供应链分析
产业链结构、上下游与瓶颈环节研究 · 共 1309 条 · 滚到底自动加载更多
-
欣讯通过鸿海垂直整合间接成为英伟达CPO供应链受益者，被市场忽视且估值偏低。
我很确定像高盛这样的机构在$15亿估值时错过了欣讯(6451)的CPO封装/测试/组装业务。这就是为什么我非常看好它，认为它是一个完全走后门的、隐藏的英伟达CPO产能提升的主要受益者。鸿海是其供应链中的主要合作伙伴，包括$TSM和$ASX。但是……欣讯是鸿海的光学部门，并承接其内部的全部光学和先进封装业务。所以它们没有在任何地方被明确列出，也没有与英伟达的直接合同（但鸿海有）。但很可能通过鸿海垂直整合吸收了英伟达的CPO和其他产能。
英文原文
Pretty sure institutions like GS missed Shunsin (6451) at $1.65B for CPO packaging/test/assembly. Which is why I'm very bullish on it as a completely backdoored, hidden + major beneficiary of $NVDA CPO ramp. Foxconn is a major supply chain partner among $TSM and $ASX. But… Shunsin is Foxconn's optical arm, and captures their captive optical and advanced packaging volume. So they're not explicitly listed anywhere or have direct contracts with Nvidia (but Foxconn does) But likely soaks up Nvidia CPO + other volumes through Foxconn vertical integration.
-
确认 Lightelligence / Lightmatter 也有新消息
@michael_rigoni 不，这确实是关于 Lightelligence 的新信息发现，尤其是考虑到他们最近的成功。 Lightmatter 也是一样。
英文原文
@michael_rigoni No this is definitely new information discovery with Lightelligence. Especially in light of their recent success. Same with Lightmatter.
-
认为 OE Solutions 还太早，几乎没开始抽样
@UmidKD OE Solutions 的 sampling 才刚刚开始。ELS 线还没有客户，也没有 active qualification。 大概还早了 8 个月，hyperscaler 供应商甚至未必会决定去认证他们，但市场现在就是在把所有光学相关的东西都往上抬。
英文原文
@UmidKD Sampling has barely started for OE Solutions. No customers or active qualification for their ELS line. Probably 8 months too early and hyperscaler suppliers might not even decide to qualify them, but markets are moving up anything optical related.
-
认为 SIVE 在 GFS 生态图里被列入极其看多
没错，市场可能漏掉了 $SIVE + $GFS SCALE 这件事。 $SIVE 被列为 $GFS 生态图里两个公开 laser 供应商之一，这对 Sivers 来说是极度看多的信号。 每当像 $AMD 这种公司使用 $GFS 做 CPO、再借助 GlobalFoundries 搭建光学方案时， Sivers 大概率就是他们的光源，给 $AMD -> hyperscaler 这条链路供能。 带着 CPO 的 hyperscaler buildout 最终会流到这家只有 17 亿美元市值的小型 laser 光源公司 $SIVE 身上。
英文原文
Yes, markets might have missed this with $SIVE + $GFS SCALE. The inclusion of $SIVE as 1 of 2 public laser suppliers in $GFS ecosystem image presentation is a highly bullish indicator for Sivers. Whenever semis like $AMD (using $GFS for CPO), builds their optical solutions with GlobalFoundries. Sivers is the likely light source for their buildout and powers companies like $AMD -> Hyperscalers. The hyperscaler buildout with CPO flows down to this tiny $1.7B MC laser light source in $SIVE.
-
认为 SIVE 也是 Lightelligence 的潜在未披露激光供应商
有个很有意思的点：$SIVE / $SIVEF 也很可能是 Lightelligence 的未披露激光供应商。 感谢 Plaskpojen 做的 OSINT 发现。 https://t.co/sj2J0b9ycN
英文原文
Fun thing to note: $SIVE / $SIVEF is the likely undisclosed laser supplier for Lightelligence as well. Credit to: Plaskpojen for the OSINT find. https://t.co/sj2J0b9ycN
-
说明 MSSCorp 在 CPO 检测里几乎是功能性垄断
@GCV_geoff 我觉得它还在交易中。
英文原文
@GCV_geoff It's still trading I think.
-
解释 MSSCorp 的功能性垄断地位
我确实说过，MSSCORP 很可能在 CPO 检测上有功能性垄断。 这几乎没有绕开的办法，而且对 $NVDA / $TSM 的良率来说都很关键。 他们自己也说过“the c...”
英文原文
I did say MSSCORP likely has a functional monopoly for CPO over inspection. Not really a way around it and it's critical for $NVDA / $TSM yields as CPO ramps. They did say "the company’s goal is to seize a 90 percent share of the CPO inspection market" Should have been the biggest indicator.
