供应链分析
产业链结构、上下游与瓶颈环节研究 · 共 1309 条 · 滚到底自动加载更多
-
TOWA 或成 HBM4 上游受益者
是啊……我越看越觉得，市值约 13.4 亿美元的 $TOWA / $TOWCF（6315），可能会成为 SK 海力士、三星、$MU 在 HBM4 支出中的一个“捡漏”受益者。 它高度周期化，跟支出周期强相关。 但我们看到存储厂商的资本开支在全面上调。 这倒不是那种能 4 倍涨的标的…… 但我的猜测是，新一轮 HBM4 存储周期开始后，必须的资金流入激增会让它被显著重估。 尽管它在这轮新周期里似乎已经被大多数人忘了，但它在压缩建模流程上有垄断地位。 混合键合虽然只是去掉了一个步骤，但仍然需要 Towa（垄断企业往往会拿到更高溢价）。 我也不觉得这轮周期和上一轮一样，因为最近存储公司赚钱的速度实在太夸张了？ 只是我的想法，但如果我的判断没错……等大家都在建 HBM4 产线时，它应该会在 11 日周一（东京时间）的财报里体现出来？
英文原文
Yeah... the more I look into things, $TOWA / $TOWCF (6315) at ~$1.34b looks like a sleeper beneficiary for HBM4 spend from SK Hynix, Samsung, $MU. It’s hyper cyclical with spending cycles. But we’re seeing hikes in capex spend across the board from memory makers. Not exactly a 4x type play... But my guess is that it should get materially re-rated from a spike in unavoidable inflows for the start of the new HBM4 memory cycle anyway. And they appear to be a largely forgotten beneficiary for this new cycle despite their monopoly in compression modeling processes. Hybrid bonding just removes one step in the process but requires Towa anyway (monopolies tend to get higher premiums) I also don't think this cycle is like the last due to how much memory companies have been printing recently? Just my thoughts, but if my guess is right... it should show up in earnings on the Monday 11th (Tokyo Time) as everyone’s building out HBM4 lines?
-
MSFT 上游供应链脆弱点
业绩不错，也很高兴看到 $MSFT 在加速美国 AI 基础设施建设。 不过，微软漏掉了一个关键脆弱点，我想听听你的看法。 微软有没有什么措施，来防止地缘政治对手把你们的上游供应链收购掉？ 比如说，RoboTechnik 收购了微软依赖的 ficonTEC。 微软依赖的自动化组装和测试设备，现在已经落到地缘政治对手手里了。 这看起来非常危险，而超大规模云厂商可能还没意识到基础设施所有权方面正在发生什么。
英文原文
Hey great earnings and nice to see $MSFT accelerating US AI infrastructure. However, there is one critical vulnerability Microsoft is missing, was wondering if I could get your take. Is there anything Microsoft is doing to prevent geopolitical adversaries from buying out your upstream supply chains? For example, a company RoboTechnik acquired ficonTEC, which Microsoft relies on. The automated assembly and testing machines that Microsoft relies on are now owned by geopolitical adversary. This just seems very dangerous and hyperscalers might be missing what's happening with ownership of infrastructure.
