供应链分析

产业链结构、上下游与瓶颈环节研究 · 共 1122 条 · 第 11 / 23 页

  1. 解释 SIVE 大涨的真正催化来自会议确认和光源瓶颈,而不是一条 X 帖子。

    所以有深度投资者(像你们),也有交易员(像我)。 投资者往往忽略的是,究竟是什么让像 $SIVE 这样的股票上涨。 对我来说,我把时间点卡在 $NVDA GTC 和 OFC 大会附近,因为 Nvidia 推进 CPO 架构大概率会有新消息。 这波抛物线式上涨的原因是: -> $SIVE 被确认是 O-Net 的光源。 -> $SIVE 又在 OFC 上被确认是 Jabil 的光源。 这些才是我那篇 thesis 发出后最大的催化。 它本来就已经很不对称了,因为还有 $POET -> $MRVL Celestial,以及 Ayar 这一层。 而 Jabil 的可插拔公告正好又和我做多的时间点重合,只是锦上添花。 这不是“发一条 X 帖子,股票就涨 300%”那么简单,事情没那么运作。

    英文原文

    So there's deep investors (like you), then there's traders (like myself). What investors miss is what causes a stock like $SIVE to go up. For me, I timed it around $NVDA GTC + OFC conference, since there's likely new news around Nvidia pushing CPO architectures. The reason for the parabolic rise is: -> $SIVE confirmed as light source to O-Net. -> $SIVE confirmed as light source to Jabil at OFC. These were the biggest catalysts after my thesis post was announced. It was already asymmetrical due to $POET -> $MRVL Celestial as well as Ayar. But the Jabil plugabble announcement timed with me going long is a cherry on top. It's not just an "X post, stock goes up 300%", doesn't work like that.

  2. 认为 SOI 估值重估较慢,但硅光和 CPO 放量会带来上行。

    @excu4e16421 $SOI 的重估速度比 $AAOI 慢,但我觉得它已经接近旧业务拖累周期的底部,而且它们提供的衬底会从硅光 / CPO 放量中受益。 一年后也许能有 2.5 倍,这其实已经很好了,不是每只票都能三个月翻三倍。

    英文原文

    @excu4e16421 $SOI is slow rerating compared to $AAOI but I do think it’s the bottom of their legacy drag cycle but they get upside from silicon photonics/cpo ramp up with the substrates they provide. Maybe 2.5x in a years time (which is still good, not everything triples in 3 months)

  3. 认为 SIVE 是光学激光瓶颈的最高暴露标的,已经是多家供应商的 CW 光源。

    @NBA_GSW_30 我真的很喜欢 $SIVE,它是光学激光瓶颈里最有暴露度的标的! 它们已经是 Jabil、Ayar、Marvell Celestial、O-Net 等很多供应商的 CW 光源。 而且我觉得这家公司还有很长的路要走。

    英文原文

    @NBA_GSW_30 I really like $SIVE as the highest exposure to the optical laser bottleneck! They’re already the CW light source for many suppliers like Jabil, Ayar, Marvell Celestial, O-Net. And I think the company has a long way to go

  4. 认为光子学并非纯投机,超大规模客户订单已排到 2028。

    我不会说这只是投机。超大规模云厂商从 $LITE 到 $COHR 的订单都已经排到 2028 年了。 就连 $TSEM 也有 70% 的产能被预订,而且它 2028 年的前瞻市盈率也才大约 20 倍。 CW / CPO 对光子学来说更像是投机部分,而不是可插拔光模块周期本身。可相比存储 / GPU,它其实也算合理。 不过我觉得大家现在盯的光子瓶颈右边,像 PCB 之类,挺有意思的。

    英文原文

    I wouldn't say it's speculative. Hyperscalers have orders completely booked until 2028 from $LITE to $COHR. Even $TSEM has 70% of their capacity booked and it's low fwd ~20 P/E 2028. CW/CPO more speculative for photonics, not the pluggable optical transceivers cycle. But compared to memory/GPUs, makes snse. But think everyones looking at photonics bottlenecks right low like pcb, which is funny

  5. 持有 VPG 作为 TSLA Optimus 传感器供给的一部分,但不确定是否会被中国伙伴替代。

    @throughthetape 我持有 $VPG,因为这曾经是我做 Tesla Optimus 的多头标的。 不过,我也不能 100% 确定它们是不是已经被中国合作伙伴在量产放大时设计绕开了…… 但这段话看起来又像是它们仍然是关键供应商?真的很难完全确定。

    英文原文

    @throughthetape I have a position in $VPG since this was my $TSLA Optimus long. But, I'm not 100% sure if they got designed out or not by Chinese partners for mass production ramp... However, this statement makes it sound like they're still a critical supplier? Really hard to know for sure.

  6. 把 VPG 视作 Tesla Optimus 传感器供应链的一环,并引用机器人量产目标。

    我很久没再聊 $VPG 了。 这曾经是我看 Tesla Optimus 放量的一个传感器标的。 本周 Sidoti 大会上有个非常重要的线索: “到 2026 年底前,每周有可能扩产到几百台机器人。” 这听起来是不是很像 Optimus? 如果按他们 CEO 那句“每台机器人 500 到 1200 美元”的说法来看:也不差。 但当然,很多机器人供应链在真正的人形机器人大规模拐点出现之前,基本都还是死钱。 也许 Unitree IPO 会让机器人供应链稍微活跃起来。

    英文原文

    Been awhile since I covered $VPG. This was a sensor play for $TSLA Optimus Ramp. Very important breadcrumb this week at the Sidoti Conference: "Potential to scale to hundreds of robots per week by the end of 2026." Kinda sounds aligned with Optimus? No? If we go off their CEO "$500-$1.2k per robot" statement: Not too shabby. But of course, a lot of robotic supply chains are kind of dead money until the actual inflection point humanoids at scale hits. Maybe Unitree IPO sparks some life into robotic supply chains soon?