-
解读 GS 光学 TAM 9 倍增长的 CPO 细节
Goldman Sachs 关于 Optical Networking 的预测。 整体 TAM 会增长 9 倍： 从 150 亿美元到 1540 亿美元。 CPO 占其中 910 亿美元。 这到底是什么情况？ 对，这也是为什么像 $LITE、$SIVE 和 $SOI 这类名字值得关注。
英文原文
Goldman Sachs Projections on Optical Networking. Aggregate TAM would increase 9x: From $15B to $154B. CPO contributes $91B. What the actual? Yep... And another confirmation why I'm so incredibly bullish on CPO names like $SIVE (laser), MSSCorps (yields), or Shunsin (packaging). This is frontrunning the next supercycle. I don't think people realize how wild these projections are yet.
-
SIVE 是激光瓶颈供应商，可能供货多家CPO厂商却未被机构发掘。
$SIVE 处于高盛所说的具有显著每股收益(EPS)上行空间的激光瓶颈环节。 但它们仍未被机构发现，尽管它们可能与 $AMD 共封装光学(CPO)、$MRVL CPO、JBL 1.6T、Lightmatter、Ayar、$POET 等公司有关联。 Ayar 被提及了，但奇怪的是他们没有找到它们的上游激光供应商。
英文原文
$SIVE is in the laser chokepoint that Goldman says has significant EPS upside. But they're undiscovered to institutions still, despite their likely mapping to $AMD CPO, $MRVL CPO, $JBL 1.6T, Lightmatter, Ayar, $POET, and others. Ayar got mentioned though, but strange they didn't find their upstream laser supplier.
-
认为 CW/EML laser chokepoint 上行空间最大
@jojowamikko 我也有我自己的看法。 但我同意 CW / EML laser chokepoint 是上行空间最大的之一。 他们漏掉了很多名字，我会更关注这些，因为它们相对更纯。
英文原文
@jojowamikko I have my own opinions. But agree that the CW/EML laser chokepoint has one of the highest upside. They missed a ton of names, and I'd focus on those since they're relatively "undiscovered" to institutions.
-
高盛看好AI驱动光通信供应链，多环节供应商名单汇总
如果你对高盛的报告感兴趣： 他们预期以下领域会出现"显著的每股收益(EPS)增长"： 1. 光学模块和引擎 2. 连续波(CW)激光器和电吸收调制激光器(EML) 3. 印制电路板(PCB)/覆铜板(CCL)制造商 以下是这份超详细清单： 激光器/光源(硅光子SiPh)：Sumitomo、Furukawa(5801)、Landmark(3081)、YJ Semi、$LITE、$COHR、$AVGO、Etern、曙光光电(https://t.co/hrQbVE0QlT)、世嘉光子(https://t.co/HvJVCM7BMm)、应用光电(AAOI)、Accelink、三菱电机(6503) 激光器/光源(EML)：Lumentum(LITE)、Coherent(COHR)、博通(AVGO)、YJ Semi、永恒光电(Etern)、住友电工(5802)、曙光光电、世嘉光子、Source Photonics(私营)、三菱电机(6503)、Accelink、古河电气(5801) 无源光学组件-耦合器/分路器：FOCI(3363)、Senko(私营)、AFR、Everprox、Lumentum(LITE)、Coherent(COHR) 无源光学组件-滤波器：世嘉光子、住友电工(5802)、Lumentum(LITE)、Coherent(COHR)、唯亚威(VIAV)、FOCI(3363) 无源光学组件-衰减器：Everprox、Accelink、FOCI(3363)、Lumentum(LITE) 无源光学组件-波分复用(WDM)：AFR、Everprox、Accelink、Ciena(CIEN)、TFC Optical、Lumentum(LITE)、FOCI(3363)
英文原文
If you're curious about Goldman's Report: They expect "significant EPS upside" among: 1. Optical modules and engines 2. CW lasers and EMLs 3. PCB/CCL manufacturers Here's the massive cheat sheet: Laser / Light Sources (SiPh): Sumitomo, Furukawa (5801), Landmark (3081), YJ Semi, $LITE, $COHR $AVGO, Etern (https://t.co/Rgl8i5e7N4), Everbright Photonics (https://t.co/hrQbVE0QlT), Shijia Photons (https://t.co/HvJVCM7BMm), Applied Optoelectronics (AAOI), Accelink (https://t.co/OjlKPLtkA4), Mitsubishi (6503.T) Laser / Light Sources (EML): Lumentum (LITE), Coherent (COHR), Broadcom (AVGO), YJ Semi (https://t.co/qWTYoY2y6z), Etern (https://t.co/Rgl8i5e7N4), Sumitomo Electric (5802.T), Everbright Photonics (https://t.co/hrQbVE0QlT), Shijia Photons (https://t.co/HvJVCM7BMm), Source Photonics (Private), Mitsubishi (6503.T), Accelink (https://t.co/OjlKPLtkA4), Furukawa (5801.T) Passive Optical Components - Couplers / Splitters: FOCI (3363.TWO), Senko (Private), AFR (https://t.co/v8bWvYaSCp), Everprox (https://t.co/PgwoMVPleI), Lumentum (LITE), Coherent (COHR) Passive Optical Components - Filters: Shijia Photons (https://t.co/HvJVCM7BMm), Sumitomo Electric (5802.T), Lumentum (LITE), Coherent (COHR), Viavi (VIAV), FOCI (3363.TWO) Passive Optical Components - Attenuators: Everprox (https://t.co/PgwoMVPleI), Accelink (https://t.co/OjlKPLtkA4), FOCI (3363.TWO), Lumentum (LITE) Passive Optical Components - WDM: AFR (https://t.co/v8bWvYaSCp), Everprox (https://t.co/PgwoMVPleI), Accelink (https://t.co/OjlKPLtkA4), Ciena (CIEN), TFC Optical (https://t.co/fjsjeuQxcq), Lumentum (LITE), FOCI (3363.TWO)
