北京时间 2026-07-16 22:54:42 · 个股论点

SIVE轻晶圆厂产能模型显示高毛利潜力，CW瓶颈加剧利好，看好其CPO地位及并购上行空间。

中文翻译

很多人对 $SIVE 通过轻晶圆厂模式（Win Semi 等）进行产能爬坡建模感到好奇：假设使用 Win 10% 的晶圆产能作为低端配置（65% 良率假设，$50-$75 ASP）： Sivers 将支撑 $3.41 亿-$5.12 亿的年度阵列营收。鉴于管理层 50-60%+ 毛利率目标的上限，大致为： $2.05 亿-$3.07 亿的年度毛利润。相对于 Sivers 当前约 $11 亿的市场资本化，若此产能情景在 2028 年实现，约为： 3.6–5.4 倍市值/毛利润。若按 15% 计算，毛利润为 $3.07 亿-$4.61 亿（2.4–3.6 倍市值/毛利润）。 Sivers CEO 也回复称他们正在与更多晶圆厂合作以获取产能。从旧版演示文稿看，自 2024 年以来似乎有更多的资格认证。因此，随着共封装光学(CPO) 的兴起，产能目标可能比此处陈述的更大。我也预计未来季度营收管道预测将上调，因为更多合格供应商进入大规模量产(HVM)。 _ 至于需求端，连续波(CW) 激光器也非常瓶颈。 Lumentum 因 ER 财报电话会议中的外延面激光二极管(EML) 义务，正在公开市场购买 CW 激光器。 $AMD 正在签署长期协议(LTA) 以锁定 CW 产能（来自 Trendforce）。所以当 Sivers 与 $GFS, $JBL, Ayar, $POET, O-NET 等合作爬坡时... 鉴于当前的限制，任何上线的独立合格产能很可能被吸收。锦上添花的是，摩根士丹利将 $SIVE (~$11 亿) 列为三大领先 CPO 激光器玩家之一。与其近期笔记中 $550 亿+ 的 Coherent 和 Lumentum 并列，是有原因的... _ 总结：Sivers 仅需 Win 的低端配置，相对于当前估值即可产生可观的毛利润。我认为最大的未建模营收上行空间在于美国纳斯达克上市后的 TAM 扩张与并购。通过复制 Lumentum 与 Cloud Light 的剧本，构建整个收发器模块或光引擎。

英文原文

Lot of people were curious about $SIVE capacity volume ramp modeling through fab-light (Win Semi + others): Using 10% of Win's wafer capacity as a low-end allocation (65% yield assumption, $50-$75 ASP): Sivers would support $341-$512M worth of annual array revenue. Given upper end of managements 50-60%+ gross margin target, would be roughly: $205–307M of annual gross profit. Against Sivers current ~$1.1B MC, would be ~: 3.6–5.4× MC/gross profit if this capacity scenario plays out in 2028. And at 15% would be $307–461M in gross profit (2.4–3.6× MC/gross profit) Sivers CEO also replied that they're working with more fabs for capacity. And from an older deck, there looks to be more qualifications since 2024. So capacity targets might be larger than what's stated here as CPO takes off. I also expect to see revenue pipeline projections hiked in future quarters, as more qualification suppliers to into HVM. _ As for demand side, CW also happens to be very bottlenecked. Lumentum are buying CW off the open market due to EML obligations from their ER transcript. And $AMD are signing LTAs to secure CW capacity (from Trendforce). So when Sivers is ramping with $GFS, $JBL, Ayar, $POET, O-NET and others... Given the current constraints, it's highly likely any independent qualified capacity that comes online would be absorbed. And as a cherry on top, Morgan Stanley named $SIVE (~$1.1B) as one of the three leading CPO laser players . Alongside $55B+ players Coherent and Lumentum in their recent note for a reason... _ TLDR: Sivers only needs a low end allocation from Win to make substantial gross income relative to current valuations. I think the largest revenue upside that isn't modeled in if they TAM expansion with M&A after US NASDAQ listing. By copying the Lumentum playbook with Cloud Light to build out entire transceiver modules or with optical engines.

在 X 上查看原推 ↗