  7. 认为 P4O 作为玻璃供应商在 LITE/COHR 供应链里有策略价值。

    @genZinvest0r 我差不多每隔一条帖子就会收到 3 条关于 $P4O 的评论,所以最后还是去看了看 DD。 从战略角度看,它作为 Bosch、Samsung 以及可能下游 $LITE、也许 $COHR 的玻璃供应商,确实挺有价值。 从财务上看就没那么强了,因为它在 BOM 里占比极低。

    英文原文

    @genZinvest0r I got like 3 $P4O comments every other post of mine, so finally ended up taking a look at the DD. Strategically it's pretty valuable as a glass supplier to Bosch, Samsung, and likely downstream $LITE maybe $COHR etc. Financially, not as much since extremely low BOM.

  8. 认为上游玻璃晶圆供应链很有趣,但自己并不推荐买入。

    @AtlasShrug1 这个供应链地图本身就很有意思。 当一位关注者发现了可能位于 $LITE OCS 上游的玻璃晶圆供应商时,它旁边还有像 $GLW 这种千亿美元公司的存在。 我不是在推荐,也从来没说过这是个好投资。

    英文原文

    @AtlasShrug1 The supply chain map is just very interesting. When a follower found likely $LITE OCS upstream suppliers on the glass wafer level alongside hundred billion dollar companies like $GLW. Not recommending it and never said it was a good investment.

  9. 详细映射 P4O 可能成为 LITE 上游玻璃供应商,但强调未必能显著转化为收入。

    我刚看了评论…… $P4O 看起来像是一个潜在的 $LITE 上游供应商,市值只有 3400 万美元。 它大致的映射可能是: -> Plan Optik $P4O(玻璃晶圆) -> Teledyne $TDY 或 Silek(存储晶圆厂) -> $LITE(OCS 交换机) -> $GOOGL(TPU) 不过,$LITE 这个层面上的玻璃晶圆供应,可能也会和 Corning($GLW)双源采购。 P4O 是 Infineon、Samsung 等公司的已知供应商,所以也不是随便的一家公司。 €335 万现金 vs. 3552 万美元市值,资产负债表还挺健康。 但这不代表它一定会带来显著收入提升,除非他们开始提价(因为玻璃晶圆大概只占 $LITE OCS BOM 的很小一部分)。 我个人在读完这些之后是持有了一些股份。 因为 $LITE OCS 潜在供应链瓶颈在战略上很有价值,也许财务上没那么明显。 但它关于 Plan Optik 的玻璃流和 MDF finishing 在 OCS 封装里几乎不可替代这一点,我觉得是成立的。 当然,超大规模云供应链保护得很严,所以也不可能 100% 确定。 这些功劳都属于跟帖的粉丝们。 但 TLDR: 供应链映射 -> $P4O -> 晶圆厂 -> $LITE -> $GOOGL,按他们的分析是很可能的。 至于它是不是 $LITE OCS 供应链里的瓶颈……我不确定它到底能不能显著转化成财务价值。 再说一遍,我完全不是在推荐它。 只是觉得这种 $LITE OCS 供应链映射很有意思,把粉丝辛辛苦苦做的研究丢到空白里也太浪费了。

    英文原文

    So just checked out the comments... $P4O does look like a potential upstream $LITE supplier at a $34M MC. How it likely maps: -> Plan Optik $P4O (Glass Wafers) -> Teledyne $TDY or Silek (Mem Foundries) -> $LITE (OCS Switches) -> $GOOGL (TPUs) However, $LITE likely dual sources with Corning ( $GLW ) on the glass wafer level. P4O is a known supplier to Infineon, Samsung, and others, so not exactly a random company. €3.35 million vs. 35.52M MC is very healthy balance sheet. That being said: -> That doesn't exactly mean this translates to material revenue boosts unless they start price hike (given glass wafers are likely a very small part of $LITE OCS BOM). I do personally own shares, after reading this. Since $LITE OCS potential supply chain chokepoints is strategically valuable. Maybe not so much so as financially valuable. But their core thesis that Plan Optik's Glass Flow and MDF finishing are effectively irreplaceable for OCS packaging checks out. Of course, hyperscaler supply chains are heavily guarded, so no way to know 100%. All credit goes to follower comments. But TLDR: Supply Chain Mapping -> $P4O -> Foundries -> $LITE -> $GOOGL is very likely from their analysis. As for being a chokepoint in $LITE OCS supply chain... dunno if it translates materially. Again, not recommending this at all. Just thought the $LITE OCS supply chain mapping like this very interesting, and that it would be a waste of the follower kinda posted all this work into the void.

  10. 认为标的当前约 4.4 亿市值,长期可看到数十亿美元级别,Jabil 放量后 MTSI 也会重估。

    @rodrimartinezgg 现在大概交易在 4.4 亿美元市值附近,但我对今天的预期是 20 到 30 亿美元。 然后在一年半后,一旦 Jabil 和其他公司开始放量,它可能会到 $MTSI 的 170 到 200 亿美元市值。

    英文原文

    @rodrimartinezgg Trading at ~$440M right now but my expectation is $2-3B today. Then could be $MTSI $17-20B MC in a year and half once Jabil and others scale up.

  11. 认为机构开始意识到 SIVE 在 CW/EML 瓶颈中的重要性,自己仍看它是下一个十倍。

    当机构开始意识到 $SIVE 在 CW / EML 激光瓶颈里的重要性时,真是有种“TFW”的感觉。 但他们手里却没有仓位…… Sivers 到现在已经涨了 220%+,但我预计它会成为我下一个像 $AXTI 那样的个人 10 倍股。 https://t.co/DmqpkZ8Lqa

    英文原文

    TFW institutions realize how important $SIVE is in the CW/EML laser bottleneck. But have no positions… Sivers is up 220%+ so far, but I expect this to be my next personal 10x like $AXTI. https://t.co/DmqpkZ8Lqa

  12. 惊讶 TSEM 在自己做多后两周涨了 63%。

    兄弟,像 $TSEM 这种 200 多亿美元市值的公司,怎么会在我做多两周后就涨了 63%…… 这家公司在我发帖前还沉寂了大概 3 个月。 https://t.co/fTYqm8ew03

    英文原文

    Bro how is a $20B+ company like $TSEM up 63%… 2 weeks after I go long? The company was stagnant for like 3 months until my post. https://t.co/fTYqm8ew03

  13. 认为 Sandisk 投资 Nanya 可能说明传统存储瓶颈比市场想的更严重。

    @Kaizen_Investor 也许对 $SNDK 来说这是战略性动作,我不觉得他们会随便到处扔 10 亿美元。 我会继续挖一挖,看看市场是不是忽略了更大的影响。

    英文原文

    @Kaizen_Investor Probably strategic for $SNDK, don't think they'd just throw $1B around here and there. Will do some digging and see if there's any larger implications markets are missing.