-
认为激光器环节是CPO供应链最被低估的瓶颈，拥有InP垂直整合等多项期权价值。
最近关于"共封装光学(CPO)供应链"的讨论很多，什么是最被低估的环节？让我发表一下我的观点： 激光器瓶颈（$LITE、$SIVE、$COHR）是目前接触光子学和CPO的最佳途径。 因为存在大量期权价值，这给了Lumentum等公司溢价空间。 具体体现在： -> 随着可插拔光收发器的出现，下游IP总可寻址市场(TAM)不断扩展 -> InP激光器晶圆厂的垂直整合，而非无晶圆厂(fabless)模式 -> 多道工序的垂直整合，例如测试甚至封装 把你自己想象成在经营一家公司，并快速推进到$100B或$300B的规模。 不要仅仅基于预测的收入和利润率来做电子表格模型。
英文原文
There's a lot of discussion around "CPO supply chains" and what's the most undervalued. Let me put my opinion out there: The laser chokepoint ( $LITE, $SIVE, $COHR ) is by far the best exposure to photonics and CPO. Because there's so much optionality that gave companies like Lumentum the premiums. With: -> downstream IP TAM expansion as seen with pluggable optical transceivers -> vertical integration of InP laser fabs rather than fabless. -> vertical integration of multiple processes eg. testing or even assembly. Treat it like you're running your own company and speedrunning it to $100B or $300B. Not just spreadsheet modeling off projected revenue/margins.
-
Sivers成为CPO关键激光供应商，NVDA引发的产能瓶颈正驱动AMD等厂商争抢上游资源。
看到$SIVE成为CPO(共封装光学)的Tier 1激光供应商，这真的很疯狂。 这是我的预测/猜测与大致映射： $NVDA用$COHR、$LITE的产能（在他们新的20亿美元以上支出计划之后）。 早期2025年EML用了同样的手法，导致了今天的瓶颈。 现在，$AMD/超大规模云商正在争抢上游激光供应商。 因此，$SIVE + Win / $GFS成为可能的主要途径。 你可以从以下看出： -> $MVL CPO通过Celestial（Nvidia签了协议，但他们没有激光器） -> $AMD CPO -> Ayar -> $POET -> Lightmatter -> 其他项目（如Jabil 1.6T） 因此，Sivers/Win成为Tier 1尖端关键独立激光供应商。 这一点的一些暗示： 1. $GFS在其生态系统中将$SIVE / $LITE列为仅有的两家公开激光供应商。 2. Ayar将$LITE / $MTSI从其网站上移除，并将$SIVE提升为主要激光供应商。 所以其他所有人最终都选择了Sivers，因为$COHR / $LITE已经完全分配。 我的猜测是，随着架构标准化，许多二级供应商也会抢占溢出市场份额。 明年年初，考虑到$NVDA的瓶颈，这场争夺关键环节的竞争将异常激烈。 而这家瑞典小型公司$1.2B规模的$SIVE将处于这场竞争的中心。
英文原文
It's pretty insane to see $SIVE become a Tier 1 laser supplier for CPO. This is my prediction/guess with est. mapping: $NVDA ate up all the capacity with $COHR, $LITE after their new $2B+ spending spree. Same playbook with EML early 2025, causing the bottleneck seen today. Now, $AMD / hyperscalers are scambling for upstream laser suppliers. Hence why $SIVE + Win / $GFS became likely primary route to go down. You can see this with: -> $MVL CPO through Celestial (Nvidia signed a deal, but they don't have lasers) -> $AMD CPO -> Ayar -> $POET -> Lightmatter -> and other programs (eg. Jabil 1.6T) As a result, Sivers/Win emerged as the Tier 1 bleeding edge + critical independent laser supplier. And there's hints for this when: 1. $GFS listed $SIVE / $LITE as the two only public laser suppliers in their ecosystem. 2. Ayar removed $LITE / $MTSI off their website and elevated $SIVE to their primary laser supplier. So everyone else ended up going with Sivers since $COHR / $LITE are fully allocated. My guess is that a lot of the secondary suppliers also capture overflow as architectures standardize. But scramble for this chokepoint will be insane early next year given $NVDA bottleneck. And a small $1.2B Swedish company in $SIVE will be in the center of it.
-
@MontagueCap CPO（共封装光学）一切都还处于极早期阶段。所有像$LITE这样的光学公司，从可插拔光模块超级周期发展至今，现在已达到数百亿美元规模。 CPO的规模将远超大量现有预测，而像$SIVE、MSSCorps这样的公司目前还在10亿美元规模左右。
英文原文
@MontagueCap Anything CPO is EXTREMELY early. All the optical companies from the pluggable supercycles like $LITE are now in the tens of billions. CPO will dwarf a ton of projections, and companies like $SIVE, MSSCorps are still in the $1B range.