-
Foxconn 让 Shunsin 进入 Nvidia 供应链
只是提醒一下：我不需要每天都发一条关于我持有标的的帖子。 不过，关于 Shunsin（6451）这件事我确实漏掉了： $NVDA 选择了富士康作为 Groq 加速器的独家机架级供应商。 这对 Shunsin（富士康的光学封装业务）非常正面。 所以 ShunSin 的 CPO 和光学封装，基本上借着这条新闻，间接搭上了 Nvidia 的下一代供应链。 市场还没有意识到这层关系（而且它在流通盘分布里，对大多数人来说是被冻结的）。 文章里明确写到，富士康正在整合这些系统，包括“……网络”。 既然富士康是独家组装商，它就会垂直整合，并很可能把 ShunSin 的光学封装方案纳入设计里，而不需要 ShunSin 直接从 Nvidia 那里拿到网络合同。 这就是我说 Shunsin 本质上是在白嫖 $NVDA CPO 生态的意思。富士康体量太大了。 而且你得到的是 Nvidia 光学供应链里的头部玩家，估值却比 $LWLG 还低，哈哈。
英文原文
Just an FYI: I don't need to post something every day about the names I own. However, I did miss this with Shunsin (6451): Foxconn picked by $NVDA as exclusive rack-scale supplier for Groq accelerator. This is highly positive for Shunsin (Foxconn's optical packing arm). So ShunSin's CPO and optical packaging basically backdoored Nvidia's next-gen supply chain with this news. Markets just haven't realized this connection yet (and it's on disposition, so frozen for most). The article explicitly notes that Foxconn is integrating these systems "including ... networking." Bc Foxconn is the exclusive assembler, they'll vertically integrate and likely use ShunSin's optical packaging solutions into the design, bypassing ShunSin's need for networking contract directly from Nvidia. This is what I meant by Shunsin is literally a free piggy back ride to $NVDA CPO ecosystem. Foxconn is massive. And you're getting the leading player in Nvidia's optical supply chain for a lower valuation than $LWLG lol.
-
测试/良率链条普遍利多
我帮你盯着这件事（测试/良率篇）： $VIAV 和 $FORM 的财报：极度看多 这意味着什么？ 像 $ONTO / $CAMT 这样的名字会起飞。再加上 $TOWA（6315），因为有迹象显示存储产量正在激进爬坡。 像 Msscorps / $KEYS 这样的名字也会起飞。 更广泛来看，存储和光学生态的上游良率、测试、验证和检测都会大幅受益。而且如果它们正在扩产，这对 $COHR、$FN、$LITE 等也都是领先指标。 对 $VIAV： -> 4.068 亿美元 vs. 3.93 亿美元（超预期）同比增长 42.8%。 -> 每股收益 0.27 美元 vs. 0.2-0.24 美元 指引为 4.27 亿至 4.37 亿美元，显示加速。 对 $FORM： -> 2.26 亿美元，按年增长 32%，每股收益 0.56 美元 vs. 0.45 美元 -> 毛利率大幅提升到 49%（这说明有定价权）。 -> 指引为每股收益 0.61 美元，中值营收约 2.4 亿美元。 “高带宽存储（HBM）需求创纪录，以及代工和逻辑网络应用更强劲” 基本上就是更小的良率/测试生态整体都在起飞。BRRR。
英文原文
Monitoring the situation for you (testing/yields edition): $VIAV and $FORM earnings: Extremely Bullish So what does this mean? Names like $ONTO / $CAMT go brr. Throw in $TOWA (6315), since there's indication of aggressive memory production ramp. Names like Msscorps / $KEYS should go brrr. Broader upstream yields, test, validation, and inspection for both memory + optical ecosystem go heavily BRRR. And it's a leading indicator for $COHR, $FN, $LITE, and others if they're ramping up production. For $VIAV: -> $406.8M vs. $393M (beat) 42.8% Y/Y growth. -> $.27 EPS vs $0.2-$0.24 Guidance was $427m-$437m, indicating acceleration. For $FORM: -> $226M, 32% Y/Y, $.56 EPS vs. $.45 -> margins increased a TON to 49% (which indicates pricing power). -> Guidance was $.61 EPS, midpoint ~$240m revenue. "Record demand for High Bandwidth Memory (HBM) and stronger "Foundry & Logic networking applications" Basically the smaller yields/test ecosystem in general. BRRR.
-
上游设备商TOWA受益于三大存储厂HBM4资本支出扩张，CPO封装尚未量产需观察ASX等信号，整体由超大规模数据中心资本支出驱动。
$TOWA最大的信号来自SK Hynix、Micron和Samsung，因为这是上游...三大厂商增加了资本支出预测+建设HBM4，所以资本支出确实会流向Towa。 MSSCORP更微妙。这是用于共封装光学(CPO)的，还没到量产的阶段...你可能需要关注$ASX和其他信号。 但它们都主要从超大规模数据中心资本支出作为总的框架来获得推导，今天这个框架基本是看涨的。
英文原文
$TOWA biggest signal is SK Hynix, Micron, and Samsung since this is upstream... and the big 3 increased capex projections + building out HMB4 so that capex does flow to Towa. MSSCORP more nuanced. This is for CPO, which is not volume ramped yet... You'd probably want to look at like $ASX and other signals. But they all get derived mainly from hyperscaler capex as an overarching umbrella, which was largely bullish today.
