2026-03-27 · 供应链分析

分析 $SIVE 在光子学 CPO CW WDM 供应链中的关键地位，估值 $3 亿相对 $LITE 的 $550 亿市值极具不对称优势。

涉及标的： $AVGO $COHR $JBL $LITE $MRVL $MTSI $POET $SIVE $WIN

中文翻译

"没有 Win（赢家）就拼不出 Winner（胜利者）。" "因此：我看多 $SIVE 供应链。" "$SIVE -> WIN (TPE:3105) -> Ayar SuperNova -> $JBL -> 超大规模数据中心，作为一个整体流程。" "如果你对比同样处于共封装光学(CPO) CW WDM 领域的 $LITE 等公司。" "原因如下：" "-> $POET / $MRVL Celestial。" "-> Ayar / $JBL 和 O-Net 使用 $SIVE。" "它被设计为超大规模数据中心下一代光子学架构的光源。" "估值约 $3 亿。" "而且我确实认为 WDM DFB 阵列是规模扩展(scale up)的优选架构。且开发难度极高。" "这不是零和博弈的架构之争，如何处理规模扩展(scale out)很可能会与使用单发射器的 $LITE 和 $COHR 分工合作。" "以及像 $MRVL 这样有内部供应链的厂商如何设计他们的架构。" "但如果你比较市值差异（$SIVE 约 $3 亿，$LITE 约 $550 亿）。" "任何人都能看清这个光子学领域 $3.3 亿美元即将到来的颠覆者 $SIVE 价值有多明显。" "尤其是 Win Semi 斥资 $40 亿的晶圆厂承接了来自 $SIVE、$AVGO、$MTSI 等客户的 fabless 激光器规模化生产。" "对于做多这条供应链以及光子学和超大规模数据中心架构发展方向来说，这是高度非对称性的机会。

英文原文

You cant spell Winner without Win. Because of that: I’m bullish on $SIVE supply chain. $SIVE -> WIN (TPE:3105) -> Ayar SuperNova -> $JBL -> Hyperscalers, as one flow. If you compare to $LITE and others that’s also in the same CPO CW WDM space. There’s a reason why: -> $POET / $MRVL Celestial. -> Ayar / $JBL and O-Net use $SIVE. It’s designed in as the light source for the next-gen photonics architecture for hyperscalers. At a ~$300M valuation. And I do think WDM DFB arrays are the superior architecture for scale up. And are incredibly hard to develop. It's not a zero-sum architectural game and will likely be split with how you handle scale out with single emitters like $LITE and $COHR. As well as captive suppliers like $MRVL and how they design their architectures. But if you look at the MC difference (~$300m with $SIVE, $55B with $LITE). Anyone can see how clear how valuable the $330M incoming disruptor in $SIVE is to the photonics space. And especially with the $4B foundry in Win Semi that captures fabless laser production from $SIVE, $AVGO, $MTSI and other players at scale. It’s highly asymmetrical to long both the supply chain as exposure to where photonics and hyperscalers architecture are heading.

在 X 上查看原推 ↗