-
CPO市场将迎抛物线式增长，是最佳早期布局机遇
以防你们想知道我为什么对CPO（共封装光学）如此看好。比如$SIVE（激光器）、Shunsin（封装）、MSSCorps（良率）、Win Semi/$TSEM（晶圆代工）。\n\n"CPO市场预计从2026年到2030年将以142%的复合年增长率强劲增长（不含ELS）"\n\n"规模扩展CPO细分市场预计将在2030年前超过规模扩展应用，成为主导市场。\n\n未来几年将呈现近乎抛物线式的增长。\n\n许多像Sivers这样的厂商之前对800G可插拔光模块几乎没有实质性敞口，但却是CPO领域最前沿的激光器供应商领导者。\n\n这是由$NVDA和$AVGO驱动的下一代光学超级周期中最好且最早的机遇之一。
英文原文
Just in case you’re wondering why I’m so bullish on CPO. Like $SIVE (Lasers), Shunsin (Packaging), MSSCorps (Yields), Win Semi / $TSEM (Foundry). “The CPO market is projected to grow sharply by a 142% CAGR from 2026 to 2030 (excl. ELS)” “The scale-up CPO segment is projected to surpass scale-out applications before 2030 and become the dominant market. You have almost parabolic growth over the next few years. With many players like Sivers having no material exposure to previous 800g pluggable optics but are the bleeding edge leaders of CPO as the laser supplier. This is one of the best and earliest opportunities of the next optical supercycle for an architecture driven by $NVDA and $AVGO.
-
分析CPO供应链中三家最有可能实现3倍涨幅的公司：SIVE(MC约$1.3B)、MSSCORP($1.2B)、Auros($210M)。
至于3倍涨幅的股票，在这些水平上我认为： 1. SIVE 2. MSSCORP (6830) 3. Auros (322310) 是我最看好的猜测。以下是我的思路： 1. SIVE：我真心认为他们明年能达到100亿美元以上的市值，他们是CPO激光器领域的绝对前沿，与LITE和COHR并列。 按130亿美元市值计算...关于可能的版图映射： 光子学领域：AMD CPO、MRVL Celestial CPO、JBL 1.6T、Lightmatter、Ayar、ALChip、GUC、O-Net (ELS)、POET。 太空与国防领域：通过YSS、RTX、ERIC、Bae Systems实现的Golden Dome项目。 硅光子学领域：AAPL（Apple Watch）。 这已经是非常多的客户了，而且还在持续增加。 他们总是可以通过IP收购向下游拓展TAM，或垂直整合来快速追赶LITE的600亿美元市值，一旦获得更多融资的话。 2. MSSCORP (6830)：CPO检测领域的垄断者，约12亿美元市值。 对CPO良率的100%垄断，TSM、AMAT、NVDA、LCRX、INTC等很可能都是客户。 "公司的目标是占据CPO检测市场90%的份额" 这基本意味着100%，只是他们不想引发反垄断担忧。如果他们守住垄断地位且CPO产能增加，很容易从12亿美元涨到50-90亿美元。 3. Auros (322310)：三星/SK海力士的供应商，约2.1亿美元市值，专注于混合键合(Hybrid Bonding)量测。 基本上专注于两类产品： -> HBM4 / HBM4e / HBM5周期，这是KLA此前在IR量测领域拥有垄断地位的市场。 ---> 目前正在三星工厂进行认证，下半年预计量能爬坡。SK海力士在升级到混合键合时很可能也会进行认证。 -> 薄膜厚度测量。 ---> 正在获得"主要本土芯片制造商"（三星或SK海力士之一）的认证，目标今年实现批量供应。 他们已经开发了十多年，终于要在下半年从数年的认证期转向两类产品的量能爬坡。 如果他们转型到量能爬坡阶段，似乎非常有希望在3倍至6.3亿美元左右，这在韩国市场中像是一颗未被发现的宝石。 当然，我不确定这些公司最终会如何发展，这全都是推测性的，但这是高置信度的供应链映射。 但凭我的直觉，这三只我自己持有的股票在这个水平上最有可能。
英文原文
As for 3x brrrs these levels: 1. $SIVE 2. MSSCORP (6830) 3. Auros (322310) Are my best guesses. Here's my thought process: 1. $SIVE: I genuinely do see them being $10B+ next year, they're the literal bleeding edge for CPO lasers alongside $LITE and $COHR. At a $1.3B MC... For likely mapping: Photonics: $AMD CPO, $MRVL Celestial CPO, $JBL 1.6T, Lightmatter, Ayar, ALChip, GUC, O-Net (ELS), $POET. For Space + Defense: Golden Dome via $YSS, $RTX / $ERIC / Bae Systems. Silicon Photonics: $AAPL (Apple Watches). This is just a stupid amount of customers and it's still increasing. They can always TAM expansion downstream through IP acquisitions or vertically integrate to speedrun $LITE's $60B MC one day once they get more funding. 2. MSSCORP (6830): CPO monopoly over inspection at ~$1.2B. 100% monopoly over CPO yields, $TSM, $AMAT, $NVDA, $LCRX, $INTC, and others are all likely customers. "The company’s goal is to seize a 90 percent share of the CPO inspection market" This basically means 100%, they just don't want antitrust. If they defend their monopoly and CPO ramps, can easily see this worth ~$5B-$9B from $1.2B 3. Auros (322310): Samsung / SK Hynix supplier at ~$210M for Hybrid Bonding Metrology. Basically pure play on two products: -> HBM4 / HBM4e / HBM5 cycles, that $KLA had a monoply over for IR metrology. ---> Getting qualified now likely in Samsung factories, H2 volume ramp est. Sk Hynix likely qualifying too when they upgrade to hybrid bonding. -> Thin-film thickness measurement. ---> Getting qualified now, with "major domestic chipmaker" (either Samsung/Sk hynix), targets mass supply this year. They've been developing for the past decade, only to volume ramp two products from years of qualification H2 this year. Seems extremely likely to 3x to $630M if they switch to volume ramp, feels like an undiscovered gem in the Korean market? Of course, not sure how they play out and this is all speculative but high confidence supply chain mapping. But off the top of my head these three that I own are the most likely ones at this level.