  14. 说自己年初看多 Nanya,没想到 SNDK 现在直接入股,暗示旧存储瓶颈可能被低估。

    哇,Nanya 年初还是我做多的标的…… 我完全没想到 $SNDK 现在居然刚刚入股了它 10 亿美元以上。 Sandisk 选择入股像 Nanya 这样的公司,说明 legacy memory 的瓶颈可能比市场想的更大、也更被低估。 这对 Macronix 以及其他台湾公司也有巨大的影响。 或者……$SNDK 只是 NAND 价格上涨赚太多了,正好找地方花钱?

    英文原文

    Woah Nanya was my long earlier in the year... Did not expect $SNDK to take a $1B+ stake in it just now. Fact that Sandisk is taking stakes in companies like Nanya, means bottleneck in legacy memory might be bigger / more underpriced than marketes thought. Massive implications for others like Macronix + other TW companies. Or... $SNDK just has too much money from NAND price hiking and needed somewhere to spend it?

  15. 欢迎别人加入 VNP/FPLSF,称其对超大规模云供应链极其重要。

    @PhotonCap 欢迎加入 < $VNP / $FPLSF >。 它们确实是中国和日本以外,对超大规模云供应链最重要的西方公司之一。

    英文原文

    @PhotonCap welcome aboard to &lt; $VNP / $FPLSF &gt;. Genuinely one of the most important Western companies to hyperscaler supply chains outside of China/Japan.

  16. 说 SIVE 在瑞典交易所,想买的人会受限。

    @bennybigbull $SIVE 需要瑞典那边的交易权限。

    英文原文

    @bennybigbull $SIVE needs trading permissions in Sweden

  17. 说感谢别人的决定,自己只是分享 thesis 和 conviction。

    @MildBullyMMA 这是你自己做的决定,不用谢我! 我只是一路分享我自己的 thesis 和我自己的 conviction。

    英文原文

    @MildBullyMMA You made the decision yourself, don't thank me! I'm just sharing my own thesis and my own conviction around $SIVE along the way.

  18. 说这些公司会跟着 Win Semi 放大产能。

    @arslang3 它们会跟着 Win Semi 一起扩产。

    英文原文

    @arslang3 They scale with Win Semi.

  19. 认为光模块 PCB 只是次级机会,最赚钱的是 EML/CW 瓶颈。

    @Antrunt 可以关注光模块 PCB,但我会把重点放在 EML / CW 瓶颈上,因为那才是最赚钱的部分。

    英文原文

    @Antrunt optical transceiver pcb, but I'd focus on the EML/CW bottleneck which is the most profitable.

  20. 分析Sivers被收购的可能性及CHIPS Act对半导体收购的监管影响

    @TGodTracer 哦,100%的公司都会想收购Sivers。他们是世界上少数几家像$MTSI这样能做到这一点的公司之一。 但收购可能会被阻止,因为$SIVE对欧洲和美国都至关重要,涉及国家安全(CHIPS Act受援企业)。

    英文原文

    @TGodTrader Oh 100% companies would want to acquire Sivers. They're one of the few companies in the world like $MTSI that can do this. But acquisitions might be blocked since $SIVE is central to national security for both Europe and US (CHIPS Act recipient).

  21. 认为 SIVE 估值仍极小,重估空间很大。

    @Ayouubb_ $SIVE 现在的市值大概还只有 4.2 亿美元,在这个瓶颈里的位置下,这简直小得离谱。 我觉得它的重估空间真的很大。

    英文原文

    @Ayouubb_ $SIVE is still only ~$420m MC which is hilariously small given their positioning in this bottleneck. There's just so much room to be rerated imo.

  22. 说 SIVE 因在瑞典交易所,很多人买不到;早买的人会很开心。

    @PhotonCap 对啊,很多人买不到 $SIVE,因为它在瑞典交易所…… 但那些早期买到的人,接下来几个月大概率会很开心。

    英文原文

    @PhotonCap Yeah, lot of people can't buy $SIVE because it's in a Swedish exchange... But people who did early are likely going to have a fun time over the next few months.

  23. 引用 Broadcom 讲话,强调激光是半导体瓶颈,应优先布局 CW/EML 与晶圆厂。

    每个行业领导者…… 尤其是今天这段话里 $AVGO(Physical Layer Products 部门)。 他们提到激光是半导体的瓶颈。 如果你还没多头布局…… -> CW 激光:$SIVE | $MTSI -> EML 激光:$COHR | $LITE -> 或者它们的晶圆厂:$TSEM / Win Semi 也许是时候醒醒了? Broadcom 的 Ramachandran 说: “虽然现在行业里有多个供应商……但激光领域确实存在明确的供给约束。”

    英文原文

    Every industry leader... Especially $AVGO (Physical Layer Products division) in this statement today. Cites Lasers as a bottleneck for semiconductors. If you aren't long... -> CW Lasers: $SIVE | $MTSI -> EML Lasers: $COHR | $LITE -> or their foundries in $TSEM/Win Semi Maybe it's time to wake up? Broadcom Ramachandran: "Even though there are multiple suppliers in the industry today... there is definitely a supply constraint in the laser space,”

