co/gs8e1z2rf1
-
英伟达投资联发科，整合ASIC生态并确立标准以对抗博通。
今天，$NVDA 宣布向联发科投资 35 亿美元。看来英伟达正通过“结网”融入更广泛的 ASIC（专用集成电路）生态系统，从 $MRVL 到联发科。并让未来的生态系统与英伟达标准保持一致。（也许也是为了抵御 $AVGO）
英文原文
Today, $NVDA has announced a $3.5B investment into Mediatek. It appears Nvidia is "networking" itself into the broader ASIC ecosystem, from $MRVL to MediaTek. And have future ecosystems aligned with Nvidia standards. (maybe to fend off $AVGO as well) https://t.co/okdGKS51U3
-
博主庆祝瓶颈论点获验证，硅光子普及促成多项新产能协议。
@AizuddinAl4653 这需要你自己估算！我不想给出市值(MC)估算，尤其是当人们可能会误解模型时。我只是在庆祝“瓶颈(chokepoint)”论点得到了外部评论的验证，这些评论提到了类似垄断的市场份额数据。而且，由于硅光子(silicon photonics)的大规模采用，这一瓶颈已经转化为可能带来10多项新长期协议(LTAs)/产能协议。
英文原文
@AizuddinAl4653 That’s for you to estimate yourself! I don’t want to give MC estimates especially if people misinterpret the modeling I’m just celebrating that the chokepoint thesis was validated by external commentary around monopoly-like market share figures. And since the chokepoint converted into likely 10+ new LTAs/capacity agreements because of siph mass adoption.
-
三星锁定70%产能至2031年，反驳内存泡沫论。
“内存是泡沫。”那么请解释一下这个，KOSPI看空者： 据报道，三星已通过长期协议(LTA)与$NVDA、$MSFT和$GOOGL等主要交易对手，锁定了其至2031年约70%的内存产能。 五年的需求可见性…… https://t.co/GUGWTkXkMc
英文原文
“Memory is a bubble.”Explain this then, KOSPI bears: Samsung has reportedly reserved ~70% of its memory production capacity through 2031 via LTAs with $NVDA, $MSFT, and $GOOGL as main counterparties. Five years of demand visibility… https://t.co/GUGWTkXkMc
-
Celestial若上市估值或达百亿，强调CPO资格认证周期价值而非短期亏损。
假设 Celestial（$MRVL 的收购标的）是上市公司。其估值在美国市场可能会是天价（我猜约 60-100 亿美元）。因为 2028 年营收预测为 5 亿美元（Q4 年化运行率），2029 年营收预测为 10 亿美元。而且 Celestial 是超大规模数据中心共封装光学(CPO) 计划的核心部分。Celestial 的 2026 年 Q2 财报会是：- 营收接近 0（Marvell 称“自收购以来的营收和利润并不重大”）- 每季度亏损 1250 万美元（每年 5000 万美元）。那些懂行的、关注资格认证周期(qualification cycle) 玩家的人会怎么说... 1. 基于 Celestial 的 Q2 市销率(P/S) 且因为它每季度亏损 1250 万美元... 说它是一只毫无价值的 meme 股？还是 2. 鉴于 2028 年 CPO 超大规模数据中心的机会，它应该被估值 60-100 亿美元？我认为这是美/欧文化 + 市场之间的核心文化分歧。如果你知道如何评估资格认证周期玩家，哪个数字才真正重要？我认为毫无疑问是后者。
英文原文
Imagine Celestial ($MRVL acquisition) was public. The valuation would probably be astronomical in US markets (if I had to guess, $6-10B) Because 2028 revenue projections were $500m (q4 annualized runrate), 2029 revenue projections were $1B. And because Celestial is a core part of hyperscaler CPO programs. Celestial's 2026 Q2 earnings would have been: - Close to $0 revenue (Marvell said "revenue and earnings since the acquisition were not material") - Losing -$12.5M/quarter ($50M/year) Would people who knew what they were talking about looking at qualification cycle players say... 1. based on Celestial's Q2 P/S numbers and because it was losing -$12.5M a quarter... That it's a worthless meme stock? Or 2. with CPO hyperscaler opportunities for 2028, it should be valued at $6-10B? And I think that's a core cultural disagreement between US / EU cultures + markets. If you knew how to evaluate qualification cycle players, which number actually matters? I'd argue overwhelmingly the latter.
