供应链分析
产业链结构、上下游与瓶颈环节研究 · 共 1122 条 · 第 9 / 23 页
-
认为 AAOI 自有供应链和产能扩张会带来更高毛利。
@Kaizen_Investor 依我看,大概能有 30%->40% 的提升,凭经验说的。 他们从 InP 激光器 fab 到组装,把整个供应链都放在内部,所以毛利率会随着时间上升。 尤其是当使用老一代 $NVDA 机型的主权数据中心,随着超大规模客户从 800G 转向 1.6T,会买更多 800G。
英文原文
@Kaizen_Investor 30->40% off the top of my head. They do the entire supply chain in-house from inp laser fab to assembly, so gross margins should go up over time. Especially since sovereign DCs using older gen $NVDA models buy more 800g as hyperscalers move to 1.6T
-
称 AAOI 收到超大客户 7100 万美元新订单,产能翻上去后营收会很夸张。
Applied Optoelectronics($AAOI)今天说,他们从超大规模客户那里又拿到了一笔 7100 万美元的新订单。 流入的订单源源不断。 如果他们成为美国最大的 800G / 1.6T 产能,而且光学需求真的起飞, 营收转化会相当夸张。
英文原文
Applied Optoelectronics ( $AAOI ) said it received a new $71 million order from its hyperscaler customer today. The orders flowing in are endless. If they become the largest 800G/1.6T capacity in the US and optical demand does take off. The revenue translation would be pretty insane.
-
认为 AXTI 掌控行业命脉,若武器化会带来巨大回报。
@americangeomap $AXTI 掐住了整个行业的咽喉。 如果它真的把这种优势武器化,我觉得没有哪只票能比得上 AXT 的回报。
英文原文
@americangeomap $AXTI has a stranglehold over the entire industry. If it decides to weaponize that, don’t think any pick can beat AXT returns
-
澄清 NBIS 不是做光互连,而是更偏 neocloud 和数据服务。
@GeorgeMakro91 $NBIS 并不做光互连。 它更像是 neocloud、robotaxi、数据标注、数据库这一类的业务。
英文原文
@GeorgeMakro91 $NBIS does not do optical interconnects. It's more in the neocloud, robotaxi, data labeling, database type space.
-
认为 AAOI 像早期 SNDK,关键取决于执行和产能目标。
@Market_Nick_ $AAOI 看起来像早期的 $SNDK。 成败很大程度上取决于执行,但他们预测自己会从 InP 激光器 fab 到组装,拿下美国最大的 1.6T 产能,这一点挺看多的。
英文原文
@Market_Nick_ $AAOI does look like an early $SNDK. Largely dependent on execution, but the fact they project they'll have the largest 1.6T capacity in the US from projections (from inp laser fab to assembly), is pretty bullish.
-
说明 SIVE 还有 OTC 代码。
@Sumojoe118 $SIVE 在美国 OTC 还可以用 $SIVEF 交易。
英文原文
@Sumojoe118 $SIVE trades US OTC as $SIVEF too.
-
消费级光子市场比AI数据中心更大,光子公司不应只盯大客户订单。
消费级激光剑产品的市场规模比AI数据中心还大。说真的不明白这些光子公司在干嘛,被$MSFT、$AMZN的超大规模数据中心订单分散了注意力。
英文原文
@tomszczypka Consumer lightsabers have a bigger TAM than AI DCs. Not sure what these photonic companies are doing, getting distracted by $MSFT, $AMZN hyperscaler orders tbh.
-
按类别列出自己喜欢的几只名字。
@FranGGlez 哦,每个类别来一只: 1. $MRVL 2. $SIVE 3. Win 4. $SOI 5. $LITE
英文原文
@FranGGlez Oh one from each category: 1. $MRVL 2. $SIVE 3. Win 4. $SOI 5. $LITE
-
认为激光瓶颈下很难选赢家,Sivers 以较低市值却进入大客户路线图而很有吸引力。
100% 是激光。 我个人不会往下游去做组装之类的业务。 但在激光瓶颈下,真的很难在 Luxnet、Furukawa、Sumitomo、$SIVE、$MTSI、$LITE、$SMTC(在 Heico 收购 CW 之后)这些公司里挑出唯一赢家…… 不过我觉得 Sivers 特别有吸引力,因为它们已经进入了 $JBL 的 1.6T 路线图,以及 $MRVL 的 CPO 路线图,而市值才大约 2.9 亿美元。
英文原文
100% lasers. I personally wouldn't go downstream into assembly and others. But with the laser chokepoint it's really hard to pick a winner between stuff like Luxnet, Furukawa, Sumitomo, $SIVE, $MTSI, $LITE, $SMTC (after heico cw acqusition) and others... But I found Sivers to be really compelling since they're in $JBL 1.6T + $MRVL CPO roadmap at ~$290m MC.
-
引用 Lightcounting 认为光互连到 2030 年可到 1000 亿美元,并列出自己偏好的链条标的。
Lightcounting:"这不是笔误" 到 2030 年,光互连市场很有可能从 2025 年的约 190 亿美元增长到 1000 亿美元以上。 如果你想看我最喜欢的名字: 复合型:$MRVL、$SMTC CW 激光:$SIVE、$MTSI、$AAOI 晶圆代工:$TSEM、Win Semi 衬底 / 外延:$AXTI、$SOI、$IQE 黄金标准:$LITE、$COHR 下一轮 TAM 扩张的乘数,主要由 CPO / Scale Up 驱动,而这大多由 SiPh 和外部 CW 激光推动。
英文原文
Lightcounting: "This is not a typo" Optical interconnects has a reasonable chance to reach $100B+ by 2030 from ~$19B (2025). If you're looking for my favorite names: Compounders: $MRVL, $SMTC CW Lasers: $SIVE, $MTSI, $AAOI Foundries $TSEM, Win Semi Substrates/Epitaxy: $AXTI, $SOI, $IQE Gold Standard: $LITE, $COHR The next TAM expansion multiplier is CPO/Scale Up, largely driven by SiPh and external CW lasers.