-
META 的 capex 利好供应链
@joestar1402 市场也许不喜欢 $META 激进地花 1450 亿美元资本开支……但我觉得小型半导体供应链公司会喜欢，哈哈。 这要看你关注什么。不过整体而言，财报普遍都还是正面的。
英文原文
@joestar1402 So markets might not like $META aggressive capex spending $145 billion… but i think the small semi supply chain players do lol. Depends on what you’re looking at. But earnings were generally positive across the board.
-
超大规模云厂商 capex 仍然很强
我已经帮你把情况盯好了。 超大规模云厂商 capex 支出的 TLDR（除了 $TSM 之外的信号）。 来自 $MSFT、$AMZN、$META 的财报： 上游半导体供应链会继续起飞。 $META：2026 财年 capex 为 1150 亿至 1350 亿美元 修正后的 Q1 财报：capex 提高到 1250 亿至 1450 亿美元。 -> 更高的零部件价格（涨价） -> 激进的 AI 基础设施支出 -> 定制模型训练 $MSFT：实际为 319 亿美元，而不是（352.9 亿美元），但原因是供应链瓶颈，而不是资源不足。 我们还在等 $AMZN 的确认，不过他们今年年初的原始 capex 预期大约是 2000 亿美元，来自 Jassy 的备忘录。 所以你可以放心，所有半导体供应链名字在下个季度大概率还是会继续起飞，因为有太多 capex 被灌进去。 然后还有新任美联储主席，他对 AI 和降息都非常积极，而 Jerome 在后面保持平稳（市场大概会喜欢这个）。
英文原文
I've monitored the situation for you. TLDR on hyperscaler capex spend (signal aside from $TSM). From $MSFT, $AMZN, $META earnings: Upstream semi supply chains go brrr. $META: 2026 FY capex $115-$135B Revised Q1 ER: $125B-$145B capex raised. -> Higher component prices (price hikes) -> Aggressive AI infra spending -> Custom Model Training $MSFT: was actually $31.9B vs ($35.29B) but due to supply chain bottlenecks, rather than lack of resources. Still waiting for confirmation around $AMZN but their original capex projection was ~$200B from Jassy's note earlier this year. So you can sleep easy, all the semi supply chain names still likely to keep going brr next quarter, since there's so much capex funneled into them. Then you get the new fed chair who is gung-ho all in on AI + rate cuts, with Jerome keeping stuff chill in the back (which markets probably like)
-
20.9M 对大公司不算太大
@ShadowTrades149 对，我看到了。2,090 万美元对一家 100 亿美元以上的公司来说其实不算太重要（但也有帮助），所以我没单独写。
英文原文
@ShadowTrades149 Yeah I did. $20.9M isn’t really very material to a $10B+ company (but helpful) so didn’t cover it.
-
TOWA 仍是必须持有
@johndoestocks 问得好，$TOWA 还是必须持有，所以不是问题。 即便到了本世纪末的 HBM5、采用 hybrid bonding，它们也许只是丢掉流程中的一个步骤，但仍然会是关键卡点。
英文原文
@johndoestocks Good question, $TOWA is still a must have so not a concern. Even with hybrid bonding end of decade hbm5, they might just lose one step of the process but remain a critical chokepoint
-
LPK 主打玻璃基板和激光刻蚀卡点
@MickaelLacoste2 不是。玻璃基板才是下一波先进封装浪潮。 而 $LPK 在激光诱导深蚀刻上几乎是功能性垄断。 这更像是对下一代封装范式的纯度敞口，而且没有 SKC 那种拖累，也没有 $INTC 那种其他业务分散。
英文原文
@MickaelLacoste2 No. Glass substrates are the next advanced packaging wave. And $LPK has a functional monopoly over laser induced deep etching. It’s more of pure play exposure to the next packaging paradigm without the drag SKC has or other segments $INTC has.