-
对比苹果与树莓派在AI供应链中的角色，质疑苹果被动获利。
$AAPL 现在显然是一家重要的 AI 公司了？ -> “OpenAI 囤积数万台 Apple Mac mini 和 Mac Studio”用于强化学习 -> Anthropic 正从 $AMZN 租赁 Apple Mac mini 单元以完成类似的 AI 任务 我从未见过一家公司通过“什么都不做”就能赢得如此彻底。 但每当我看到 Apple + AI，我就会想到我挚爱的 $RPI。据悉，Anthropic 自 27 日起正在与 Raspberry Pi 合作，为其新模型制定硬件标准。 目前尚无关于 Anthropic/OpenAI 购买 Raspberry Pi 设备用于类似任务的消息，但也许某些开源情报(OSINT)研究者可以深入挖掘一下。
英文原文
$AAPL is now an important AI company apparently? -> "OpenAI Hoarding Tens Of Thousands Of Apple Mac mini and Mac Studio" for reinforcement learning -> Anthropic is renting Apple Mac mini units from $AMZN to complete similar AI tasks I've never seen a company win so hard by doing nothing. But whenever I see Apple + AI, I think of my beloved $RPI. Apparently Anthropic is working with Raspberry Pi now for their new model Hardware Standard, as of the 27th. There's nothing yet about Anthropic/OpenAI buying Raspberry Pi devices for similar tasks, but maybe some OSINT researcher can do some digging.
-
台积电高管指出光引擎非瓶颈，激光器、光纤及测试环节才是部署关键。
$TSM 先进封装副总裁： “晶圆代工厂可以扩大光引擎的产能，但大规模部署的真正瓶颈在于其他地方…… 包括激光器、光纤、光纤连接器和测试”（UDN） 以下是一些分类思路： 激光器：$LITE, $SIVE, $COHR, LandMark (3081), LuxNet, Sumitomo, Furukawa 光纤/FAU（光纤阵列单元）：Foci, Browave, $GLW, Sumitomo 光纤/CPO（共封装光学）连接器：BizLink, $GLW, Fujikura, Nextronics, COXOC CPO测试：MPI (6223), Advantest, $FORM, Msscorps, $KEYS, Chroma （披露：我持有上述主题的相关敞口，具体个股名称并非由台积电指定） 我也喜欢其他标的，例如ShunSin，它现在明确拥有TSMC COUPE（共封装统一封装引擎）的敞口，但可能属于封装/集成等不同类别。 例如：“富士康旗下ShunSin确认与台积电COUPE合作” 我认为相比上半年并没有太多新内容，除了持续的新闻（例如上周报告指出FAU在2027年可能出现供需失衡）。 但偶尔重申整体受益者还是很有帮助的……
英文原文
$TSM Vice President of Advanced Packaging: "Foundries can scale up the optical engine, but the real bottlenecks to large scale deployment lie elsewhere... Including lasers, optical fiber, fiber-optic connectors, and test" (UDN) Just some bucket ideas: Lasers: $LITE, $SIVE, $COHR, LandMark (3081), LuxNet, Sumitomo, Furukawa Optical fiber / FAU: Foci, Browave, $GLW, Sumitomo Fiber Optic / CPO Connectors: BizLink, $GLW , Fujikura, Nextronics, COXOC CPO testing: MPI (6223), Advantest, $FORM, Msscorps, $KEYS, Chroma (disclosure: exposure to theme above, individual names not named by TSMC) There are others I like as well, such as ShunSin, which now has explicit TSMC COUPE exposure but probably belongs under a different category like packaging/integration. eg. "Foxconn unit ShunSin confirms TSMC COUPE partnership" I don't think there's anything too new vs. H1, aside from ongoing news (eg. FAU looks like there will have a demand imbalance in 2027, from a report last week). But it's helpful to reiterate overall beneficiaries from time to time...