-
认为 AAOI 的美国一体化制造和转商用机会很有价值。
执行风险永远都存在!但我真的很喜欢 $AAOI 这一点:他们把 InP 激光器从 fab(制造)到组装全部都留在美国本土完成。 这不仅让他们有机会拿到类似美光或 SK 海力士那样的估值溢价,还给了他们从商用模式(merchant model)转型,以及向仍然被低估的 TAM 扩张方向切换的选择权。
英文原文
実行リスクは常につきものです!しかし、$AAOI がInPレーザーのファブ(製造)から組み立てに至るまで、すべてをアメリカ国内で完結させているところが本当に気に入っています。 そしてこれにより、彼らはマイクロンやSKハイニックスのようなプレミアム(プレミアム評価)を獲得できるだけでなく、マーチャント・モデル(外販)への移行や、まだ過小評価されているTAM(獲得可能な最大市場規模)の拡大という新たな機会へピボット(方向転換)するためのオプショナリティ(柔軟な選択肢)も得られるのです。
-
重申 AAOI 可能拥有美国最大的 1.6T 产能,市场仍低估其前景。
再提醒一下: $AAOI 预计会拥有美国最大的 1.6T 光收发器产能。 而且在牛市情景下,光学 TAM 预计会扩张到 1000 亿美元(Lightcounting)。 AOI 早先的业绩电话会里就说过:“我们预计很快会在美国拥有最大的……1.6T 收发器产能。” 他们预计 2026 年第三季度开始发货,并且从 2027 年下半年开始,产能会达到每月 3.78 亿美元,也就是如果超大规模客户愿意买他们做出来的任何东西,年化就是 43.5 亿美元。 现在还是 2026 年上半年,所以因为市场波动就开始恐慌还稍微早了点。 因为 AOI 把整个供应链都自己包了,从 InP 激光器 fab、设计到组装全都自有。 而且他们在美国 + 台湾正在做回流(而不是继续从中国 / 马来西亚 / 泰国采购)。 市场尤其会漏掉这一点: -> 他们在供应链垂直整合的任何环节,都有向 ELS 甚至 merchant model 扩张 TAM 的可选性。 $AAOI 交易在 64 亿美元市值左右,主要是因为执行不确定性和宏观环境。而只要还没真正兑现,执行风险就一直会是个投机性压制。 但如果他们有任何迹象表明能完成目标,而且宏观转好,我们大概就看不到现在这种估值了。
英文原文
Just a gentle reminder: $AAOI expects to have the largest production capacity of 1.6T optical transceivers in America. Just as optical TAM is expected to ramp to $100B in a bull case (Lightcounting). AOI: “We expect that we will soon have the largest production capacity for … 1.6T transceivers in the US” from their former ER transcript. They expect shipments to begin Q3 of 2026 and projected $378M / month capacity starting H2 2027, which is $4.35B annualized if hyperscalers buy anything they make. We’re in H1 of 2026, so it’s a tad early to be panicking from market volatility. Since AOI does the entire supply chain from InP laser fab, design, and assembly all captive. And especially in the USA + Taiwan operations that they’re reshoring (rather than sourcing to China/Malaysia/Thailand) Markets miss this part especially: -> They have TAM expansion optionality for ELS to merchant models in any parts of their vertically integrated supply chain. $AAOI is trading at $6.4B largely because of execution uncertainty and macro. And execution will always be an speculative overhang until it comes time to deliver. But if there’s any indication that they hit their targets and macro clears up, we probably won’t see current valuations again.
-
认为美国企业采购瑞典激光供应商很有意思,欧洲高技术并不意外。
@Sofigoodboy 看到美国企业正在使用像瑞典的 $SIVE 这样的尖端激光供应商,供应链这件事真的很有意思! 不过欧洲本来就以像 $ASML 这样的顶尖技术著称,所以这也不算特别意外。
英文原文
@Sofigoodboy 미국 기업들이 스웨덴의 $SIVE 와 같은 최첨단 레이저 공급업체를 이용하고 있다는 사실을 보면 공급망이란 정말 흥미롭습니다! 하지만 유럽은 $ASML 처럼 뛰어난 기술력을 보유한 것으로 잘 알려져 있으니, 그리 유별난 일도 아니지요.
-
认为下一轮光子学升级里,SIVE 这类纯激光公司仍有空间。
当然还有 Sumitomo、Furukawa、$MTSI、$LITE、$COHR 这些公司…… 但像 $SIVE 这样,作为下一轮光子学放量的纯 CW 激光标的,真的很少见。 很多人现在看的是当前周期的可靠性,但下一代架构不一样,老版本可能会过剩。所以我觉得在 CPO / scale up 里,新的小公司还有机会冒头。
英文原文
Yes there’s others out there like Sumitomo, Furukawa, $MTSI, $LITE, $COHR, and these types of players… But very rare to you see a pure play cw laser for the next photonics ramp like $SIVE. People are looking at current cycles for reliability, but next architectures are different and might glut older versions. So I think there’s room for new smaller players to shine for cpo/scale up
-
同意供应链视角很重要,自己更偏好小市值的 AEHR。
@KawzInvests 这份供应链总览做得不错!是的,$VIAV、$AEHR 都很好。 我个人只是更喜欢像 Aehr 这种更小市值的名字,因为这意味着上行空间更大,但 beta 也更高,因为它还没那么被验证。
英文原文
@KawzInvests Nice high level overview on the supply chain! Yeah $VIAV $AEHR are all great. I just personally like the smaller MC names like Aehr, since it implies higher upside, but higher beta since it's less proven.
-
认为对 MU 和 AMD 来说都是好消息,只是不喜欢市场把合同义务武器化。
@Mismo_Mktg 对 $MU 到 $AMD 这些公司来说,当然都是好消息。 我只是对他们为了走到这一步而把合同义务武器化感到很烦。
英文原文
@Mismo_Mktg Oh it’s great news for sure from everything like $MU to $AMD. I’m just annoyed how they weaponized contractual obligations to get there
-
认为大多数人根本不知道 Celestial 及其供应链。
@Akskn7 我觉得 99% 的人甚至都不知道 Celestial 是什么,更别说他们的组件供应商了。 不过他们可能知道 $MRVL,主要是因为 Jim Cramer,但即便那也算比较小众。
英文原文
@Akskn7 I don’t think 99% of people even know what Celestial is let alone their component suppliers. They might know $MRVL because of Jim Cramer though but even that is considered niche.
-
认为市场终于开始意识到,Nvidia 投资 Marvell 背后对应的是上游光子链重估。
市场终于开始意识到的事情: 下游:$NVDA 向 $MRVL 的 Celestial CPO 项目投资了 20 亿美元: $POET(Celestial 的 interposers):财报里说今年要出货 3 万台光引擎。 $SIVE 是 Poet 和 Celestial 在 Marvell 光子路线图中近中期的光源。 Win Semi 是 InP CW 激光生产的晶圆代工。 而更上游的还有:$SOI / $AXTI 这些衬底,以及 $IQE 的外延片。 这就是下一轮光子学超级周期的起点,而且估值便宜得离谱。 而且这发生得比市场和机构在 2026 年预期的还要早,Nvidia 正在抢跑整条转变路径。 99.9% 的人只看到了 Marvell 下游的变化,但还没意识到上游的供应冲击会在哪里发生。
英文原文
Something that markets are finally starting to realize: Downstream: $NVDA invested $2B into $MRVL Celestial CPO program: $POET (Celestial’s Interposers): To ship 30,000 optical engines to ship this year from earnings today. $SIVE is the light source of Poet and Celestial for near/mid term Marvell photonics roadmap. Win Semi is the foundry for InP CW laser production. And even more upstream: $SOI / $AXTI for substrates and $IQE for epiwafers. This is the very start of the next photonics supercycle, at dirt cheap valuations. And it’s happening earlier than markets and institutions expected in 2026, with Nvidia frontrunning this entire shift. 99.9% of people see the changes happening downstream with Marvell, but don’t realize yet where the supply shocks will happen upstream.
-
认为 Nvidia 给 Marvell 20 亿美元投资后,更看好 Celestial 生态。
@B38B37 @PhotonCap $NVDA 给 $MRVL 20 亿美元,确实让我对 Celestial 生态里的 $POET 和 $SIVE 更看多了一点。
英文原文
@B38B37 @PhotonCap $NVDA handing $2B to $MRVL does make me a bit more bullish on the celestial ecosystem like $POET and $SIVE.