-
AXTI 的上游与垂直一体化风险被低估
以当前估值推动 $AXTI 的主要是机构；做空它的散户很可能会大错特错。 如果 AXT 因出口管制而下跌，整个光子业务扩张也会受影响。（例如 AXT -> IQE -> Macom/lite）而且更值得担心的是别的问题。 如果把 AXT 10 多个上游合资企业单独拆成一家纯西方商社，它从战略角度看，价值可能会比 5N 还高。 另外，营收数字有误导性，因为这些是 TTM 合同，不是新签合同。像 7n 高纯铟这类上游材料，相比去年价格已经大涨。 这条供应链是垂直整合的，所以可能会像 NAND 一样，在有限基板下出现涨价；但不同于住友，AXT 除了少数环节（比如红磷）之外几乎包揽了整条供应链。 所以它比你想的更危险。
英文原文
It’s majority institutions moving $AXTI at these valuations, retail shorts might turn out terribly wrong. If AXT goes down from export controls the photonics buildout does too. (Eg. AXT -> IQE -> Macom/lite) and there are bigger problems to worry about. If you just split out AXT’s 10+ upstream JVs into one merchant Western company it would probably be worth more strategically than 5N. Also revenue numbers are misleading since these are TTM contracts, not new ones signed. All the upstream materials like 7n high purity indium has spiked immensely since last year. It’s all vertically integrated so there might be a NAND style price hike off limited substrates, but unlike Sumitomo, AXT does the entire supply chain minus a few like red phosphorus. So it’s more dangerous than you think.
-
不建议做空 AXTI
我不是说 $AXTI 现在是值得买入的标的。 我只是说别做空它，因为它们（以及中国）在整个美国超大规模云厂商所需流程里，实际上处于功能性垄断地位。 而且如果他们通过涨价进行反制，事情可能会变得非常糟。
英文原文
@bc1aa1fa3932407 I'm not saying $AXTI is a good buy at these levels. I'm just saying not to short it because they (and China) are a functional monopoly over this entire process required by US hyperscalers. And it could go very wrong if they weaponize through price hikes.
-
AXTI 不只是 InP 基板那么简单
散户空头真的不该碰 $AXTI。 我一直看到有人把 $AXTI 误解成只有 “InP 基板”，总是拿住友来对比，但我一直没出声。 这是错的。 它们实际上掌握了整个 InP 供应链 40% 以上，这也是机构愿意给它们融资的原因。 它们握有很多卡点，不只是 InP 基板。 正如海湾地区那样，削减大约 20% 的供应就会是灾难；如果再叠加另外 3-4 个卡点，后果更严重。 -> 所有原材料（铟、镓、锗）。 Vital 和 $AXTI（自有、金美）与中国公司一起构成寡头；而西方的 5N 等只占很小一部分份额。 由于中国（Vital、AXT）掌握多数份额。 -> pBN 坩埚卡点： $AXTI 也是另一个巨大的卡点（博宇）。还有信越/摩根等少数玩家，但这又是精炼层面的一个瓶颈。 InP 基板： $AXT / 住友在这里是双寡头，JX、$COHR 只占很小比例。 这里面还有很多其他工艺（比如红磷）处于中间环节（例如日本化学/理沙）。 这是一个材料问题。 $COHR 不能在没有 AXT 控制的所有上游材料和精炼环节的情况下，直接做出 InP 基板。 如果 $LITE 的 CEO 因为 InP 基板睡不着觉，那么 50 亿美元估值对维持西方超大规模云厂商的建设上线来说根本不算什么。 日本的上游材料正在被出口管制，而 AXT 把这些都牢牢握在手里。 所以总的来说：我看到很多空头把事情混同成 “只是 InP 基板” 了。 实际上是所有铟、镓 -> 精炼加工 -> pBN 坩埚，再到最终由它们与住友共同完成的 InP 基板生产过程。 如果 Vital 和多个上游卡点停止供货，下游西方 InP 基板厂商就会受影响。 AXTI 会变成一个小型垄断。 估值来自掌控整条供应链，而不只是连机构分析师都搞不懂的 “InP 基板制造”。
英文原文
Retail short sellers should really not touch $AXTI. I keep seeing people misunderstand $AXTI as only "InP Substrate" with Sumitomo, over and over but just never made a comment. This is wrong. They are literally 40%+ of the InP supply chain, which is why institutions are funding them. They hold MANY chokepoints, not just InP Substrates. As seen with the Gulf, removing ~20% of a supply is catastrophic, but do that with 3-4 other chokepoints. -> All the raw materials (indium, gallium, germanium). Vital and $AXTI (captive, JinMei) are the duopoly alongside Chinese companies. With Western 5N and others only holding a small fractional share. Since China (Vital, AXT) controls majority. -> pBN Crucible chokepoints: $AXTI is a another massive