-
韩国监管加码致杠杆芯片ETF交易量暴跌，微软Windows成散户交易障碍。
我会怀念韩国散户对 $EWY / 三星 / $SKHY 每日波动反应强烈的日子。 新文章：“随着监管限制生效，韩国日内交易者逃离杠杆芯片ETF” - 旨在提供三星/ SK海力士 2倍每日回报的ETF交易量已降至6月峰值的4% - 该要求紧随此前旨在抑制投机活动的高最低存款等措施之后 - 然后出现了最终BOSS……一个为期5天的交易课程，其中： “投资者必须下载仅限Windows的程序，并每天至少花一小时使用虚拟资金进行交易” $MSFT Windows 击溃了韩国散户。
英文原文
I'm going to miss the days of Korean retail reacting heavily to $EWY / Samsung / $SKHY daily volatility. New piece: "Korean day traders flee leveraged chip ETFs as regulatory curbs bite" - Trading value in ETFs designed to deliver 2x daily returns of Samsung / SK Hynix has fallen to just 4% of its June peak - The requirement follows earlier measures, including a higher minimum deposit, aimed at cooling speculative activity - And then came the final boss... A 5 day trading course where: "Investors must download a Windows-only program and spend at least one hour each day trading with virtual funds" $MSFT Windows wiped out the Korean retail.
-
纠正观点：解决AI瓶颈比单纯找电力更难。
@Sovranoq 谢谢指正，这比**仅仅**找到电力要难得多。
英文原文
@Sovranoq Thanks for the correction, harder than **just** finding power.
-
马斯克指出液冷、变压器等基础设施是AI建设难点，博主列出相关概念股。
来自埃隆推文的几个随机想法： $AAON - 冷水机组/液冷 $NVT - 冷却 $MOD - 冷水机组/冷却 $IESC - 布线 $MTRS - 冷水机组 $JCI - 冷水机组 $HPS.A - 变压器（披露我持有该头寸） $PRY.MI - 网络/布线（光纤） $CRDO / $AAOI (同上) / $LITE / $COHR - 网络 既然埃隆·马斯克指出： - 变压器 - 布线 - 液冷 - 大型冷水机组 + 复杂网络 是AI算力建设中的故障点。 这“比寻找电力更难”。
英文原文
Just some random ideas from Elon’s post: $AAON - Chillers / liquid cooling $NVT - Cooling $MOD - Chillers / Cooling $IESC - Wiring $MTRS - chillers $JCI - chillers $HPS.A - Transformers (disclosure I own positions in this) $PRY.MI - networking/wiring(fiber) $CRDO / $AAOI (same) / $LITE / $COHR - networking Since Elon Musk cited: - transformers - wiring - liquid cooling - massive chillers + complex networking As the point of failure for AI compute buildout. Which is “harder than finding power”
-
AI周期虽有局部泡沫，但以NVDA和TSM为核心的结构性趋势将持续加速。
@Omardrisss 更广泛的AI超级周期（如前沿实验室估值、部分Neoclouds）中存在泡沫。偶尔会有些局部回调。但以$NVDA和$TSM为核心的整体主题看起来是结构性的，并将持续加速上行。
英文原文
@Omardrisss There’s bubbles within the broader AI supercycle (eg. frontier lab valuations, some Neoclouds). And some pop here and there. But the overall theme with $NVDA and $TSM in the center of it all looks structural and would keep ramping up.