  24. AAOI在光收发器超级周期中拥有巨大的重新估值潜力,其美国本土完整供应链是核心优势。

    如果你是$AAOI的新手: 在 ATH 附近以 $109 进场看起来很吓人,毕竟今年已涨了 175%。 然而,这看起来就像是光子学领域的 $SNDK。 而碰巧的是,光收发器超级周期的中心... 在 $8.2B 估值下,我们可以看看预测: (预计产能 * 平均售价预测) 2026年Q2:~$312.1M 2026年Q4:~$1.41B 2027年Q2:~$1.53B 2027年Q4:~$1.97B 如果$AAOI最终在营收上超越了 $55B 的 $LITE... $8B 市值对于这种收入增长来说看起来有点荒谬,利润率可达 30-40%... 重新估值潜力是巨大的。 最坏的情况如果内部激光晶圆厂失败,就得从 $COHR 采购。 或者最终成为一个美国制造版的 Innolight / Eoptolink(两家都是 $66B-$90B+ 的中国公司)? 在美国建立完整供应链的选择权被低估了。 $AAOI 是我少数高度看多的标的之一,除了 $SIVE 和 $LITE,如果他们能兑现预测的话。

    英文原文

    If you're new to $AAOI: It looks scary entering positions near ATH at $109 after a 175% YTD increase. However, this looks like the photonics equivalent of $SNDK. And it so happens the center of the optical transceiver supercycle... At a $8.2B valuation we can look at projections: (est. Capacity * ASP Projections) Q2 2026: ~$312.1M Q4 2026: ~$1.41B Q2-2027: ~$1.53B Q4-2027: ~$1.97B If $AAOI ends up leapfrogging the $55B $LITE in revenue... $8B MC looks a little absurd for revenue growth off 30-40% margins… Re-rating potential is enormous. Worst case scenario if they fail internal laser fab and buys off $COHR. Is it ends up a Made in America Innolight / Eoptolink (Both $66B-$90B+ Chinese companies)? The optionality of making the entire supply chain in America is understated. $AAOI is one of my few high conviction longs aside from $SIVE and $LITE if they can deliver on projections.

  25. 披露自己持有 VNP,并称 NEO 对机器人供应链很重要。

    @VacuityNomad 先声明一下:我持有 $VNP。 我知道 $NEO,而且我在之前的 DD 里就把这家公司标为机器人供应链里很重要的一环。很不错的选择!

    英文原文

    @VacuityNomad Just a disclosure above: I do own $VNP. I'm aware of $NEO and I flagged that company as important to robotic supply chains in my earlier DD. Great pick!

  26. 回顾 AXTI 10 倍成功后,列出 SIVE、AAOI、VNP、SOI 等下一批高信念供应链标的。

    所以我的 10 倍 Moonshot $AXTI 最后真的飞到了月球。 如果你觉得自己错过了火箭: -> $SIVE 是我现在最喜欢的下一个 AXT 式 10 倍 Moonshot。 -> $AAOI 自从我提到之后已经涨了 4 倍,但我还是觉得它从现在这个位置还能再涨 4 倍。 如果你想玩得更稳一点: -> $VNP 是西方版本的 $AXTI -> $SOI 是硅光 / CPO 版本的 $AXTI,但它还没真正起量。 -> $IQE 更偏向重组后的二元结果。 -> $TSEM、$LITE、$COHR 则是更稳的复利型标的。 我之前还没怎么提过 $VNP(5N Plus),但它确实是西方供应链里最重要的公司之一,是 $AXTI 的对照面。 它主要供应:铟、锗、镓、碲、铋。也供应 LEO 卫星太阳能电池,以及光子学的衬底 feedstock。 市场里永远还有更多机会!

    英文原文

    So my 10x moonshot $AXTI ended up landing on the moon. If you felt like you missed the rocket: -> $SIVE is my personal favorite next 10x AXT moonshot at today's prices. -> $AAOI already up 4x since I've mentioned it, but I still think it can 4x again from these levels. If you want to play a bit safer: -> $VNP is the Western $AXTI -> $SOI is the $AXTI for silicon photonics/CPO, but it hasn't really ramped up yet. -> $IQE is more binary based on restructuring. -> $TSEM, $LITE, $COHR are the steadier compounders. I haven’t mentioned $VNP (5N Plus) much yet but it’s genuinely one of the most important companies in Western supply chains as the counterbalance to $AXTI. Basically it supplies: indium, germanium, gallium, tellurium, bismuth. Solar cells for LEO. Substrate feedstock for photonics. There's always more opportunities in the market!

  27. 强调 SIVE 给 Jabil 的 1.6T 供光已被 OFC 确认,并非猜测。

    @siv_s7 这不是猜测?$SIVE 为 Jabil 1.6T 供光这件事已经在 OFC 上被确认了。 这还不包括 Ayar、$POET Celestial,以及其他 CPO 玩家。

    英文原文

    @siv_s7 It’s not speculation? $SIVE powering Jabil 1.6T is confirmed from OFC. This is not even including Ayar, $POET Celestial, and other CPO players.

  28. 认为瑞典媒体在把当地散户吓出 SIVE,结果把筹码转移给更懂供应链的美国投资者。

    我真的觉得很有意思…… 你会看到像《Privata Affärer》这样的瑞典本地媒体,几乎没有先进半导体背景: 却在把这家公司大部分持股的瑞典本地散户吓出 $SIVE。 它大概是他们国家里市值最高的上行空间科技股,市值大约 3.2 亿美元。 然后这些仓位就转到了更懂超大规模云 CPO / photonics 供应链的美国投资者手里。 本地媒体居然在替美国投资者干活? 谢谢?

    英文原文

    I honestly find it amusing... You have Swedish local journalists like "Privata Affarer" with no background in advanced semis: Scaring the company's majority-owned local Swedish retail investors out of their $SIVE positions. Probably the highest upside tech-stock in their country at ~$320m MC. And transferring positions to US investors who understand hyperscaler CPO/photonics supply chains better. The local media is doing all the work for American investors? Thanks?

  29. 认为太空里做数据中心和追踪导弹有现实应用,但未必像光子学那样转化为现金流。

    @yuntungshieh 这种应用还要几年,但确实是真实存在而且很重要。比如从太空追踪高超音速导弹,或者在太空做数据中心。 问题在于它能不能转化成 FCF,我的看法是……可能不如光子学。

    英文原文

    @yuntungshieh Few years away but applications are definitely real and important. Like tracking hypersonic missiles from space or DCs in space. Problem is if it translates into FCF, which I’d argue… probably not compared to photonics.