-
继续偏爱激光链条,而不是只看 POET。
@PhotonCap 我以前提过 $POET!不过就我个人而言,我还是更喜欢 $LITE 和 $SIVE 这种激光侧的标的。
英文原文
@PhotonCap I’ve talked about $POET before! But personally I’ve liked the laser side of things more with $LITE and $SIVE
-
说明 AAOI 做光器件的全链路业务,不会在转型中被替代。
$AAOI 会先做激光 -> 设计 -> 组装 -> 再把收发器卖出去。所以它们把整条供应链都包了。 可插拔光器件不会消失,所以他们会填补那一块。 然后在另一边,他们还做 ELSFP(外部光源),所以在 CPO 侧就是激光 -> 设计 -> 封装。 也就是说,这两边他们都做,不会在行业转型里被替代掉。
英文原文
$AAOI does laser -> design -> assembly -> then sells the transceiver. So they have the whole supply chain. Pluggables won't go away so they'll fill that gap. Then on the side, they build ELSFPs (external light sources), so laser -> design -> packaging on the CPO side. So they do both. They don't get replaced in the transition.
-
认为去年 NVDA 先把产能拿走了,时间大概是 2024 末或 2025 初。
@LIama002 去年这事其实是在关门后发生的,当时 $NVDA 已经把所有产能都拿走了。 Trendforce 还发过文,说明这事在去年就更早发生了。 也许是 2024 年末,或者 2025 年初?
英文原文
@LIama002 It was behind closed doors last year with $NVDA securing all the capacity. Trendforce published that it happened earlier last year. Maybe late 2024, early 2025? https://t.co/WbwGaQ6Gwm
-
认为市场低估了 Nvidia 投资下游后带来的上游供应冲击。
市场漏掉了 Nvidia 投资的含义: 今天它向光学领域的 $COHR 投了 20 亿美元。 向光学领域的 $LITE 投了 20 亿美元。 今天又向光学领域的 $MRVL 投了 20 亿美元。 Nvidia 去年就用过同样的剧本。 他们意识到 800G/1.6T 可插拔器件的推进会耗尽全球 EML 供给。 所以他们接触了 $LITE、$COHR、Sumitomo,并提前预定了大部分产能。 我们已经在他们股价里看到了反映:自从那轮严重供给紧张以来,$LITE 已经涨了 955%。 现在我们看到同样的剧本在过去一个月里再次开始。 只是这次是一个新的架构。 而且这些交易都包含了几十亿美元的采购承诺和未来产能权利。 所以……下一步是什么?CW / EML 和 CPO 瓶颈。 Nvidia 只是更喜欢投资下游玩家,但真正的供应紧张发生在上游。 来自 $MTSI、$SIVE、$LITE、$COHR、Furukawa 和 Sumitomo 的激光供应商都在全速运转。 Win Semi、$TSEM、$GFS 这些晶圆代工也大概率在满负荷运转。 整条供应链都会受益(比如 $AEHR 的测试、$SOI 的衬底)。 但从晶圆代工到 ELS / CW 激光瓶颈的这两个环节,可能会是最大的受益者。 Nvidia 是下一轮发生什么的最大信号;现在只是等拐点到来的时间而已。
英文原文
The market are missing the implications from $NVDA investing: $2B into $COHR for optical. $2B into $LITE for optical. and $2B into $MRVL for optical today. Nvidia did this exact same playbook last year. They realized the push to 800G/1.6T pluggables would exhaust the global supply of EML. So they approached $LITE, $COHR, Sumitomo, and preallocated majority of production. And we've seen this reflected in their share price with $LITE rising 955% since the major supply squeeze. We're seeing the beginning of the same playbook happen now over the last month. Just for a new architecture, this time. As these deals included multibillion-dollar purchase commitments and future capacity rights. So... what's next? CW/EML and CPO bottlenecks. Nvidia just prefers to invest in downstream players. But the supply crunch happens upstream. Laser suppliers from $MTSI, $SIVE, $LITE, $COHR, Furukawa, and Sumitomo are on overdrive. Foundries from Win Semi, $TSEM, $GFS are likely on overdrive. The entire supply chain benefits (eg. testing from $AEHR, substrates with $SOI). But these two segments from foundries to ELS/CW laser chokepoints are likely to be the biggest beneficiaries. Nvidia is the biggest signal of what's coming next; it's just a waiting period for the inflection point to hit.
-
AEHR获光收发器厂商认证订单,处于类似AAOI早期阶段,面临量产拐点。
$AEHR 从一家领先的光收发器(optical transceiver)公司获得了资格认证订单。 它正处于早期阶段,类似于 $AAOI 正在被主要超大规模云服务商(hyperscaler)供应链测试用于光收发器/硅光子学(silicon photonics)阶段。 在可能随时到来的大批量量产拐点(mass volume inflection point)之前。 $AEHR 盘前上涨 12.73%。
英文原文
$AEHR new qualification orders from a leading optical transicever company. It’s in the early $AAOI stage where it’s getting tested by major hyperscaler supply chains for optical transceivers/silicon photonics. Before the mass volume inflection point that may be at any time. $AEHR up 12.73% premarket.
-
认为 POET 受益,但 SIVE 对 Nvidia 和 Marvell 的 Celestial 更关键。
@marioyep5 是的,$POET 确实是受益者,但 $SIVE 作为光源,对 $NVDA 和 $MRVL 的 Celestial 现在来说非常关键。
英文原文
@marioyep5 Yes, $POET is beneficiary but $SIVE is incredibly important to $NVDA and $MRVL Celestial now as the light source. https://t.co/mn3oESTKlw
-
认为 Nvidia 正在强推 CPO 路线图并锁定主要参与者。
@TheValueist 是啊,$NVDA 真的在强力推动他们的 CPO 架构路线图,而且正在把最大的几家玩家锁进自己的标准里。 我们正在实时看到光子学里下一步会发生什么。
英文原文
@TheValueist Yeah $NVDA is really pushing hard on their CPO architectural roadmap and locking in all the biggest players into their standards We’re seeing what’s coming next in photonics real time.
-
认为 Nvidia 可能更希望 SIVE 保持独立,也可能出现股权收购或被 Broadcom 并购。
@yuntungshieh 我觉得 Nvidia 可能更希望 $SIVE 保持独立。 所以我们可能会看到 $MRVL、AlChip 对它做股权收购……或者也可能被 $AVGO 收购,用来卡死竞争对手的 CPO 路线图。现在还不确定最后会怎么演。
英文原文
@yuntungshieh I think Nvidia would rather have $SIVE stay independent. So we might see some share acquisition from $MRVL, AlChip… or maybe an acquisition by $AVGO to chokepoint their competitors CPO roadmap. Not sure how this plays out.
-
调侃如果 Broadcom 买下 SIVE,就能控制后续多家公司的路线图。
如果 $AVGO 真的花 3 亿美元买下 $SIVE,那就太可惜了…… 这样一来,它就能从上游光源层面控制 $NVDA 和 $MRVL 的近端路线图。 我相信博通肯定不会在想这个……对吧?
英文原文
It would be a shame if $AVGO bought $SIVE for $300m… Then controlled the near term roadmap for $NVDA and $MRVL from the upstream light source level. I’m sure Broadcom isn’t thinking about this… right? https://t.co/wrS431DG8n
-
认为 POET 受益巨大,但很多分析师忽略了其光源来自 SIVE。
@InvestWithFelix 对,$POET 是巨大的受益者。它在 Celestial 体系里更广为人知。 但连分析师都会忽略,Poet 的 ELS 光源其实来自 $SIVE。 不过 Nvidia 对这个生态的验证,对整个链条都很重要。
英文原文
@InvestWithFelix Yep $POET is a massive beneficary. It’s more well known with Celestial. Even analysts miss though Poet’s light source for ELS comes from $SIVE. But huge for their whole ecosystem that Nvidia is validating it
-
把 Nvidia 备货 EML 产能的历史剧本,类比到下一代光子瓶颈。
这就是 $NVDA 当年提前锁定全部 EML 产能时的那套剧本,在那之前超大规模客户那边就已经出现了巨大的供给冲击。 这也是为什么我一直在为下一处瓶颈里的 $SIVE 拉响警报。 因为 $NVDA 现在也在下一代架构上做同样的事。
英文原文
This is the exact playbook $NVDA did securing all the EML capacity before there was a massive supply shock for the rest of the hyperscalers. That’s why I’ve been sounding the alarm with $SIVE in the next chokepoint. Since $NVDA is doing the same with the next gen architectures.