-
NVDA指引超预期，AI供应链瓶颈环节将迎来爆发。
提醒AI看空者：$NVDA 预测2027年超大规模云厂商支出将达1.3万亿美元。较2026年约8000亿美元大幅增长，尽管产能仍受限，营收增速仍达约70%，远超华尔街此前约45%的预期。这些资本支出流向AI供应链……从存储、网络到晶圆代工厂。尤其是那些未从旧世代中受益的新瓶颈/卡点环节。我们就像派对刚开始2小时，大家起初喝的是札幌生啤。有人往情境(Situational)的酒里加了烈酒，我们不得不暂停派对30分钟。但现在派对重启，桌上摆满了龙舌兰酒。我认为真正的狂欢即将开始。
英文原文
FYI to the AI bears: $NVDA projects $1.3T of hyperscaler spend for 2027. Up from ~$800B in 2026, alongside ~70% revenue growth despite remaining capacity constrained… far above prior Street expectations of ~45%. That capex flows across the AI supply chain… from memory and networking to foundries. Especially through the new bottlenecks/chokepoints that didn’t benefit from older generations. We’re like 2 hours into the party where people started off drinking Sapporo draft beer. Someone spiked Situational’s drinks and we had to pause the party for 30 min. But now it’s back on with tequila shots on the table. I think the real party is about to begin.
-
用胡萝卜比喻说明需求转移及华邦退出加剧供应短缺。
@CryptoBankz_65 两件事：我现在很想吃胡萝卜。第二点是……这只是一个简化的 TLDR（太长不看版）。 因为我没提到对红色(2)、绿色(3)、蓝色(4)胡萝卜需求增加的连锁影响，以及那些只吃蓝色胡萝卜的人。 因为太饿了，开始改吃绿色胡萝卜。 而那些原本只偏爱绿色胡萝卜的人，也开始改吃红色胡萝卜。 或者那位著名的胡萝卜农夫 Winbond（华邦电子），放弃了一座较老的红色胡萝卜农场，导致胡萝卜供应更加短缺。
英文原文
@CryptoBankz_65 Two things: I'm hungry for carrots now. The second was was that... this was simplified TLDR. Since I didn't include the implications of more carrot demand across red (2), green (3), blue (4) carrots and how people who only ate blue carrots. Started to eat green carrots because they were so hungry. And people who only preferred green carrots started to eat red carrots. Or the famous carrot farmer named Winbond, ditching one of their older red carrot farms, causing more carrot shortages.
-
DRAM产能转向导致短缺，ESMT因垄断利基市场而受益。
简单总结一下： 有个叫 $NVDA 的家伙食量巨大。英伟达只想从农民手里买高价土豆。 土豆农 $MU 原本卖胡萝卜……现在把农田改种土豆了。 英伟达导致了一场严重的土豆饥荒，因为它吃得太多。所以其他种胡萝卜的，比如 SK 海力士和三星，都把产能转去种土豆了。 但是……人们仍然需要吃饭。胡萝卜是日常健康饮食的一部分。 现在除了像“ESMT”这样的传统胡萝卜农，没人种胡萝卜了。 所以现在大家都从 ESMT 买胡萝卜，导致短缺，价格上涨。 所以 ESMT 需要更多胡萝卜种子（来自 PSMC）。但 PSMC 的胡萝卜种子也快不够用了。 所以 PSMC 向 ESMT 收取更高的种子费。 但 ESMT 卖胡萝卜的价格远高于 PSMC 提高的种子成本。 现在……ESMT 卖胡萝卜赚得足够多，基于7月财报的交易市盈率仅为1.9倍。 ESMT 能一直卖胡萝卜到2027年吗？ 其他所有农民都这么认为，因为把宝贵的土豆农场改回胡萝卜农场既困难又缺乏动力。
英文原文
Okay just to simplify this: There is a guy who eats a lot called $NVDA. Nvidia only wants to buy high-priced potatos off the farmers. A potato farmer $MU, that was selling carrots... now shifted farmland to potatos. Nvidia caused a great potato famine cause they can eat a lot. So everyone else who farmed carrots like SK Hynix and Samsung moved their farmland capacity to grow potato. But... people still need to eat. And carrots are a healthy part of their everyday diet.. Now there's no more carrots aside from a player like "ESMT", the legacy carrot farmer. So now, because everyone buys carrots off ESMT, there's a shortage and the price goes up. So ESMT needs more carrot seeds (from PSMC). But PSMC is running low on carrot seeds to grow the carrots. So PSMC charges ESMT more for the seeds. But the price ESMT sells the carrots at are much higher than what PSMC hikes the carrod seed price. now... ESMT is making enough from selling carrots that it traded 1.9x P/E off July's earnings. Can ESMT keep selling these carrots through 2027? All the other farmers thinks so since it's both hard + low incentive to migrate their valuable potato farms back to carrot farms.