  30. 将卫星周期和光子超级周期对比,认为后者更能支撑收入和盈利增长。

    这里有一个太空卫星超级周期: -> 由 $PL 触发。 随后 $BKSY / $SATL / $SPIR 等跟上。 然后还有光子超级周期: -> 由 $LITE 和 $COHR 触发。 随后 $AAOI / $TSEM / $MTSI / $SIVE / $IQE / $SOI 等跟上。 前者是对太空应用的炒作。 后者是对 AI 带来的极端收入和盈利增长的炒作。 后者通常更容易被证明是合理的。

    英文原文

    There's the Space Satellite MegaCycle: -> Triggered by $PL. With $BKSY / $SATL / $SPIR and others following. Then there's the Photonics SuperCycle: -> Triggered by $LITE and $COHR. With $AAOI / $TSEM / $MTSI / $SIVE / $IQE / $SOI and others following. One is hype over applications from Space. The other is hype over extreme revenue and earnings growth from AI. The latter tends to be more defensible.

  31. 认为 SOI 只是慢慢重估,但衬底业务仍有 2-3 倍空间。

    @miraclemaster07 $SOI 没什么变化,我只是继续拿着,它现在也只涨了适度的 20%。 我觉得它在接下来一年里还能涨 2 到 3 倍。

    英文原文

    @miraclemaster07 Nothing's changed about $SOI, just sitting on it and it's up a modest 20%. I think it can 2-3x in the next year.

  32. 认为 SOI 可能供应 HIMX->TSM 链条,并在混合树脂上几乎垄断。

    @zephyr_z9 这其实不太是资产负债表的问题。 他们更可能是 $HIMX -> $TSM COUPE 链条的供应商,而且在 WLO 所需的混合树脂上几乎形成了半垄断。

    英文原文

    @zephyr_z9 Not really about the balance sheet per say. They're the likely supplier to $HIMX -&gt; $TSM COUPE, and had a semi-monopoly over hybrid resins needed for WLO

  33. 强调 SIVE 通过 CHIPS Act 和国安项目对美国很重要,并给出更高估值判断。

    如果你到现在还不知道,$SIVE 通过 CHIPS 法案对美国国家安全很重要。 而且尤其是,它还是超大规模云 AI 供应链的光源。 到目前为止,他们已经拿了大约 1160 万美元的两笔 CHIPS 法案资助: 1. 和 $ERIC 以及 Raytheon $RTX 一起做 FR3 beamformer ICs 2. 和 $BA(LSE)BAE Systems 一起做 EW 技术 CEO 还暗示 2026 年会有更多来自 CHIPS 的资金。 小盘股几乎从来拿不到美国政府这种支持。 你手里有一家市值 3.3 亿美元的公司: 它对美国半导体业务的国家安全至关重要, 也对超大规模云供应链($MSFT、$META、$AMZN、$ORCL 等)里的光子学至关重要, 通过 $MRVL Celestial、Ayar、Jabil、ONet 等链条。 关于扩产和 capex 的问题,现在已经由 Win Semi 合作基本回答了。 至于“等待 CPO”的机会成本,现在也被 OFC 上 Jabil 1.6T 可插拔模块放量这件事回答了。 $SIVE 是光子学里最被低估的宝石之一,我个人认为它今天就应该值 20 亿美元以上。

    英文原文

    If you didn't know by now $SIVE is important to US national security via CHIPS act. But especially, as the light source for hyperscaler AI supply chains. They got two CHIPS acts grants ~$11.6M so far: 1. FR3 beamformer ICs with $ERIC and Raytheon $RTX. 2. EW Tech with $BA (LSE) BAE Systems With more coming funding from CHIPS in 2026 hinted by the CEO. Small caps almost never get this sort of support from the US government. You have a company sitting at $330M: That's critical to US National Security for their Semi arm. and Critical to Hyperscaler supply chains ( $MSFT, $META, $AMZN, $ORCL, etc) for Photonics through $MRVL Celestial, Ayar, Jabil, ONet, and others... The main question regarding scale vs. capex is now answered by Win Semi partnership. And the question regarding "CPO waiting" opportunity cost is now answered by the Jabil 1.6T pluggable ramp from OFC. $SIVE of the most unknown gems in the photonics space and I personally think this company should be $2B+ today.

  34. 列出多项 CHIPS Act 资助名单,并强调国家安全和政府资金的重要信号。

    美国 CHIPS Act 资金是最大的信号之一。 对美国国家安全的重要性。 不太出名的名单: 1. $SIVE(3.3 亿美元) 2. $AKTS(9.2 亿美元) 3. $BLG(3400 万美元) 4. $QUBT(16.5 亿美元) 5. $RGTI(52.7 亿美元) 6. $MRAM(2.1 亿美元) 7. $SKYT(已被收购) 8. $PDFS(13.5 亿美元) 9. $ATOM(1.77 亿美元) 10. $SLVC(2.27 亿美元) 11. $NVTS(21.2 亿美元) 12. $WOLF(7.5 亿美元) 13. $ALMU(2.34 亿美元) 14. $ACLS(26 亿美元) 直接资助 | DoC CHIPS PMTs: 1. $WOLF(Wolfspeed):7.5 亿美元 2. $INFN(Infinera):9300 万美元(DoC PMT)- 后来被 Nokia 收购 3. $SKYT(SkyWater Technology):1600 万美元(DoC PMT)- 后来被 IonQ 收购 直接资助 | 联邦研发合同: 4. $RGTI(Rigetti Computing):1128 万美元(AFRL / 量子网络与 AFOSR 芯片 fab) 5. $QUBT - DoC / NIST 针对 TFLN PICs 的直接合同 国防微电子 / 分包商: 6. $SIVE:1160 万美元(NEMC Hub) 7. $MRAM - 870 万美元项目合作方(通过 SCMC Hub 的“CHEETA”项目) 8. $ACLS - 780 万美元合作项目(由 GE Aerospace 通过 CLAWS 领导) 9. $BLG - 650 万美元(迄今通过 CLAWS Hub 获得的国防分包) 10. $AKTS - 400 万美元(NEMC 项目“LADDER”的主导资助) 关联公司: 11. $ATOM:SWAP Hub 成员 12. $NVTS:SCMC Hub 的嵌入式合作方 13. $PDFS:CA DREAMS Hub 的行业合作方 14. $ALMU:CA DREAMS 和 MMEC Hubs 的附属成员 -> 电磁战(EW):在 6 个项目中合计分配了 5100 万美元。 -> AI 硬件:在 7 个项目中合计分配了 4200 万美元。 -> 5G/6G 无线通信:在 5 个项目中合计分配了 4200 万美元。 -> 商业跃迁技术:在 7 个项目中合计分配了 3800 万美元。 -> 量子技术:在 4 个项目中合计分配了 3200 万美元。 -> 安全边缘计算 / IoT:在 4 个项目中合计分配了 2500 万美元。 提醒一下: 对美国技术重要,并不等于就是好投资(比如 Wolfspeed)。 我只是研究了一下 $SIVE 的 CHIPS Act 资助,然后顺手看看还有哪些较不知名的公司也拿了补助或成了受益者。