-
认为 SIVE 已经进入 Nvidia 路线图,且市值仍很小。
@degenlurker_ 有意思的是,$SIVE 现在已经被写进 $NVDA 的路线图里了,而市值还只有大约 3 亿美元。 只是现在还没人把 Sivers -> Poet -> Celestial -> Marvell 这种多级关系串起来,虽然 Nvidia 已经砸了 20 亿美元去加速他们的 CPO 路线图了。
英文原文
@degenlurker_ It’s funny $SIVE is now baked into $NVDA roadmap now at ~$300M MC. Just no one associates multi hop relations between Sivers -> Poet -> Celestial -> Marvell yet when Nvidia invested $2B into accelerating their CPO roadmap yet.
-
Nvidia 对 Marvell 的 20 亿美元投资,也验证了 SIVE 的光源位置。
刚刚消息出来:$NVDA 向 $MRVL 投了 20 亿美元。 这笔交易主要是为了“联合硅光工作”,也验证了 Marvell 和 Celestial AI 的光互连 CPO 路线图。 但……谁在给 $MRVL 的光源供电,这点市场可能漏了?(提示:$SIVE)
英文原文
Just in: $NVDA invests $2B in $MRVL. This deal is largely for “joint silicon photonics work” and validates Marvell and Celestial AI's optical fabric CPO roadmap. But… guess who powers the light source for $MRVL that markets might have missed? (Hint: $SIVE) https://t.co/9lqJLIx7nG
-
以Marvell收购Celestial为例,分析CPO供应链中Sivers作为CW激光器关键瓶颈供应商的巨大TAM扩张潜力。
看看$MRVL收购Celestial($SIVE是其~2.9亿美元的光源供应商)这个例子就知道了。Marvell以55亿美元收购Celestial(32.5亿美元+22.5亿美元里程碑付款)。2026/2027年预期:零收入,亏损5000万美元。2028年预期:5亿美元收入。2029年预期:10亿美元收入。这是巨大的TAM(市场规模)扩张和收入加速增长。Sivers只是Celestial之上的客户之一,此外还有O-Net、Ayar、其他未公开的CPO(共封装光学)厂商,Jabil负责桥接量产爬坡并赢得良率。那些对光子学毫无理解的散户……正在做的事相当于:因为Celestial在2025、2026年没有CPO收入,就把它建模为5000万美元。因为它目前处于净亏损状态,而不是着眼于2027年之后的大规模爬坡。$SIVE是我抢先布局下一个架构超级周期(位于最早期阶段)的首选标的,涉及ELS TAM扩张+CPO+连续波(CW)激光器瓶颈。如果你想找后期介入的标的,可以看$LITE,它已经涨了3800%。
英文原文
Just look at $MRVL Celestial ( $SIVE is the light source at ~$290m ) as an example. Marvell bought it for $5.5B. ($3.25B + 2.25B milestones). 2026/2027 expectations: 0 revenue, -$50m loss 2028 expectations: $500m revenue 2029 expectations: $1B revenue. This is massive TAM expansion and revenue acceleration. This is just one of Sivers customer on top of O-Net, Ayar, other undisclosed CPO players, with Jabil bridging volume ramp and Win with yields. Random retail users with no understanding of photonics… are doing the equivalent of modeling Celestial at $50m because of no CPO revenue in 2025 2026. And because it’s operating at a net loss instead of looking at the massive ramp 2027 onward. $SIVE is my #1 pick for frontrunning the next architectural supercycles at the very beginning for ELS TAM expansion + CPO + CW laser chokepoints. If you want a later stage pick, you can look at $LITE that already ran 3800%.
-
列出几个关键瓶颈点,认为 VNP 最接近西方版 AXTI。
是的,我花了很多时间把这个图整理出来。 瓶颈 1:$VNP 可能是西方世界里最接近 $AXTI 的单一瓶颈,但在镓、铟、砷等材料上,它还达不到同样的规模。 瓶颈 2:pBN 坩埚,也许是 Shin-Etsu Chemical? 瓶颈 3:第二层可能是 Sumitomo / $COHR。
英文原文
Yep, I spent a lot of time making this on the chart. Chokepoint 1: $VNP is probably the closest you get to $AXTI in the West for one chokepoint, but nothing to the same scale for (gallium, indium, arsenic, etc. ) Chokepoint 2: pBN Crucibles, maybe Shin-Etsu Chemical? Chokepoint 3: Sumitomo / $COHR for the second.
-
表示自己只是做高层分析,不想把整条光子链讲得太细。
@swisssong 不客气。我觉得大家应该把这份分析当成非常高层次的理解,而不是对整个行业的深度拆解。 我也不太明白 Mirae 怎么漏掉了从 $MTSI 到 Luxnet 这么多名字……
英文原文
@swisssong Anytime. I think people should treat the analysis as extremely high level, not a deep dive into the industry. I'm not quite sure how Mirae missed so many names from $MTSI to Luxnet...