-
LPK 受益于玻璃核心基板与 CPO
我猜市场也认同我对 $LPK 的看法？ -> 当玻璃核心基板是下一波先进封装浪潮 -> 又是一个垄断卡点 -> 再叠加 CPO 应用。 通常就会一路起飞。 https://t.co/46u7msyDIl
英文原文
I guess markets agree with my take on $LPK? -> When you have glass core substrates as the next advanced packaging wave -> be a monopoly chokepoint -> add in CPO applications. It typically goes brrr. https://t.co/46u7msyDIl
-
SIVE 可能迎来 CHIPS Act 第二轮
提示一下……$SIVE 的 CHIPS Act 第二轮资金很快就要来了（这和 CHIPS Act 的续周期相一致） 美国政府不会随便把钱发给一家市值 10 亿美元的瑞典公司。 这非常不寻常。市场很可能漏掉了什么。 因为他们拿到了这些资金： -> $RTX 和 $ERIC 的波束成形器（空间/电信）。 -> BAE Systems 的 STAR 双工阵列（电子战中的雷达干扰）。 国防主承包商很可能正在用 Sivers 的微芯片 IP 来造最终产品。 至于最终产品……还有什么空间 + 军工承包商应用会涉及 LEO 卫星？我想你应该能猜到。 这些国防主承包商并不是要在太空里播放泰勒·斯威夫特的 YouTube 视频…… 但商业衍生品也可以用于 SpaceX / Amazon LEO Kuiper 这类应用。 不管怎样，你还拥有 AI 光子这条故事线，而这才是核心增长垂直领域。 我最兴奋的还是光子……但你也顺手得到了一家由美国 CHIPS Act 背书、嵌入 Raytheon 和 BAE Systems 的公司，去做那些他们在太空里想搞的黑科技。 只是没人知道精确时间线和应用，因为 CHIPS Act 国防合同就是这种性质。 但它也许会在未来的财报里突然出现。
英文原文
Hint... $SIVE CHIPS Act round 2 incoming soon (consistent with the CHIPS Act renewal cycle) The U.S. government doesn't just give out funding to random $1B Swedish companies. It's highly unusual. And there's probably something markets are missing. Since they got funding for: -> $RTX and $ERIC Beamformers (space/telecom). -> BAE Systems for STAR duplex arrays (radar jamming during electronic warfare) Defense Primes are likely using Sivers microchip IP to build the final products with this. As for the final products... What other space + military contractor applications are there involving LEO satellites? And I think you can guess. The defense primes aren't trying to play Taylor Swift Youtube videos in space... But the commercial spinoff can also be used for SpaceX/Amazon LEO Kuiper type applications. Regardless, you have the side of the AI photonics story, which is the core growth vertical. I'm most excited for photonics... but you also happen to get a company backed with U.S. CHIPS ACT embedded in Raytheon and BAE Systems for whatever black magic they want to do in Space. Just nobody knows exact timeline + applications because due to the nature of CHIPS ACT defense contracts. But it might appear randomly in financial statements down the road.
-
美国优先也要保住上游供应链
@Winterrose America first 也包括要从亚洲合作伙伴或者欧盟机器供应商那里，保护自己的超大规模云供应链。 如果美国对欧盟股票的控制权不够，中国就会试图接管上游垄断，这过去十年我们已经看到了。
英文原文
@Winterrose America first also includes securing its hyperscaler supply chains from partners in Asia or machine suppliers in EU. If US doesn’t own enough control of equities in EU, China will try and take over upstream monopolies, as we’ve seen this past decade.
-
把资金集中到长期卡点
@NabQ321 直接把全部仓位押到那种会随着时间大幅扩张 TAM 的卡点里，比如先进封装或光子领域的玻璃基板。
英文原文
@NabQ321 full port it into a chokepoint that has massive TAM expansion over time like glass substrates for advanced packaging or photonics.