    英文原文

    US Chip ACT Funding is one of the largest signals. For importance to America National Security. Lesser known list: 1. $SIVE ($330m) 2. $AKTS ($920m) 3. $BLG ($34m) 4. $QUBT ($1.65B) 5. $RGTI ($5.27B) 6. $MRAM ($210m) 7. $SKYT (Acquired) 8. $PDFS ($1.35B) 9. $ATOM ($177M) 10. $SLVC ($227M) 11. $NVTS ($2.12B) 12. $WOLF ($750m) 13. $ALMU ($234m) 14. $ACLS ($2.6B) DIRECT FUNDING | DoC CHIPS PMTs: 1. $WOLF (Wolfspeed): $750 Million 2. $INFN (Infinera): $93 Million (DoC PMT) - Bought by Nokia 3. $SKYT (SkyWater Technology): $16 Million (DoC PMT) - Bought by IonQ DIRECT FUNDING | Federal R&D Contracts: 4. $RGTI (Rigetti Computing): $11.28M (AFRL/quantum networking and AFOSR Chip fab) 5. $QUBT - Direct DoC/NIST contract for TFLN PICs DoD Microelectronics / Sub-Contractors 6. $SIVE: $11.6 Million (NEMC Hub) 7. $MRAM - $8.7 Million project partner ("CHEETA" project via SCMC Hub) 8. $ACLS - $7.8M partner project (Led by GE Aerospace via CLAWS) 9. $BLG - $6.5M (DoD sub-contracts secured to date via CLAWS Hub) 10. $AKTS - $4 Million (Lead grant for the "LADDER" project via NEMC) Adjacents: 11. $ATOM: Member of the SWAP Hub 12. $NVTS: Embedded partner in the SCMC Hub 13. $PDFS: industry partner in the CA DREAMS Hub 14: $ALMU: affiliate member in the CA DREAMS and MMEC Hubs -> Electromagnetic Warfare (EW): $51 million allocated across 6 projects. -> Artificial Intelligence (AI) Hardware: $42 million allocated across 7 projects. -> 5G/6G Wireless Communications: $42 million allocated across 5 projects. -> Commercial Leap-Ahead Technologies: $38 million allocated across 7 projects. -> Quantum Technology: $32 million allocated across 4 projects. -> Secure Edge Computing / IoT: $25 million allocated across 4 projects. Just a warning: Being important to US technology (eg. Wolfspeed) does not translate to being a good investment. Was just doing research into $SIVE CHIPS act funding and decided to see what other lesser known companies got grants or were beneficaries too.

  35. Jabil采用$SIVE光源的1.6T可插拔模块填补了等待CPO的空窗期。

    @lxjourney_ Jabil的1.6T光收发器(pluggables)使用$SIVE作为光源。 之前你必须等待CPO(共封装光学)方案,但现在有了一家最大的光子学厂商之一——继承了$INTC(英特尔)的部门,拥有超大规模云厂商(hyperscaler)终端用户,填补了等待的空窗期。

    英文原文

    @lxjourney_ Jabil 1.6T transceivers for pluggables using $SIVE as the light source. Before you had to wait for CPO but now you have one of the largest photonics players that inherited $INTC's division with hyperscaler end-users, bridging the waiting gap.

  36. 称 SIVE 是小盘光子学里最有吸引力的名字,远超其他可重估公司。

    $SIVE 是小盘光子学领域里最有说服力的名字,我个人会继续把更多仓位集中到它上面。 我知道还有别的名字看起来也被低估、可能会重估得更高,但你不会从它们那里拿到像 CW 激光瓶颈那样的 10 倍空间。

    英文原文

    $SIVE is the most compelling name in the small-cap photonics sector, and I'd personally just put more concentration into that. I'm aware of the other names that look undervalued and could get re-rated higher, but you won't get 10x upside like you would with CW laser chokepoints.

  37. 说这些公司会跟着 Win Semi 一起扩产。

    @thejoemando 它们会跟着 Win Semi 一起扩产。

    英文原文

    @thejoemando They scale with Win semi

  38. 认为 SIVE 作为 CW 激光瓶颈,按同行估值应明显高于当前市值。

    我老是被问: $SIVE 的目标价是多少?它就是一家 3.2 亿美元市值的公司,能跑多远? 它是关键的 CW 激光公司,也是光子学里下一个瓶颈。 最接近的对标是 $MTSI 的 170 亿美元市值。 然后是 $LITE 的 519.7 亿美元市值。 我觉得它今天就应该交易在 20 到 30 亿美元。 但它现在只有 3.2 亿美元。 这种巨大的估值差距,说明它一旦放量,潜在上行空间非常大。

    英文原文

    I keep getting asked: What's the PT of $SIVE? It's a $320m MC company, how far can it run? It’s a critical CW laser company, the next chokepoint in photonics. Closest comparison is $MTSI is $17 Billion. Then $LITE at $51.97 Billion. I think it should trade at $2-3B, today. Yet it’s $320 million. The sheer valuation gap points to massive upside when it scales.