-
只是举了一个机器人供应链的例子。
@RonDeSantis 这只是机器人供应链的一个例子:
英文原文
@RonDeSantis This is just one example for the robotics supply chain: https://t.co/ALmJSwnZ4L
-
整理Mirae Asset的CPO价值链全图,涵盖激光源到EDA各环节上市公司
来自Mirae Asset的CPO价值链概览: 激光源: Coherent < $COHR > Lumentum < $LITE > Furukawa Electric (TYO: 5801) Yuanjie Semiconductor (SHA: 688498) Innolight Technology / 中际旭创 (SZSE: 300308) PIC代工: TSMC < $TSM > GlobalFoundries $GFS Samsung Electronics (KRX: 005930) Tower Semiconductor < $TSEM > EIC、驱动IC: Broadcom < $AVGO > Marvell < $MRVL > NVIDIA < $NVDA > ELS、光引擎: Innolight / 中际旭创 (SZSE: 300308) TFC / 苏州天孚光通信 (SZSE: 300394) O-Net Technologies Eoptolink Technology (SZSE: 300502) FAU(光纤阵列单元): Senko Advanced Components(私企) Sumitomo Electric (TYO: 5802) TFC (SZSE: 300394) FOCI Fiber Optic Communications (TWO: 3363) FAU、对准工具: ficonTEC(私企) All Ring Tech (TWO: 6187) ADST(私企) FAU、引擎组装: Fabrinet < $FN > Hon Hai / Foxconn (TWSE: 2317) ASE Technology < $ASX > FOCI (TWO: 3363) OSAT、先进封装: ASE Technology < $ASX > Amkor < $AMKR > Kyocera (TYO: 6971) Powertech / PTI (TWSE: 6239) Shinko Electric (TYO: 6967) Fabrinet < $FN > 连接器、套管: Senko Advanced Components(私企) Sumitomo Electric (TYO: 5802) US Conec(私企) T&S Communications (SZSE: 300570) Molex(私企) Browave (TWO: 3163) 光纤: Corning < $GLW > Sumitomo Electric (TYO: 5802) Nittobo / 日东纺绩 (TYO: 3110) 电/光测试: Keysight < $KEYS > Teradyne < $TER > FormFactor < $FORM > Chroma ATE (TWSE: 2360) Multilane(私企) 交换、系统: NVIDIA < $NVDA > Broadcom < $AVGO > Marvell < $MRVL > Google < $GOOGL > EDA: Synopsys < $SNPS > Cadence < $CDNS > Ansys < $ANSS > 对这些名单中的一些名称感到困惑,他们可能混淆了一些公司,比如把Innolight和激光源公司(如$MTSI、Sumitomo、$SIVE、Luxnet)与实际终端模块混在一起了(除非有什么非公开材料或者我遗漏了)。 但对于对这个领域感兴趣的人来说,这是一个很好的高层概览。 (该推文引用了 @aleabitoreddit 的推文,引用内容仅供理解语境): CPO领域Mirae分析师笔记: 规模化方向: CPO ASIC:$AVGO、$MRVL 光收发器:$COHR、$LITE、中际旭创 DSP/PAM4:$AVGO、$MRVL 相干DCI:$CIEN、$NOK OCS设备:iPronics、Polatis 光缆/光纤:$GLW、Prysmian、Furukawa HCF:$LITE、OFS DCI相干:Ciena、Nokia、Huawei 光放大器:$LITE、$COHR OCS网关:KDDI 规模提升方向: SiPh代工:imec、$GFS、$TSM SiPh调制器:$NVDA(内部MRM)、$INTC ELS:NTT、Furukawa、$LITE、$COHR 太赫兹互连:研发阶段? CPO测试:"新玩家进入增加" 微透镜/光学系统:"新玩家进入增加" TLDR:规模提升的CPO即将到来。我认为分析师笔记遗漏了一堆上游公司名称,并把ELS和光源混为一谈了。 但对于了解他们认为谁是领先玩家来说,这是个很好的高层概览。
英文原文
CPO Value Chain Summary from Mirae Asset: Laser Source: Coherent < $COHR > Lumentum < $LITE > Furukawa Electric (TYO: 5801) Yuanjie Semiconductor (SHA: 688498) Innolight Technology / Zhongji Innolight (SZSE: 300308) PIC Foundry: TSMC < $TSM > GlobalFoundries $GFS Samsung Electronics (KRX: 005930) Tower Semiconductor < $TSEM> EIC, Driver IC: Broadcom < $AVGO > Marvell < $MRVL > NVIDIA < $NVDA > ELS, Optical Engine Innolight / Zhongji Innolight (SZSE: 300308) TFC / Suzhou TFC Optical Communication (SZSE: 300394) O-Net Technologies Eoptolink Technology (SZSE: 300502) FAU (Fiber Array Unit): Senko Advanced Components (Private) Sumitomo Electric (TYO: 5802) TFC (SZSE: 300394) FOCI Fiber Optic Communications (TWO: 3363) FAU, Align Tools: ficonTEC (Private) All Ring Tech (TWO: 6187) ADST (Private) FAU, Engine Assembly: Fabrinet < $FN > Hon Hai / Foxconn (TWSE: 2317) ASE Technology < $ASX > FOCI (TWO: 3363) OSAT, Advanced Packaging: ASE Technology < $ASX > Amkor < $AMKR > Kyocera (TYO: 6971) Powertech / PTI (TWSE: 6239) Shinko Electric (TYO: 6967) Fabrinet < $FN > Connector, Ferrule: Senko Advanced Components (Private) Sumitomo Electric (TYO: 5802) US Conec (Private) T&S Communications (SZSE: 300570) Molex (Private) Browave (TWO: 3163) Fiber: Corning < $GLW > Sumitomo Electric (TYO: 5802) Nittobo / Nitto Boseki (TYO: 3110) E/O Testing: Keysight < $KEYS > Teradyne < $TER > FormFactor < $FORM > Chroma ATE (TWSE: 2360) Multilane (Private) Switch, System: NVIDIA < $NVDA > Broadcom < $AVGO > Marvell < $MRVL > Google < $GOOGL> EDA: Synopsys < $SNPS > Cadence < $CDNS > Ansys < $ANSS > Confused by some of names of the list, they might have conflated a few names like Innolight with laser source like $MTSI, Sumitomo, $SIVE, Luxnet, with the actual end module (unless there's something that's not public material or I missed)? But just for people interested in the landscape, this is a good high-level overview.
-
AI数据中心供应链严重依赖亚洲,中东冲突加剧流动性风险,稀土应成美国国安优先事项
Hi Ron,这个话题分为两个不同的部分: 1. 超大规模云服务商($ORCL、$META、$AMZN):他们的支出超过了盈利水平($GOOGL 是罕见的例外,而 $AAPL 相对而言支出并不多)。 因此市场担心这并没有转化为实质性的收入。 尤其是像 Oracle 这样负债累累的公司,降息是使其稍微可持续的重要因素(如果你看债务利息,数目非常庞大),而降息既能推动产能扩张,也能带来实质性的成本节省。 由于伊朗战争和原油价格上涨,原定2026年的降息预期已经落空。 雪上加霜的是,人们担心他们的 AI 数据中心建设供应链会中断。 如果你看看他们的 AI 数据中心部件产地: - 来自台湾 - 来自韩国 - 来自日本 这是一个全球化的供应链。 虽然美国看起来没问题,因为它在石油、液化天然气、氦气和其他半导体供应链所需的资源方面是绝缘的。 但我们的亚洲合作伙伴并不处于长期冲突中。 所以当你看到亚洲那些规模较小、专业化的公司时,它们正在苦苦挣扎。 这种影响会层层放大,一直传导到终端的亚马逊网络服务 AI 数据中心(成本上升、供应短缺)。 2. AI 整体而言:投资者对此非常乐观。 问题在于实现路径,而美国在表面上看起来被中国卡住了。 - 美国投入过多,而中国和其他国家正在蒸馏我们的最新模型(需要 KYC 端点管控),坐享其成。 - 随着 AI 项目的规模扩大,它正变得越来越依赖中国公司。 我认为这是因为美国缺乏稀土,这应该成为我们的头号国家安全优先事项。 这是制造新一代 AI 硬件、机器人供应链和太空探索所需的。 这些资源都由中国和俄罗斯控制。即使是领先的"西方"公司,我已发现它们仍然主要依赖中国/俄罗斯。 这是我们确保前沿项目安全最大的漏洞之一。 如果你看光子学(AI 硬件如何通过光加速),我们没有西方的供应链来满足需求。主要材料来源是中国的 $AXTI 和中国的 Vital 两家公司,然后流转到日本或其他公司,最后进入美国。 如果你看我们的人形机器人项目(我们如何将 AI 带入现实世界),$TSLA Optimus 的所有躯体都是在中国制造的。 美国供应链缺乏制造这些组件所需的稀土,因为成本太高。 3. 中东地区的冲突 近期下跌的一个重要原因是流动性问题。我们在阿联酋、中东的合作伙伴主要在为私募市场提供资金(想想 OpenAI),他们大量的支出流向 AI 市场。 他们也是从 Meta 到微软等科技七巨头最大的投资者之一。 如果他们的油田继续受到干扰,他们可能不得不从美国市场撤出流动性。 这既影响这些公司在 AI 上的支出能力,也会导致美国市场的抛售下跌。 如果有takeaway(要点)的话:请把稀土列为美国头号国家安全优先事项。特朗普总统已经投资了 $USAR、$MP 和几个单独的公司。 但还有许多更多至关重要的、尚未盈利的公司,它们对于制造 AI、自动驾驶汽车和机器人都是必需的。我们需要打破对中国/俄罗斯的依赖。
英文原文