-
AI需求激增致传统封装产能紧缺，引线键合设备成瓶颈。
之前漏看了这条提及，但IC设计业内人士透露，OSAT（外包半导体组装与测试）厂商已警告客户：“到2027年，传统封装产能可能出现超过20%的缺口”（例如 $AMKR、Powertech 可能受益）。 随着：“引线键合设备”成为主要瓶颈（ASMPT、$KLIC） -> AI服务器出货激增引发了成熟制程芯片需求的飙升 -> OSAT现有的成熟封装产能正被快速吸收 -> 提前锁定传统封装产能正成为一大挑战 级联瓶颈模式正在内存（如DDR2、DDR3）领域出现，现在似乎也蔓延到了封装产能。
英文原文
Missed this mention earlier but IC design insiders reveal that OSAT providers have already warned clients: "traditional packaging capacity could face a deficit exceeding 20% by 2027" (eg. $AMKR, ethereum:0x041ff0e49f6f774e7dc7bd10ee4a14c00b1d80b2, Powertech as potential beneficaries) With: "wire bonding equipment" emerging as the primary bottleneck (ASMPT, $KLIC) -> influx of AI server shipments has triggered a spike in demand for mature chips -> existing mature packaging capacity at OSATs is being rapidly absorbed -> securing early access to traditional packaging capacity is now becoming a challenge Cascading bottleneck patterns are occurring across memory (eg. DDR2, DDR3) and now packaging capacity I guess.
-
传统存储厂商远期市盈率极低，估值取决于层级
@iammichaelz 取决于你关注的是存储器的哪一层级 许多传统 DRAM/NAND 和 NOR Flash 厂商目前的远期市盈率(P/E)仍然仅为 1.8-3 倍
英文原文
@iammichaelz 取决于你看的是哪一层存储 很多传统 DRAM/NAND 和 NOR Flash 厂商目前的远期 P/E 仍然只有 1.8-3 倍
-
存储需求呈结构性短缺，看好板块远期估值上升。
所以……回到存储这边，我们现在有： $SKHY CEO 表示：“我们预计[存储]短缺将持续到 2030 年底” $SNDK 表示：“我们看到直到 2030 年，NAND 闪存都存在结构性的巨大需求”，非 GAAP 毛利率将维持在约 80%。 $NVDA 表示供应承诺从 1,190 亿美元增加到 2,790 亿美元，主要由存储采购推动……与此同时，存储定价处于“极端”水平且还在继续上涨。 Winbond 正在与多家客户就延伸至 2029-2030 年的配额展开讨论。 我之前就说过，存储需求看起来是结构性的，而现在从传统 DRAM 到 HBM 高带宽内存，需求可见度已经延伸到了未来数年…… 所以我确实认为，整个板块的远期 p/e 倍数都有进一步上升的机会。
英文原文
所以… 回到存储这边我们现在有： $SKHY CEO 表示: “我们预计[存储]短缺将持续到 2030 年底” $SNDK 表示: “我们看到直到 2030 年，NAND 都存在结构性的巨大需求”非 GAAP 毛利率将维持在约 80%. $NVDA 表示 供应承诺从 1,190 亿美元增加到 2,790 亿美元，主要由存储采购推动…与此同时, 存储定价处于“极端”水平 而且还在继续上涨. Winbond 正在与多家客户就延伸至 2029-2030 年的配额展开讨论. 我之前就说过, 存储需求看起来是结构性的, 而现在从传统 DRAM 到 HBM，需求可见度已经延伸到了未来数年.. 所以我确实认为, 整个板块的远期 p/e 倍数都有进一步上升的机会.