  39. 用一句话概括 CW 激光从 SIVE 到 MTSI 的链条。

    @KiroIkigai CW 激光这条链就是 $SIVE 到 $MTSI。

    英文原文

    @KiroIkigai Cw lasers with $SIVE to $MTSI

  40. 认为在 Jabil 公告后 SIVE 应值 20-30 亿美元,如今仍只有约 3.35 亿。

    @ThatIsIt11 我个人认为,在 Jabil 公告之后,$SIVE 今天就应该值 20 到 30 亿美元。 而现在它大概只在 3.35 亿美元附近。

    英文原文

    @ThatIsIt11 I personally think $SIVE should be valued at $2-3B today after the Jabil announcement. Right now it’s sitting at ~$335m

  41. 高信念看好 SIVE,认为体量更大后机构会成为核心持有人。

    我对 $SIVE 的未来有很高的信念! 我确实认为,一旦市值再大一点(加上成交量),很多机构也会有能力成为核心持有人,从而支持它的成长。 我也收到了很多反馈,说基金目前还因为一些 mandate 的限制没法买进。

    英文原文

    I have high conviction in $SIVE future! I do think once the market cap gets a tad larger (with volume), many institutions would be able to be core holders as well to support the growth. Getting a lot of feedback from that funds aren’t able to buy in just due to certain mandates so far.

  42. 说 SIVE 是自己的下一步。

    @Igosuki $SIVE 是我的下一步……

    英文原文

    @Igosuki $SIVE is my next play…

  43. 说 Soitec 似乎更担心整体市场,还需要时间建立信心。

    @dimitrisnatsis Soitec 似乎更关心整体市场。 也许它还需要一点时间来建立自己的信心。

    英文原文

    @dimitrisnatsis Soitec seems more concerned with the overall market. Maybe takes some more time for it build up its confidence.

  44. 说 SIVE 上涨 6.7%,在大盘普跌时说明这是选股者市场。

    $SIVE 今天涨了 6.7%。 当从 KOSPI 到 Nikkei 的整个市场都在跌的时候,你就知道自己做对了什么。 而 $TSEM 到 $AXTI 这些个股还在涨。 很明显,现在已经是选股者市场了。 像石油、稀土、半导体这些全世界都在争夺的主题标的,已经明显跑赢了更广泛的市场。

    英文原文

    $SIVE is up 6.7% today. You know you're doing something right: When the entire market from KOSPI to Nikkei is red. While individual longs from $TSEM to $AXTI are green. It's clear that this is now a stock picker's market. As thematic individual names in Oil, Rare Earths, to Semis, that the world is fighting over... Have strongly outperformed the broader market.

  45. 同意 SIVE 是光子超级周期里被低估的光源,并已进入 Jabil 1.6T 光模块周期。

    是的,完全同意!$SIVE 是一个非常被低估的名字,它是即将到来的光学超级周期里的光源。 它甚至都还没等到 Celestial 或 Ayar 的 CPO,就已经在 Jabil 的 1.6T 光模块周期里被点名为光源,这正好填补了可插拔等待的空窗。

    英文原文

    Yes agreed! It $SIVE is such an undiscovered name that's the light source for the up and coming optical supercycle. It's not even waiting for CPO with Celestial or Ayar either. They're the named light source for Jabil's 1.6T optical transcivers cycle, which bridges the waiting gap with pluggables.

  46. 说那家公司就是生产铟、锗、镓、碲、铋的那家。

    @inversorBetico 就是那家生产铟、锗、镓、碲、铋的公司。

    英文原文

    @inversorBetico That's the company that makes indium, germanium, gallium, tellurium, bismuth

  47. 梳理机器人供应链所需的稀土、结构金属等关键材料公司,涵盖30余只小市值标的

    通常这个会发在「淋浴灵感」板块,但今天放主时间线。 $ALOY - 稀土粉末 -> 高纯度金属。用于液态冷却泵的磁铁 $NB - 钪,用于服务器机架的轻量化金属框架 $UURAF - 稀土分离 $ARA - 磁铁上游原料 $MEI - 中国境外的IAC矿床 $NTU - 硬岩重稀土 这些都是比较低调的,市值都不到10亿美元。 然后这是我之前整理的机器人供应链: $UUUU - 从独居石砂加工高纯度钕铁硼 $MP - 在Mountain Pass提取氟碳铈矿,垂直整合到国内NdFeB磁铁制造 $ALOY - 将重稀土氧化物转化为国防级合金和高温金属如钐和钆 $USAR - 加工重稀土元素并制造烧结NdFeB磁铁 $LYSDY - (Lynas Rare Earths Limited) 中国境外唯一商业化生产分离重稀土元素的公司 $NEO - (TSX): 他们是西方公司中目前唯一能大规模商业化生产实际NdFeB磁粉和合金的 $ILU - (稀土精炼): 澳大利亚的稀土精炼厂 $ARU - (ASX): 「矿石到氧化物」钕镨设施 2. 结构冶金(铌、钒、钛、铍) $ATI - 美国高性能钛和特种合金的主导生产商,用于机器人关节 $CRS - 美国特种结构合金供应商,包括高强度钢、钛和磁性材料 $FCX - 全球最大的钼生产商,行星滚柱螺丝的结构钢绝对必需 $NB - 关键的纯正公司,开发内布拉斯加州的Elk Creek项目,目标是供应国内的铌、钒和钛 $MTRN - 全球主要的铍加工商 $LGO - 领先的钒公开上市加工商 $BMM - 锗和镓的陆上供应和加工 $VNP - 用于先进传感器和电子的镓、锗和铟 $TECK - 中国境外最重要的锗生产商 $ALB - 锂提取 $EAF - 用于电池阳极的高纯度石墨 $ALTM - 供应电池所需的西方锂 $SYR - 巴拉马矿的石墨 $FCX - 人形机器人需要高达6.5公斤铜 $AW1 (ASX): 推进犹他州的West Desert项目,为高品位镓和铟提供国内地质来源 我知道还有很多... 只是有些在地缘政治上有点危险,怕我暴露给地缘政治对手看(说的就是你$AXTI) 如果你要做多西方阵营的一家,我最喜欢的前五。 图片来源:Crossdock Insights, Visual Capitalist 顺便说一下,这不是潘多拉魔盒的内容。 只是在我完成某个特定做多想法的研究之前,随意提一些有趣的名字。 半导体封装所需的关键元素,以及其他接近完全进口依赖的元素 -> 所以可能会有资金流入 只是随意的想法