Hi Ron, so there's two different parts to it: 1. Hyperscalers ( $ORCL, $META, $AMZN): They're spending more than they have with profit ( $GOOGL is the rare exception, and $AAPL isn't really spending much relatively). So markets are worried less worried it's translating into material revenue. Especially with companies like Oracle taking on immense amounts of debt, rate cuts were a large driving factor in why it's slightly more sustainable (if you look at debt interest, it's massive), and cutting rates drives both forward expansion and material amounts of savings. Due to the War in Iran, rising crude, former projections with cutting rates in 2026 are now gone. This is compounded that there are worries that their AI buildout supply chain gets disrupted. If you look at where their AI datacenter parts are made: - it comes from Taiwan - it comes from Korea - it comes from Japan It's global. While the US might look like it's fine, since it's insulated to Oil, LNG, helium, and others needed for the semiconductor supply chain. Our Asian partners are not in a long drawn out conflict. So when you look at companies in Asia that are more small and niche they are struggling. And this gets compounded tens of times until it reaches all the way down to the end Amazon Web Services AI datacenter (increased costs, lack of supply). 2. AI as a whole: Investors are very bullish on it. It's just how we get there, and America looks strangled by China optically. - America is spending way too much on it, with China and others distilling our latest models (need KYC endpoints), and reaping all the benefits. - As we scale our AI programs, it's becoming increasingly reliant on Chinese companies. I'd argue it's because America lacks the rare earths, which should be our #1 national security priority. This is needed to make new generations of AI hardware, robotic supply chains, and Space. They're all controlled by China and Russia. Even the leading "Western" companies, I've identified they're mainly relying on China/Russia still. And this is one of the biggest vulnerabilities in securing our frontier programs. If you look at photonics (how AI hardware is sped up by light), we no Western supply chain to fulfill our needs. It's $AXTI (in China), and Vital (in China) as the two main sources of materials needed to make them, that gets passed down to Japan, or other companies before they end up in US. If you look at our humanoid program (how we move AI to the real world), all the bodies of $TSLA Optimus are made in China. American supply chains lack the rare earths needed to make the components because it gets too expensive. 3. Disruption in the Middle East A large part of the recent drop is mainly due to liquidity. Our partners in the UAE, Middle East are largely funding private markets (think OpenAI) and a lot of their spend going to AI markets. They're also one of the biggest investors in Mag7 from Meta to Microsoft. If their oil fields continue getting disrupted they may have to pull out liquidity from US markets. This impacts both the amount of money these companies can spend on AI. As well as causing a drop in American markets from selling. If there's any takeway: Please make Rare Earths Amercia's #1 national security priority. President Trump already invested in $USAR, $MP, and a few individual names. But there are many more extremely critical, yet unprofitable companies that are needed to make AI, to self driving cars, to robotics. And we need to break our reliance on China / Russia.
-
希望 SIVE 保持独立,担心被竞争对手低价买下。
@Matrix_B0SS 我其实希望 $SIVE 保持独立。 只是随手写了些洗澡时冒出来的想法,觉得另一家超大客户或竞争对手可能会用区区 3.5 亿美元就控制 $MRVL 的近端光子项目。
英文原文
@Matrix_B0SS I actually hope $SIVE remains independent. Just posting some shower thoughts about how another hyperscaler or competitor could control $MRVL near term photonics program for a cheap price of $350m.
-
认为 SIVE 可能被多方盯上,也可能因持股结构保持独立。
洗澡时冒出来的一个想法…… SIVE 现在可能已经被 $QCOM、$MRVL、$AVGO、AlChip、联发科、$AMZN、$META、$MSFT 的风投部门盯上了? 既然现在大家都公开知道了,Broadcom 只要花 3.5 亿美元就能买下 Marvell 在光子学上的咽喉点(Amazon / Microsoft 的 ASIC 项目)…… 或者 Qualcomm 也可以用同样的方式收购 $SIVE,然后在收购 Alphawave 之后把激光 IP 做垂直整合。 又或者……超大规模客户会通过持有足够多股份,想办法让 $SIVE 继续保持独立?
英文原文
It’s funny a shower thought… Probably put $SIVE on the radar for $QCOM, $MRVL, $AVGO, AlChip, Mediatek, $AMZN, $META, $MSFT venture arms now? Since everyone publicly knows now Broadcom can just go and buy Marvell’s photonics chokepoint vulnerability (Amazon/Microsoft ASIC programs) for a rounding error of $350m… Or Qualcomm can do the same buying $SIVE then vertically integrate laser IP after Alphawave aquisition. Or maybe… hyperscalers figure out a way $SIVE remains independent by owning enough shares?
-
把 7n 高纯铟视作 InP 衬底的一个代理信号。
@ParadisLabs 7n 高纯未标准铟,在 SMM 上是 InP 衬底的一个不错代理指标。 不过对 CPO 放量来说,这件事复杂得多,我会看更多其他信号。
英文原文
@ParadisLabs 7n indium high purity nonstandard on SMM is a good proxy derivative for InP substrates. However, it’s much more multifaceted for CPO ramp and I’d look for other signals
-
只是分享想法,猜测 Marvell 的 CPO 供应链会不会被竞争对手收购卡脖子。
@Jornka329996 我只是分享一下洗澡时的想法,不知道人们或者市场会怎么反应。 也许 $MRVL 的风投方更愿意让他们的 CPO 供应链卡口被 $AVGO 或 AlChip 这样的竞争对手收购和控制?
英文原文
@Jornka329996 I’m just sharing my shower thoughts, don’t know how anyone or markets react. Maybe $MRVL ventures prefers their CPO supply chain chokepoint to be bought out and controlled by competitors like $AVGO or Alchip?
-
把近期机构轮动理解为 CPO/ELS 上游链条的机会。
尽管指数和巨头都在跌,比如 $META / $MSFT,但市场里依然明显有一些超额收益者。 这大概说明了机构轮动,或者说某个非常强的新兴行业已经在形成。 最明显的就是 CPO / ELS 供应链。 我个人的标的包括: $TSEM / Win Semi - 光子学里的 TSM $AAOI / $SIVE / $COHR / $LITE - 光源 $SOI / $AXTI - 衬底 $IQE - 外延片 当然也有更大的名字,比如 $MRVL、Sumitomo、$AVGO。
英文原文
There’s clearly some market outperformers despite index’s and mega caps from $META / $MSFT crashing. This probably signals institutions rotation or an extremely strong up and coming sector. The obvious one is into CPO/ELS supply chains. My personal picks were: $TSEM / Win Semi- TSM of photonics $AAOI / $SIVE / $COHR / $LITE- Light Sources $SOI / $AXTI - Substrates $IQE - Epiwafers But of course there are larger names out there like $MRVL, Sumitomo, $AVGO.