chokepoint (BoYu). there's a few other players like ShinEtsu/Morgan/etc, but this is another bottleneck in refinery layers. InP substrates: $AXT / Sumitomo are duopoly here, with JX, $COHR holding just a tiny percentage. There's many other processes (eg. Red Phosphorus) here too (eg. Nippon Chemical/Rasa) in the middle. This is a materials problem. $COHR can't just make InP substrates without all the upstream materials + refineries that $AXT controls. If $LITE CEO can't sleep at night because of InP substrates, $5B valuations mean nothing to the Western hyperscaler buildout stay online. Japan is actively getting export controlled on upstream materials, while AXT holds them all captive. So TLDR: It's not just "InP Substrates" that i see a lot of short sellers conflating. It's all the indium, gallium -> refinery processing -> pBN Crucible that goes into the final duopoly InP substrate creation process they share with Sumitomo. if Vital and multiple chokepoints upstream stops shipping, downstream to the Western InP substate makers. AXTI becomes a mini-monopoly. The valuation comes from owning the entire supply chain, not just "InP substrate creation" that even institutional analysts dont understand.
-
机构开始买入光子链条
@ravietweet 摩根士丹利现在最终持有了 $SOI 6.5% 的股份，这也是我另一只硅光多头。 我就是要把 AI 的结构性卡点展示给散户看，赶在机构理解之前…… Point72、$MTSI 和其他机构在我发出那篇 thesis 之后，也买了 $IQE 的流通股。
英文原文
@ravietweet Morgan Stanley now ended up holding 6.5% of $SOI, which was my other silicon photonics long. I’m do show retail the structural chokepoints in AI before institutions figure it out… Point72, $MTSI and others ended up buying $IQE float after my thesis post too.
-
西方机构开始关注 SIVE
有个有意思的事实：我不认为任何大型美国机构之前持有过 $SIVE / $SIVEF，因为它原本主要是瑞典本土散户持有的股票。 另外，美国基金也有一些投资限制，阻止它们入场（例如纳斯达克上市 / 10 亿美元市值）。 但如今情况已经变了，原因包括欧洲媒体把本地持有人洗出来、Sivers 即将登陆纳斯达克，以及市值突破 10 亿美元以上。 而且 13 亿美元对超大规模云厂商和美国机构来说真的只是零钱。 我很有把握他们很多都会想要接触 CPO、硅光和激光领域，因为全世界能做这类东西的公司本来就很少，比如 $LITE、$COHR 和 $MTSI。 尤其是……一家具备 CHIPS Act 背景的激光供应商，可能会出现在 $AMD、$MRVL、$AAPL、Lightmatter、Ayar、AlChip、GUC、O-Net 以及其他像 Golden Dome 这样的供应链里。 这部分也有博弈论因素，是为了积累股份。 因为现在美西方散户已经持有了相当大一部分流通股（来自瑞典卖盘）。 而美国机构现在也想要敞口。
英文原文
Fun fact: don’t think any large US institution had positions in $SIVE / $SIVEF since it was a majority owned local Swedish retail stock. There’s also fund mandates preventing US funds from entering (eg. Nasdaq listing / $1B MC) That’s definitely changed now thanks to European media shaking out the local holders, Sivers upcoming NASDAQ listing, and hitting $1B+ MC. Again $1.3B is literal spare change for hyperscalers and US institutions. And I have high confidence many would want exposure to CPO, silicon photonics and lasers, since there’s only a few in the entire world like $LITE, $COHR, and $MTSI. Especially… a CHIPS Act laser supplier likely in $AMD, $MRVL, $AAPL, Lightmatter, Ayar, AlChip, GUC, O-Net and other supply chains like the Golden Dome. Part of it is game theory for share accumulation. Since US/Western retail now owns a large part of the float (off of the Swedish sellers). And US institutions now want exposure.