    英文原文

    Normally this goes in shower thoughts but will post main timeline today. $ALOY - Rare earth powders -> high-purity metals. Magnets used in the liquid cooling pumps $NB - Scandium, lightweight metal frames for server racks. $UURAF- Separation of rare earths. $ARA - Upstream feedstock for magnets $MEI - IAC deposit outside of China. $NTU - hard-rock heavy rare earths These are all more under the radar stuff like sub <$1B. Then for robotics supply chains I made this earlier: $UUUU - Processes monazite sand into high-purity Neodymium $MP - Extracting bastnäsite at Mountain Pass and vertically integrating into domestic NdFeB magnet manufacturing. $ALOY - Converting heavy rare earth oxides into defense-grade alloys and high-temperature metals like Samarium and Gadolinium $USAR- Process heavy rare earth elements and manufacture sintered NdFeB magnets $LYSDY -(Lynas Rare Earths Limited) - Only commercial producer of separated heavy rare earth elements outside of China. $NEO -(TSX): They are the only Western company commercially producing the actual NdFeB magnetic powders and alloys at scale right now. $ILU- (rare earths refinery): rare earths refinery for Australia $ARU- (ASX): "Ore-to-oxide" NdPr facility 2. Structural Metallurgy (Niobium, Vanadium, Titanium, Beryllium) $ATI - Dominant US producer of high-performance titanium and specialty alloys required for robotic joints. $CRS - US supplier of specialty structural alloys, including the high-strength steels, titanium, and magnetic $FCX - World's largest producer of Molybdenum, which is strictly necessary for the structural steel in planetary roller screws. $NB - Critical pure-play company developing the Elk Creek project in Nebraska, aimed at supplying domestic Niobium, Scandium, and Titanium $MTRN - Major global processor of Beryllium $LGO - Leading publicly traded processors of Vanadium $BMM - Onshore supply and processing for like Germanium and Gallium $VNP - Gallium, Germanium, and Indium for advanced sensors and electronics $TECK - Most significant producer of Germanium outside of China $ALB - Lithium extraction $EAF - High-purity Graphite for battery anodes $ALTM - Western lithium required to supply the batteries $SYR - Balama mine for Graphite $FCX - Humanoid requires up to 6.5 kilograms of copper $AW1 (ASX): Advancing the West Desert project in Utah, for domestic geological sources for high-grade Gallium and Indium I'm aware of a lot of more... It's just some are a little dangerous to mention geopolitically in case I broadcast vulnerabilities to geopolitical adversaries reading this (looking at you $AXTI) If you're going long on a western one 5n my favorite. Image source: Crossdock Insights, Visual Capitalist FYI this is not the pandoras box stuff. Just name dropping some interesting names informally before finishing up some research on a specific long idea. Critical elements required for semiconductor packaging and others close to near-total import reliance -> so prob influx of funding going toward it. Just informal thoughts

  48. 说 SIVE->AEVA 的映射已经知名,但自己更看好其与 Boston Dynamics Atlas 供应链的关系。

    @JuhongH $SIVE -> $AEVA 这条映射已经很有名了。 人们常把 $AEVA 更多和卡车、国防这类工业用途联系起来。 但我更喜欢它在 humanoids 里的位置。而且我觉得目前根本没人把 $SIVE 映射到潜在的 Boston Dynamics Atlas 供应链上。

    英文原文

    @JuhongH $SIVE -&gt; $AEVA is already well known. People often cite $AEVA for more industrial stuff like trucking and maybe defense. But I like it for humanoids in specific. And I don't think anyone mapped $SIVE to possible Boston Dynamics Atlas supply chains at all yet.

  49. 说自己对 AEVA 的兴趣也让他关注到 SIVE 这个上游供应商。

    @AscendingLife3 我只是觉得这挺有意思,因为我本来就已经在投像 $AEVA 这样的前沿科技公司了。 而 $SIVE 这个名字恰好也是它们的潜在上游供应商。

    英文原文

    @AscendingLife3 I just find it interesting when I’m already invested in a frontier tech company like $AEVA. The other one in $SIVE happens to be their likely upstream supplier as well.

  50. 把 SIVE->AEVA->LG->Boston Dynamics 视为 4D Physical AI 到光子学的交叉链条。

    如果你不记得的话: $AEVA 是我做 4D Physical AI + World Models 的多头。 -> LG 向 $AEVA 投资了 5000 万美元,共同开发 FMCW 4D LiDAR,并明确提到了 humanoids。 -> LG(Boston Dynamics 的视觉供应商)。 但…… 谁在生产这些 FMCW LiDAR 用的 CW 激光器? 就是那个同样负责光子学放量、并在 $JBL 和 $MRVL Celestial 的 1.6T+ 可插拔模块里扮演关键角色的公司。 它大概率也是 humanoids 和 4D physical AI 的 4D AI CW 激光供应商。 一个很可能的映射是: -> $SIVE -> $AEVA -> $LG -> Boston Dynamics。 $SIVE 和 $AEVA 都是我的两只多头,但 4D Physical AI 到光子学这些前沿赛道往往会重叠。

    英文原文

    If you don't remember: $AEVA was my long for 4D Physical AI + World Models. -> LG $50m in $AEVA to co-develop FMCW 4D LiDAR, explicitly citing Humanoids -> LG (Boston Dynamics vision spuplier). But... Guess who makes those CW lasers for FM-CW Lidar? The very same company for scaling photonics with 1.6T+ pluggables with $JBL to CPO in $MRVL Celestial. Is the likely 4D AI CW laser supplier for humanoids and 4D physical AI. One highly possible mapping: -> $SIVE -> $AEVA -> $LG -> Boston Dynamics. Both $SIVE and $AEVA were my two longs, but frontier sectors in 4D Physical AI to photonics tend to overlap.