-
指出美国在InP衬底、人形机器人、高端半导体等领域严重依赖外国供应商,无法自给自足。
这在事实上并不正确,而这正是我们陷入困境的完全相同的原因。 美国完全依赖日本/中国。例如:用于AI光子学的InP(磷化铟)原料/衬底。 美国依赖中国制造人形机器人的机体。 美国通过一些关键瓶颈点依赖俄罗斯/其他对手(不会透露具体名称,因为涉及国家安全风险)。 美国依赖台湾/韩国制造高端半导体。 美国依赖欧洲制造高端机器($AXIA、$AMSL类型)或特定衬底($SOI)。 再次强调,供应链是相互关联的。这种观点正是当前政府认为自己掌握了所有牌的精确原因。 请列举一家能够供应AI所需InP衬底足够原材料的美国公司。 请列举一家今天能够完全替代$ASML的美国公司。 你做不到,因为它们还不存在。
英文原文
This is factually not true and this is the exact same reasoning why we're in this mess. US is completely reliant on Japan/China. eg. InP feedstock/substrates for AI with photonics. US is reliant on China for manufacturing bodies of humanoids. US is reliant on Russia/other enemies (won't disclose specific names since it's national security risk) through some niche chokepoints. US is reliant on Taiwan/Korea for high end semiconductors. US is reliant on Europe for high-end machines ( $AXIA, $AMSL type) or specific substrates ( $SOI ). Again, supply chain is interconnected. This type of take is the exact reason why the current administration thought it had all the cards. Name one US company that can supply enough raw feedstock required for InP substrates for AI. Name one US company that can completely replace $ASML today. You can't because they don't exist yet.
-
全球供应链高度互联,战争将摧毁从AI到消费电子的产业链,受益者仅是以色列和俄罗斯。
我仍然看多市场(目前85%多头,15%现金/对冲)。即使在战争时期,我预期从$AAOI到$LITE的许多股票仍能跑赢大盘。然而...如今世界高度互联。现任政府可能没有意识到上游供应链的严重性和脆弱性。他们可能看到$AAPL的"美国制造"。或者$TSLA的"美国制造",但如果你看看特斯拉人形机器人的供应链来源:-> 在中国。执行器(Actuator)材料来自俄罗斯的原料。材料来自加拿大的合作伙伴。-> 对于星链(Starlink),大量供应链来自台湾稳懋(Win Semi)到韩国的公司。对于OpenAI、私募股权、大型科技公司和其他公司的资金来源:-> 很大一部分来自阿联酋、中东和其他地方(随着油田被炸毁,他们可能需要撤出流动性)。持续数月的美国地面入侵造成的混乱是是非常看跌的。我最初预期的是"震撼与敬畏",美国宣称胜利,战争冲击被买回。但我不会鼓励美国走向一场旷日持久的战争。这对亚洲伙伴(韩国、日本)或我们在欧洲的盟友都没有好处。战争的主要受益者是:#1:以色列 #2:俄罗斯世界其他地区,包括美国,都被战争拖累。看起来"油价/能源价格上涨对美国有利":但它摧毁了从$NVDA到$AAPL的超大规模数据中心(Hyperscaler)所依赖的大量供应链。
英文原文
I'm still long on markets (now 85% long, now 15% cash/hedges). Even during wartime, I expect many names from $AAOI to $LITE to outperform. However... The World is largely interconnected now. The current administration might not realize the severity/fragility of upstream supply chains. They might look at $AAPL "Made in America". or $TSLA "Made in America" but if you look at the source for Tesla Humanoids: -> It's in China. Actuator materials stem from Russia feedstock. Materials come from partners in Canada. -> For Starlink, massive amount of supply chains stem from Win Semi Taiwan to companies in Korea. For where the funding from OpenAI, private equities, megacap, and others come from: -> Large part of it comes from UAE, Middle East, and others (that might need to pull out liquidity) as their oil fields get blown up. Extended disruptions (US ground invasion) for spanning multiple months, is very bearish. I was initially expecting the "Shock and Awe" and US to claim victory and for short term War shock to be bought back. But I would not encourage US to head toward a drawn out War. This especially is not good for Asian partners (SK, Japan) or our allies in Europe either. The main beneficiaries of War are: #1: Israel #2: Russia The rest of the World, including the US, are getting dragged down from War. It looks like "Oil/Energy prices going up is good for US": But it takes down a lot of supply chains that the hyperscalers from $NVDA to even $AAPL relies on.
-
对比光通信供应链各玩家对CPO光源业务的直接敞口差异
当然,这更像是直接敞口与市值的对比。 光源环境如下: 头部玩家:Lumentum (550亿美元)、Broadcom (1万亿美元)、Coherent (550亿美元)、Sumitomo (规模庞大) 大型玩家:$AAOI (约80亿美元)、Furukawa (约70亿美元)、$MTSI (约190亿美元) 新兴玩家:$INTC (约2300亿美元+)、$SIVE (2.3亿美元)、Luxnet (17亿美元) 你可以买$AVGO来获得光子学敞口,但因为它规模太大……他们的其他业务部门可能会稀释敞口。 另一方面,如果你看$SIVE,由于规模很小,它几乎是CPO光源升级的纯概念股。 所以我最终选择了最直接的敞口。 很多人对市值(MC)感到不适,所以Furukawa、$MTSI、Sumitomo这类公司可能更受机构分析师青睐。
英文原文
Sure, it's more like direct exposure vs. marketcap. The light source environment looks like this: Leading players: Lumentum ($55B), Broadcom ($1T), Coherent ($55B), Sumitomo (massive) Large Players: $AAOI (~$8B), Furukawa (~$7B), $MTSI (~$19B) Emerging players: $INTC (~$230B+), $SIVE ($230M), Luxnet ($1.7B) You can buy $AVGO for photonics exposure, but because it's so large... their other segments will probably be diluting exposure. On the other hand, if you look at $SIVE, it's almost pure-play to the light source ramp for CPO because it's so small. So I ended up picking the most direct exposure. Lot of people are uncomfortable with MC, so companies like Furukawa, $MTSI, Sumitomo might be more popular with institutional analysts.
-
把 AEHR、TSEM、AAOI 分别对应测试、代工和激光供应链。
@mi20483980476 所以这三者其实分属三个不同类别(当然有重叠),我都喜欢。 $AEHR 做测试。 $TSEM 更像是晶圆代工。 $AAOI 则更像是激光 fab -> 设计 -> 组装。 这就是一条供应链。
英文原文
@mi20483980476 So they all sit in three separate categories (with some overlap) and I like all three. $AEHR does testing. $TSEM is more like a foundry then $AAOI is more like laser fab -> design -> assembly It's a supply chain.