-
SIVEF 因 Golden Dome 叙事走强
如果你想知道为什么 $SIVEF 今天又涨得更多。 $SIVE 作为 CHIPS Act 受益者： 现在很可能在为 Golden Dome 提供支持。 作为上游半导体供应商，因为 Sivers 的主客户 ALLSPACE 被 $YSS 收购了。 York 恰好又是一家国防主承包商，和太空军、太空发展局、国防部以及 Golden Dome 都有关联。 所以，Sivers 这家小小的瑞典公司，基本上通过一条后门，接入了美国太空建设链条。
英文原文
If you’re curious why $SIVEF is up even more today. $SIVE, as a CHIPS act recipient: Now likely powers the Golden Dome. As the upstream semi supplier since Sivers’ lead customer ALLSPACE got acquired by $YSS. York happens to be a national defense prime contractor with links to the Space Force, Space Development Agency, DoD, and Golden Dome. So Sivers, a mini Swedish company, basically got a backdoor to powering the US space buildout.
-
LPK 是卡点而非纯投机
@jpm7019 $LPK 不是那样。按开发进度来看，$ALMU 更接近“投机”这个定义。 $LPK 已经主导了一个卡点，而且几乎所有半导体公司都在做认证。 现在只是猜这些半导体公司的量产爬坡会是什么样，以及这会怎样影响市值。
英文原文
@jpm7019 $LPK not exactly. $ALMU fits the definition more of speculation just going off development. $LPK is already dominates a chokepoint and is qualifying with almost every semi. It’s just guessing what volume ramp looks like for all the semi companies and how that affects MC.
-
Lightmatter 直接买 SIVE 激光器
你提供的照片只是例子。但我一直猜 Lightmatter 很可能是客户，因为他们以前就提到过它。 但我指的是那句关于“客户”的报价：它们融资了 2.5 亿美元（Lightmatter 1.54 亿 + Celestial 1 亿），这更像是 Lightmatter 直接购买 $SIVE 激光器的证据。
英文原文
The photo you provided were just examples. But I always guessed Lightmatter was a likely customer because they referenced them before. But I was about the quote around "customers" that raised $250m (Lightmatter $154m + Celestial $100m), so that's more of evidence of direct relationships of Lightmatter buying $SIVE lasers.
-
Lightmatter 似乎在用 SIVE 激光器
从这些参考资料看……Lightmatter 似乎确实在用 $SIVE 的激光器？ 这是全新的信息发现，对 Sivers 来说极其正面。 Lightmatter 是一家体量巨大的私营龙头（2024 年估值大约 44 亿美元），而 $TSM、$GFS、$TSEM、$AMKR 和 $ASX 都在扩展它的光学项目。 而 Lightmatter 的确需要光源…… $SIVE 还恰好和 $LITE 一起出现在 $GFS 的激光源供应商名单里，所以这些线索开始拼起来了？ 终端用户很可能就是你的超大规模云厂商，比如 $MSFT、$GOOGL（他们投资了）、$META 等，通过 Lightmatter-GUC 等路径进入。 我不认为市场已经把这些都定价进去了，因为整个供应链 BOM 都非常保密，而且很多只是推测。 但当 CPO 和下一代光子架构开始放量时，量产爬坡带来的营收会突然出现在资产负债表上。
英文原文
From the reference... it does look like Lightmatter uses $SIVE lasers? Which is brand new information discovery and extremely positive for Sivers. Lightmatter is a massive private leader (~$4.4B valuation back in 2024), with $TSM, $GFS, $TSEM, $AMKR, and $ASX scaling their optical program. And Lightmatter does require a light source... $SIVE also happens to be on the $GFS laser source suppliers alongside $LITE, so starting to put the dots together? End users are likely your hyperscalers like $MSFT, $GOOGL (they invested), $META, etc through Lightmatter-GUC and others. I don't think markets have priced all of this in, since all the supply chain BOM is very confidential + speculative. But when CPO and next-gen photonic architectures scale up, volume ramp revenue will appear out of nowhere on the balance sheet.
-
1.4 亿市值对激光供应链太小
@TVAFR786 14 亿美元市值算小吗？ 尤其如果他们的激光器很可能服务于 $AMD CPO、$MRVL CPO、$JBL 1.6T LRO、$AAPL 硅光、Ayar（AlChip）、Lightmatter、O-Net 终端客户（亚洲超大规模云厂商）以及其他项目。
英文原文
@TVAFR786 $1.4B MC is tiny? Especially if their lasers likely power $AMD CPO, $MRVL CPO, $JBL 1.6T LRO, $AAPL Silicon Photonics, Ayar (AlChip), Lightmatter, O-Net End users (Asian Hyperscalers), and others.
-
Celestial 可能是 SIVE 直接客户
很棒的发现。基本上从旧投资人材料和融资金额来看： 看起来 Celestial 曾经就是 $SIVE 的直接客户。 所以：$MRVL 现在大概率是在直接从 $SIVE 买激光器，而不是通过 Poet 购买（Poet 再买 Sivers 激光器来封装）。 因为 Celestial 最初就是围绕 Sivers 的激光器设计的。 Marvell 可能早就准备好通过垂直整合，绕开封装 IP 和组装环节，以获得更高利润率。 现在直接供货给 $MRVL，对 $SIVE 来说非常利好。
英文原文
Great find. Basically from old investor deck + fundraising amounts: Looks like Celestial has been direct customers of $SIVE. So: $MRVL is likely buying lasers direct from $SIVE now instead of through Poet (which bought Sivers lasers to package them). Since Celestial is designed around Sivers lasers originally. Marvell probably prepared to vertically integrate away packaging IP + assembly for higher margins anyway. Going direct to $MRVL is very bullish for $SIVE.
