2026-03-17 · 个股论点

$SIVE是硅光子/CPO架构转变中的InP CW DFB激光器核心供应商，可能复制$LITE的估值扩张路径。

涉及标的： $AAOI $AVGO $AYAR $COHR $FN $INTC $LITE $LWLG $MRVL $MTSI $NVDA $POET $SIVE $TSEM

中文翻译

$SIVE今天已上涨+73.78%（市值2.31亿美元）。市场正在对下一个潜在的 photonics 瓶颈进行信息综合定价。如果我必须解释区别： Lumentum的一个激光源主要受益于当前的光学瓶颈。而$SIVE的激光源则面向即将到来的CPO/硅光子学(SiPh)瓶颈。 Lumentum目前主要从$NVDA和超大规模云服务商那里获益，这些公司正在确保EML激光器的产能，用于当前的可插拔(optical transceiver)光收发器周期。从当前的EML瓶颈可以看出，超大规模云服务商正在抢购任何800G/1.6T光收发器及上游产能，从： -$AAOI（内部供应） -$COHR、$LITE（EML激光器+设计）-> $FN（组装） -$COHR、$LITE（EML激光器）-> Innolight / Eoptolink 下一步是什么？硅光子学(Silicon Photonics)和共封装光学(Co-Packaged Optics,CPO)。向CPO的架构转变需要大规模的高功率连续波(CW) DFB激光器阵列。而这可能引发一次完全的、突然的需求量范式转变。 $SIVE受益于用于SiPh和CPO的磷化铟(InP) CW DFB激光器：即将崛起的公司如： $AYAR、$POET使用$SIVE的激光器，但主要做先进封装(advanced packaging)。然后它们向上输送给更大的公司如$MRVL Celestial（后者购买$POET的interposers）。然而，如果你追溯到上游，光源是$SIVE。 CW DFB激光器是光引擎($SIVE)；硅光子学封装($POET等)是传输方式。 CPO规模尚未达到。但我们知道它即将到来。正如从当前光收发器周期所见： -$LITE和$COHR的光源需求获得的估值远高于$FN等专注先进封装的公司。市场一直在关注$POET，但错过了它们获取Starlight实际$ITE型光源的地方。风险是存在的，包括与$LITE、$COHR、$AVGO等公司面临多源竞争。所以再次强调，请务必做自己的研究。但我对此的反驳： Sivers足够早地为$POET、Ayar和其他公司定制激光器以满足规格要求——早在这两家公司变得流行之前（类似于$POET与$MRVL Celestial的类比）。也存在产量风险：但Win Semi认证的潜在成功可以抵消这一点。为扩大产能而稀释股权的风险，对于每个早期阶段的公司来说始终存在。我去年做过$LITE的研究，仍看好这只股票用于Google TPU ramp/光电路开关(OCS)。但今年，我专注于： $SIVE，作为新一轮光子学架构转变的个人CW DFB激光器敞口。我分享自己对捕捉从当前EML周期向即将到来的CW DFB/硅光子学周期轮动的思考。

英文原文

$SIVE is now up +73.78% today ($231M MC). As markets price in information synthesis of the next potential $LITE of photonics. If I had to explain the difference: One laser source in Lumentum primarily benefits from current optical bottlenecks. The other in $SIVE is for the upcoming CPO/Silicon Photonic bottleneck. Lumentum is largely benefiting right now from $NVDA and hyperscalers securing capacity of EML lasers for current pluggable optical transceivers cycles. As seen with the current EML bottleneck, hyperscalers are buying out any 800G/1.6T transceiver + upstream capacity from: - $AAOI (in-house) - $COHR, $LITE (EML lasers + design) -> $FN (assembly) - $COHR, $LITE (EML lasers) -> Innolight / Eoptolink What's next? Silicon Photonics and Co-Packaged Optics. The architectural shift to CPO requires massive arrays of high-power CW DFB lasers. And this would likely trigger a complete, sudden paradigm shift in volume demand. $SIVE benefits from InP CW DFB lasers for SiPh and CPO: The up and coming companies like: $AYAR, $POET source $SIVE lasers, but primarily do advanced packaging. Then they feed up to larger companies like $MRVL Celestial (that buy $POET's interposers). However, if you go upstream, the light source is $SIVE. CW DFB lasers are light engine ( $SIVE ); the silicon photonics package ( $POET and others) is how it gets transmitted. CPO scale is not there yet. But we know it's coming. And as seen with current optical transceiver cycles: - Light sources from $LITE and $COHR demand much higher valuations than companies like $FN that focus on advanced packaging. Markets have been focusing on $POET, but missed where they get the actual $LITE type light source for Starlight. The risks are present including facing multi-source competition with $LITE, $COHR, $AVGO, and others. So again, make sure to do your own research. But my argument against that: Sivers been early enough to tailor custom lasers to fit $POET, Ayar, and other specifications before they got popular (sort like the $POET to $MRVL Celestial analogy). There's volume risks as well: But the potential Win Semi qualification offsets that. Dilution risk to scale capacity, is always present with every early-stage company as well. I did my thesis on $LITE last year and still love the stock for Google TPU ramp/OCS. But this year, I'm focusing on: $SIVE, as my personal CW DFB laser exposure for the new photonics architectural shift. I’m sharing my own thoughts on capturing the rotation from the current EML cycle to the upcoming CW DFB/Silicon Photonics cycle.

在 X 上查看原推 ↗