-
博主分享其对AI供应链瓶颈轮动的精准判断,从存储到光子学再到CPO/ELS,并引用研报分析AI投资超级周期的瓶颈解决序列和CPO商业化拐点
我认为我精准把握了机构层面的瓶颈轮动。 -> 在存储股的上涨尾声捕捉到了 $SNDK、三星、SK海力士、$MU -> 在光子学领域领先机构布局了 $AAOI、$AXTI、$LITE、$COHR -> 现在再次行动,大量加仓SiPh、ELS和CPO:$SIVE、$TSEM、$SOI、$AEHR、Win Semi以及其他 当然如果你想稳健操作: $MRVL(产能自用)、$AVGO(产能自用)、$TSM和$NOK也都有涉及,但他们是更大的玩家。直接曝险下一个超级周期才是理想的。 我仍然认为从存储到EML光收发器的增长空间巨大,但最大的繁荣出现在新架构周期起点/拐点处。 除了宏观因素干扰了一些交易外,预计资本将很快轮动到CPO/ELS供应链。 券商研报的摘要:AI基础设施投资超级周期遵循严格的"瓶颈解决"序列: 算力/GPU(2023)-> 存储/HBM(2024)-> 互联/网络(2025+) 翻译: "我们判断AI价值链的第三轮投资周期已正式启动。继GPU(2023)和HBM(2024)之后,2025年后,光互联将成为增长最快的核心环节。 2027-2028年将是关键拐点,CPO商业化、1.6T标准化和Scale-Up光互联将形成共振,结构性扩大相关行业TAM。" 一些要点: 1. CPO(共封装光学)正从实验室走向商业化量产 2. 2027-2028年是重大结构性拐点(现在在2026年提前布局是个好主意,从测试端$AEHR到ELS端$SIVE或封装端$POET) 3. 光学部件、材料和测试设备的市场总规模(TAM)预计将结构性扩大3到5倍(我认为这还是保守估计)。
英文原文
I think I nailed the institutional bottleneck rotation. -> Caught the tail end of memory name rise with $SNDK, Samsung, SK Hynix, $MU -> Frontran institutions with photonics with names like $AAOI, $AXTI $LITE, $COHR. -> Doing it again now by adding heavily toward SiPh, ELS, and CPO: $SIVE, $TSEM, $SOI, $AEHR, Win Semi, and others. Of course if you want to play it safe: $MRVL (captive), $AVGO (captive), $TSM, and $NOK all do it as well, but they're larger players. And getting direct exposure to the next supercycle is ideal. I still think there's tons of room to grow for memory to EML optical transceivers, but the largest boom is at the start/inflection point of a new architectural cycle. Macro messing up some trades aside, expecting capital rotation soon toward CPO / ELS supply chains. TLDR from analyst note: The AI infrastructure investment supercycle follows a strict "bottleneck resolution" sequence: Compute/GPUs (2023) -> Memory/HBM (2024) -> Interconnect/Networking (2025+). Translation: "We judge that the third investment cycle of the AI value chain has officially begun. Following GPUs (2023) and HBM (2024), post-2025, optical interconnects will become the fastest-growing core segment. 2027–2028 will be a critical inflection point where CPO commercialization. 1.6T standardization, and Scale-Up optical transitions align, structurally expanding the TAM of related industries" Some takeaways: 1. CPO (Co-Packaged Optics) is moving from the lab to commercial mass production 2. 2027-2028 is the major structural inflection point (good idea to frontrun this now in 2026 from testing with $AEHR to ELS with $SIVE or packaging with $POET). 3. The Total Addressable Market (TAM) for optical components, materials, and testing equipment is expected to structurally expand 3x to 5x (I think this is sandbagging a bit).
-
Mirae分析师梳理CPO产业格局,作者认为遗漏上游公司但框架有参考价值
CPO产业格局 Mirae分析师笔记: 横向扩展: CPO ASIC:$AVGO、$MRVL 光收发器:$COHR、$LITE、Innolight DSP/PAM4:$AVGO、$MRVL 相干DCI:$CIEN、$NOK OCS设备:iPronics、Polatis 光缆/光纤:$GLW、Prysmian、Furukawa HCF:$LITE、OFS DCI相干:Ciena、Nokia、Huawei 光放大器:$LITE、$COHR OCS网关:KDDI 纵向扩展: SiPh代工:imec、$GFS、$TSM SiPh调制器:$NVDA(内部MRM)、$INTC ELS:NTT、Furukawa、$LITE、$COHR 太赫兹互联:研发阶段? CPO测试:「新玩家不断进入」 微透镜/光学系统:「新玩家不断进入」 总结:纵向扩展的CPO即将到来。认为分析师笔记遗漏了一些上游公司名字,并将ELS与光源混为一谈。 但作为高层视角,能看到他们认为的领先玩家还是有参考价值的。
英文原文
CPO Landscape Mirae Analyst Note: Scale-Across: CPO ASIC: $AVGO, $MRVL Optical Transceiver: $COHR, $LITE, Innolight DSP/PAM4: $AVGO, $MRVL Coherent DCI: $CIEN, $NOK OCS Equipment: iPronics, Polatis Optical Cable / Fiber: $GLW, Prysmian, Furukawa HCF: $LITE, OFS DCI Coherent: Ciena, Nokia, Huawei Optical Amplifier: $LITE, $COHR OCS Gateway: KDDI Scale-Up: SiPh Foundry: imec, $GFS, $TSM SiPh Modulator: $NVDA (in-house MRM), $INTC ELS: NTT, Furukawa, $LITE, $COHR THz Interconnect: R&D Stage? CPO Test: "Expanding entry of new players" Micro Lens / Optical Systems: "Expanding entry of new players" TLDR: Scale Up CPO is coming next. Think the analyst note missed a bunch of upstream names and conflated ELS with light source. But it's helpful to see who they think the leading players are as a very high-level view.
-
看好HIMS与TEM结合布局个性化AI医疗,认为市场潜力巨大
个性化AI医疗很可能是一个巨大的市场。没有人想早死。将 $HIMS 与 $TEM 结合具有巨大潜力。这里有一个GitLab CEO的例子:2023年,Sijbrandij被诊断出骨肉瘤;2024年7月,他宣布癌症复发了,尽管进行了多种治疗;到2024年底,他已经用尽了标准医疗方案的全部选择。没有选择接受死亡,Sijbrandij决定像构建GitLab一样将他的癌症作为一个工程问题来处理。他开始向ChatGPT输入大量的个人医疗数据,包括扫描结果、血液检测结果和组织样本数据。ChatGPT和AI在识别突变后提出了新的靶向治疗方案,然后确定如何攻击这些突变。在此基础上,他申请了未获批准、处于试验阶段的药物并获得了用药。这个方法奏效了。Sijbrandij用AI阻止并扭转了晚期预判。我们已经看到了这个成功案例( $TEM 明确列为他用于肿瘤批量RNA测序的合作伙伴)。最近也有一个人用AI治愈了自己的狗。如果你能够将 $HIMS(直销医疗分销网络)和 $TEM(个性化测序和AI治疗)结合成一个可扩展的产品,潜力是巨大的(尽管极具挑战性)。全球人口都是潜在市场。这类似于 $AXTI InP(磷化铟)在超大规模数据中心的情况:人们会不惜代价避免死亡(这就是瓶颈)。这只是未来医疗发展方向的一个缩影。
英文原文
Personalized AI medicine will likely be huge. Nobody wants to die early. Combining $HIMS to $TEM has immense potential. Here's just one example with Gitlab's CEO: > In 2023, Sijbrandij was diagnosed with osteosarcoma > In July 2024, he announced the cancer had returned, despite multiple treatments > By late 2024, he had reached the absolute end of standard of care medical options. > Rather than accepting death, Sijbrandij decided to treat his cancer the same way he built GitLab as an engineering problem > He began feeding ChatGPT massive amounts of his personal medical data from scans, blood test results, and tissue sample data. > ChatGPT and AI made novel targeted treatments after identifying mutations, then how to attack them > From there, he applied for unapproved, experimental drug and received it > The approach worked. Sijbrandij managed to stall and turn around a terminal prognosis with AI. We've seen this success story with Gitlab's CEO ( $TEM was explicitly listed the partner he used for the bulk RNA sequencing of his tumor ). As well as someone curing their own dog with AI recently. The potential if you can combine $HIMS (DTC distribution network) and $TEM (Personalized sequencing and AI treatment) into a scalable product is massive (though extremely challenging). The entire human population is the TAM. And similar to the $AXTI InP situation with hyperscalers: People will pay anything to not die (bottleneck). This is just a hint into where the future of healthcare is heading.