-
质疑Sivers混合制造策略，支持AAOI/LITE垂直整合模式。
@shawnyang1128 这是他们代工厂合作伙伴在2024年披露的晶圆产能。也许有某家美国厂商被纳入了。不过我不喜欢Sivers在这个阶段转向混合制造模式（我认为其目标仅为1/3的产能）。但我确实同意需要像$AAOI或$LITE那样实现垂直整合。
英文原文
@shawnyang1128 This was disclosed wafer capacity from their foundry partners back in 2024. Maybe one of the US players got added. I’m not a fan of Sivers switching to hybrid manufacturing at this stage though (think it targeted 1/3rd of capacity). But I do agree it’s needed to be vertically integrated like $AAOI or $LITE.
-
驳斥激光供应充裕论，指出市场误解认证周期玩家。
@Kr_Soze 我从未见过有人在瓶颈期声称激光供应有“对 Sivers 的‘巨大产能分配’……‘立即可用’”。 并且披露除了 $JBL 之外还与 6 家新的可插拔(pluggable)制造商合作（共 7 家）。 “表现糟糕”。我认为像你这样的人大多只是误解了如何评估资格认证周期(qualification cycle)的玩家。
英文原文
@Kr_Soze I've never seen someone say new "tremendous capacity allocation for Sivers" for laser supply... "available now" during a bottleneck. And a disclosure working with 6 new pluggable manufacturers beyond $JBL (so 7). "was awful". Think majority of people like yourself just misunderstand evaluating qualification cycle players.
-
澄清定制与可插拔光模块厂商的量产时间表差异。
@carrioresearch 你把定制产品与常规时间表（例如可插拔光模块厂商）混淆了。回顾他们2023-2024年关于Lightmatter、Lightelligence和Celestial的帖子（这些厂商预计2028年起飞），这就说得通了。对于可插拔厂商，他们提到“部分可以实现”非常快速的认证和产能爬坡，可能在2027年。
英文原文
@carrioresearch You're conflating custom products with regular timelines (eg. pluggable players). And this makes sense when you look at their post about Lightmatter, Lightelligence, and Celestial back in 2023-2024 (where these players take off 2028) For the pluggable players they mentioned "some can achieve" very rapid qualification and ramp" likely in 2027.
-
CW激光器瓶颈持续，多厂商产能与新技术进展令博主基本面看好。
CW（连续波）激光器瓶颈持续存在。 -> 2家晶圆厂（Win + 另一主要玩家）为 $SIVE 提供大量产能 -> $JBL 明确的爬坡时间表 - 6家新的可插拔光模块厂商处于评估/后期讨论阶段，带来近期机会 -> $GFS, Ayar, $POET, (可能包括 Celestial + 其他 CPO 共封装光学玩家)。 -> 甚至未包含他们与 TFLN 等其他技术合作的工作。 对此我基本面非常看好。
英文原文
Massive CW laser bottleneck ongoing. -> 2 foundries (Win + another major player) with substantial capacity for $SIVE -> $JBL clear ramp timing - 6 new pluggable players in eval/later stage disc for near term opportunities -> $GFS, Ayar, $POET, (potentially Celestial + other CPO players). -> not even including the other work they do with TFLN and others. I'm very positive fundamentally after this.
