2026-03-11 · 个股论点

看多 Tower Semi 作为光子代工层的非对称机会

中文翻译

我做多 $TSEM，把它看成光子领域里的 $TSM。我最看好的两只 CPO 标的是 $SOI 和 Tower Semi。考虑到下周 $NVDA GTC 会有关于新光子架构的催化，我预计 Tower Semi 会迎来很强的催化。 Nvidia 上个月已经直接和 Tower 合作扩展 1.6T 硅光，这几乎就是在给 GTC 提示，随后很可能推动下游玩家采用它。而现在，Tower 已经是 1.6T SiPh PIC 的领先供应商，也是规模化 CPO 架构的主要代工厂（另一个是 GlobalFoundries）。根据我的前瞻估算： 2028 年前瞻 PE 约为 16.8 倍到 18.1 倍。（Tower 目标到 2028 年实现 28.4 亿美元收入，经营利润率约 31.7%，净利润约 7.5 亿美元。）要注意的是，他们计划中的 SiPh 产能已有 70% 以上被预订到 2028 年了，而光子业务甚至还没真正起量。所以我预计，随着光子产能极度扩张以及分配价格上调，而这些还没被模型完全计入，Tower 的盈利会明显超预期。另外，$TSEM 因为 70% 产能已被预订，实际上风险已经大幅下降。它的市值之所以还低，很大程度上是因为它是光子供应链里非常冷门的上游玩家。但我预计 $NVDA GTC 会成为把它推向溢价估值的催化剂。我做多 $TSEM，是把它当成上游光子代工层的非对称上行标的。

英文原文

I'm long $TSEM, the $TSM of photonics. My top two picks for CPO are $SOI and Tower Semi. Given the $NVDA GTC catalyst on new photonic related architecture next week: I expect Tower Semi to get a huge catalyst. Nvidia laready directly collaborated with Tower to scale 1.6T silicon photonics last month (hint hint for GTC), likely pushing the downstream players to use it. And now, Tower is the leading supplier of 1.6T SiPh PICs and the primary foundry for scale-up CPO architectures. (the other being global foundries) From my forward est: 2028 Forward P/E: ~16.8x to ~18.1x (Tower set a target $2.84B revenue by 2028, with ~31.7% operating margin, ~$750M in net profit) The thing to note is over 70% of their planned SiPh capacity is already reserved through 2028. And photonics haven't even ramped up yet. So, I expect them to strongly beat earning projections due to extreme photonics scaling + allocations price hikes that's not modeled into projections. Also, $TSEM is heavily de-risked by 70% of capacity already being reserved. MC is likely due to $TSEM being a very obscure upstream player in the photonics supply chain. But I expect the $NVDA GTC conference to be that catalyst that brings it to premium valuations. I'm long $TSEM as an asymmetrical upside for upstream photonics foundry layer.

在 X 上查看原推 ↗