-
NVDA/TSM 生态还有上行
@latent_value7 嗯……$NVDA / $TSM 之类现在都在用 MSSCorps，我觉得这还有很大上行空间。
英文原文
@latent_value7 Yeah... $NVDA / $TSM and others all using MSSCorps has pretty big upside imo.
-
CPO 检测市场的天花板很高
CPO 是下一代光子架构的圣杯。 这位 CEO 的说法是要“拿下 90% 的 CPO 检测市场”（其实就是接近 100%，但他们不能这么说，不然会涉及反垄断）。 如果能维持这个地位，从 14 亿美元市值涨到 50 亿至 100 亿美元，并不离谱。
英文原文
CPO is the holy grail of next-gen photonics architectures. The CEO's claim was to "seize a 90 percent share of the CPO inspection market" (basically 100% but they can't say that otherwise antitrust). $5-10B+ MC isn't far fetched from $1.4B if they manage to maintain their status.
-
MSSCorps 在 CPO 检测上像垄断
我现在做多 MSSCorps（6830），市值大约 14 亿美元。 这看起来像是 CPO 检测环节的功能性垄断。 但市场可能把它和材料/失效分析混淆了，后者更像 MA-tek 和 iST 那样的寡头格局。 按客户映射来看： 1. $TSM 2. $NVDA 3. $AAPL 4. $AMAT 5. $LRCX 6. $ASML 7. $INTC 还有很大概率包括 $AVGO、联发科、三星、$MRVL 等等（他们也提到了欧盟）。 如果你感兴趣： - 台北时报把 TSM 说成是客户，Msscorps 为其提供先进材料和失效分析，还点名了苹果、英伟达、Lam、AMAT（也感谢 Latent 和我一起做尽调。英伟达那边还有通过 LinkedIn 之类看到的其他供应链关系） - 对 $AMSL 而言，《台北时报》在 2024 年 9 月 10 日写到：“ASML 采用了 Msscorps 超高灵敏度的光刻胶材料分析” - 对英特尔来说，材料分析实验室 MSSCORPS 已经拿下了英伟达和英特尔等大厂的订单（工业技术研究院） 至于检测（非破坏性红外漏检），他们几乎是垄断。 而且他们还积极使用诉讼（比如 Enli Tech 诉讼）来挡住竞争对手，我把这看成正面因素。 这带来了巨大的定价权和收益率，而且每个主要玩家都要经过他们。 CPO 检测市场也极其关键，和 $AXTI 在 InP 基板环节一样，这种巨大卡点具有纯粹的定价权和提价能力。 风险在于专利诉讼不一定按计划推进，但英伟达和其他超大规模云厂商大概率不会在 MSSCorps 赢了之后去找别家，因为他们不想承担风险。 所以这本身就构成了一个多年期的巨大优势。 超大规模云厂商不会等一个新兴竞争对手证明自己能赢再行动，也不会愿意承担这个风险。 我确实看到了 MSSCorps 因为在 CPO 的关键良率卡点上有机会重估，所以我选择做多（非投资建议，自己做功课），这就是我的思路。
英文原文
I am now long MSSCorps (6830) ~$1.4B MC This appears to be a functional monopoly in CPO for inspection. But markets might have conflated that with Material/Failure Analysis with MA-tek and iST (oligopoly). For customers from mapping: 1. $TSM 2. $NVDA 3. $AAPL 4. $AMAT 5. $LRCX 6. $ASML 7. $INTC And high probability $AVGO, MediaTek, Samsung, $MRVL, and others (they did mention EU too). If you're curious: - Taipei times names TSM as a client that Msscorps provides them with advanced material and failure analysis and name drops Apple, Nvidia, Lam, AMAT (also S/O to Latent for doing DD with me on this. there's other supply chain relationships to Nvidia through things like linkedin) - For $AMSL, Taipei Times, Sept 10, 2024 "ASML adopted Msscorps' ultra-sensitive materials analysis of photoresists" For Intel - Material analysis lab MSSCORPS has secured orders from major manufacturers such as Nvidia and Intel (Industrial Technology Research). For inspection (non-destructive infrared (IR) leakage detection), they're a monopoly. And have aggressively used litigation (like the Enli Tech lawsuit) to lock out rivals, which I view as a positive thing. This creates massive pricing power with yields and every major player goes through them. CPO inspection market is also extremely critical and like $AXTI in the InP substrate section, this massive chokepoint has pure pricing power with price hikes. The risk is the patent suit doesn't go as plan, but Nvidia and other hyperscalers aren't likely to go with other parties in case MSSCorps wins, so this creates a massive multi-year advantage anyway. Hyperscalers aren't going to wait to see how an emerging competitor is going to win or not + take the risk. I do see the massive re-rating potential with MSSCorp holding a critical yields chokepoint over CPO, so I went long (NFA, DYOR), this is just my thought process.