-
Sivers CEO确认CW激光器产能现已可用，可立即供货。
@sebiisb 是的，如果现在开始，新的晶圆厂(Foundry)认证通常耗时很长…… 但他们表示，用于连续波分布反馈(Continuous Wave Distributed Feedback, CW DFB)激光器生产的巨大产能“现已可用”。 鉴于连续波(CW)激光器短缺，这真是个惊人的消息。 为了佐证这一点，管理层被问及 Sivers 是否能利用新合作伙伴的产能立即销售激光器。 CEO 表示可以，称他们与该合作伙伴“合作已久”。
英文原文
@sebiisb Yeah, usually new foundry qualifications might take a lot of time if they started now... But they said tremendous capacity is "available now" for CW DFB laser production. Which is amazing news given CW laser shortages. Just to support that too, management was asked if Sivers could sell lasers right away using the new partner capacity. The CEO said yes, that they have been working with the partner "for a while."
-
SIVE是GFS和Ayar激光供应商，1.6T/CPO量产需待JBL等通过认证。
@degenlurker_ @nimishx $GFS 将 $SIVE 作为其参考激光设计，并展示其为共封装光学(CPO)/近封装光学(NPO)/可插拔模块在 GFS 规模化时的激光供应商。Ayar（领先的 CPO 玩家）也将 $SIVE 列为其主要激光供应商。再次强调，所有这些较新的 1.6T 和 CPO（规模化扩展）架构将在 2027/2028 年量产，如果像 $JBL 这样的厂商尚未通过超大规模客户认证，很难提供具体的量产合同/预测。
英文原文
@degenlurker_ @nimishx $GFS has $SIVE as their reference laser design and showcased them as a laser supplier for CPO/NPO/pluggables for GFS scale. Ayar (leading CPO player) has $SIVE listed as their primary laser supplier. Again all these newer 1.6T and CPO (scale up) architectures ramp in 2027/2028, it's hard to provide concrete volume contracts/projections if players like $JBL haven't qualification with their hyperscaler customers.
-
分析师关注超大规模客户供应链及上游瓶颈，而非争论机会管道。
@nimishx KPI 一直是一致的。它只是产能爬坡情况的指标。分析师主要关注的是其超大规模客户供应商，如 $GFS、$JBL、Ayar、$POET 以及其他可插拔/共封装光学(CPO)合作伙伴。以及它们的计划如何扩展，而 $SIVE 处于上游。这比争论机会管道意味着什么更有意义。
英文原文
@nimishx KPI has always been the consistent. It's just an indicator of how things would ramp. Main thing analysts look at are their hyperscaler suppliers like $GFS, $JBL, Ayar, $POET, and other pluggable/CPO partners. And how their programs scale, with $SIVE sitting upstream. Rather than trying to argue with what opportunity pipelines mean.
-
LPK因GCS延迟推迟，SIVE供应1.6T/CPO激光器，架构预计27-28年落地。
@whatdahellson 情况略有不同，鉴于三星电机/SKC Absolics 等行业光组件(GCS)的延迟，$LPK 的收入可能已被推迟。 $SIVE 是 $JBL 1.6T 长距光学(LRO)、Ayar 共封装光学(CPO)及其他厂商的激光器供应商。 1.6T/CPO 并没有出现可信的延迟，只是在等待资格认证完成以便批量合同下达。这些架构将在 2027-2028 年落地。
英文原文
@whatdahellson It's slightly different since $LPK revenue has likely been delayed given industry GCS delays with Samsung Electro-mechanics/SKC Absolics $SIVE is the laser supplier to $JBL 1.6T LRO, Ayar for CPO, and others. There's been no credible delays to 1.6T/CPO, it's just waiting for their qualifications to finish before volume contracts flow. Those architectures are 2027-2028.
-
解释半导体从开发到量产的周期，类比AAOI和AEHR的历史。
@Finishurbkfst23 因为它们正处于资格认证(qualification)阶段。 $AAOI 或 $AEHR 在出货量拐点到来之前也是同样的情况。 这仅仅是理解半导体从开发到资格认证再到量产爬坡(ramp)的周期。
英文原文
@Finishurbkfst23 Because they're qualification stage. Same thing with $AAOI or $AEHR before their volume inflections. It's just understanding semi ramp cycles from development -> qualification -> volume ramp.
-
华邦退出DDR2加剧短缺，南亚科成关键供应商，ASP看涨。
@Berlinergy 华邦电($WIN) 实际上正在退出 DDR2 领域，转而专注于 DDR4/DDR3，并表示预计 2027 年供应将更加紧张。他们已开始讨论 2029-2030 年的产能规划，并向高端产品转型，因此不太可能再向旧产线分配太多资源。至于其他厂商转向旧产线，认证耗时极长……且激励极低。我查看了晶圆代工依赖度，力积电($PSMC) 是台湾企业，地缘政治上没问题。关于晶圆分配，如果戴上供应链韧性这顶帽子，如果南亚科($ESMT) 停止供应 DDR2，下游将遭受毁灭性打击，所以我不认为会出现这种情况（因为如果华邦退出，他们可能是该产线全球最大的独立供应商）。所以我不太认为会发生……但如果他们获得的晶圆分配减少，平均售价(ASP) 只会进一步上涨，因此盈利可能依然强劲。
英文原文
@Berlinergy Winbond is actually withdrawing from DDR2 to focus on DDR4/DDR3 and said they expect 2027 to be even tighter. They're beginning capacity discussions for 2029-2030 + moving toward higher end, so don't quite see them having much to allocate toward older lines again. As for other players moving toward older lines, qualification takes a lot of time... and there's just very low incentive to. I looked for foundry dependence, PSMC is Taiwanese, so all good geopolitically. As for wafer allocations, if I had to put my supply chain hat resiliency hat on, if ESMT stopped supplying DDR2, it would just be absolute destruction downstream, so don't see that scenario (since they're probably the world's largest merchant supplier for this line if Winbond exits) So don't quite think it would happen... but if they get less wafer allocations, the ASP just gets hiked even more, so earnings would likely remain strong.
-
CW激光器产能紧缺，$SIVE作为剩余现货供应商受益。
我知道大家都在关注 $NVDA 的财报…… 但你们知道 Landmark 已与美国客户签署为期4年的连续波(CW)协议以确保“充足供应”吗？ 所以现在的格局是： - $LITE（产能耗尽） - $COHR（产能耗尽） - Landmark（产能已锁定） - $AAOI（光模块产能） - $MTSI（尚未量产） - $SMTC（产能有限） 我记得今年早些时候说过，CW激光器将是下一个光学变革，且会被 Nvidia 严重瓶颈化？ 看着这一切上演很有趣，参与者已签署至2030年的长期协议(LTA)（暗示其他玩家可用产能更少）。 这对 $SIVE / Win Semi 的影响非常重大，因为它是少数拥有产能（+ CPO级激光器）的剩余CW现货供应商之一。 让我们看看他们的执行情况。
英文原文
I know everyone’s watching $NVDA earnings right now… But did you know Landmark signed a 4 year CW agreement with US customer to ensure “sufficient supply”? So now you have: - $LITE (capacity gone) - $COHR (capacity gone) - Landmark (capacity committed) - $AAOI (capacity for transceivers) - $MTSI (not online) - $SMTC (limited) I remember saying earlier this year CW lasers would be the next optical shift and heavily bottlenecked by Nvidia? So fun watching this play out, with players signing LTAs already to 2030 (signaling less availability for other players). Implications for $SIVE / Win Semi are very material as one of the few remaining CW merchant suppliers with capacity (+ CPO-grade lasers). Let’s see how they execute.
-
博主因DDR技术影响兴奋，无暇关注英伟达财报。
@4intheflames 我对DDR3/DDR2的影响过于兴奋，以至于无暇关注$NVDA的财报或睡觉。
英文原文
@4intheflames I’ve been too excited about DDR3/DDR2 implications to look at $NVDA earnings or sleep.