-
指出Shunsin就是富士康，垂直整合是其核心优势。
@troyofsparta @tw_crypto_ Shunsin其实就是富士康。他们实际上是垂直整合(vertical integration)的受益者。
英文原文
@troyofsparta @tw_crypto_ Shunsin is literally Foxconn. They’re actually the benefit of vertical integration.
-
CPO TW 名字在回调后可分批布局
@tw_crypto_ 像 Shunsin 这类 CPO 台湾名字，很多都从持仓分布上回调过头了，所以我在那个赛道里做定投。 像 $AAOI 这类的 CSP 之类的东西，看起来也挺有希望。
英文原文
@tw_crypto_ A lot of the CPO TW names like Shunsin overcorrected from disposition so I’m cost averaging in that sector. CSP on stuff like $AAOI and others seems promising too.
-
Win Semi 是 HBM4 量产核心
所以 $SIVE 不负责产能扩张，真正负责量产爬坡的是 Win Semi。 Win 的体量非常大。$AVGO、$LITE、$QCOM、联发科、$MTSI、$NXPI 全都在用他们。 他们也在 $AAPL 和 SpaceX 的供应链里。所以只要你锁定 Win 的产能，超大规模云厂商的量产放量就是可去风险的。
英文原文
So $SIVE doesn't handle capacity scaling, Win Semi does volume ramp. Win is massive. $AVGO, $LITE, $QCOM, Mediatek, $MTSI, $NXPI, all use them. They're in $AAPL, SpaceX supply chains too. So volume scaling for hyperscalers is derisked if you secure allocation and do it through Win.
-
POET 被替代意味着 Marvell 直连 SIVE
所以 $POET 做的是封装那一层的事情。 如果 $MRVL 放弃了 Poet，意思就是 Marvell 认为自己可以把那一层垂直整合进 Celestial，并且直接自己买激光器（不只是 nda 纠纷那么简单）。 Marvell 还是得买激光器。我相信他们会准备多来源合同，但我不认为 $NVDA 会把分配额度分给 $MRVL。 这也是为什么 $SIVE 很可能仍然是直接供货给 Marvell 的激光供应商，因为它们已经符合 Celestial 的规格，这样就能直接去掉一个中间层玩家。
英文原文
So $POET does the packaging side of things. If $MRVL dropped Poet, it means Marvell felt they could vertically integrate that layer into Celestial and just buy lasers directly themselves (not just the nda dispute). Marvell would still need to buy lasers. I'm sure they're planning multi-source contracts but I don't think $NVDA would share allocation to $MRVL. Which is why $SIVE is likely still the laser supplier direct to Marvell since they fit celestial specs already, and that just cuts out a middleware player.
-
日本散户持仓是好信号
我很高兴日本社区在研究之后开始建 $SIVE 仓位！ 更强的国际股东基础总是正面的。 至于我的一些看法，市场现在看起来像这样： 1. $NVDA 把 $LITE / $COHR 的分配额度买走了 2. $AMD 的 CPO 走的是 $GFS + $SIVE / Win，剩下的激光供应可能会给 $LITE，如果还有分配的话。 3. 还有……$MRVL 的 CPO 不管怎样都需要激光器。 $SIVE 看起来是最后几家纯 merchant laser 供应商之一。 所以 Marvell 大概率会直接用 $SIVE（本来就符合 Celestial 规格），后面再多来源（也许是 $MTSI）。之后他们再把进入 Celestial 的中介封装 IP 垂直整合掉。 支持这个判断的有一些有意思的点： -> Ayar 把 $MTSI 和 $LITE 从官网上删掉了，改用了 $SIVE 作为主要供应商。Ayar 和 AlChip/GUC 等都有连接。 -> 如果看 $GFS 的幻灯片，在 $AMD 把 CPO 项目交给 GlobalFoundries 之后，只剩下 $SIVE 和 $LITE 两家公众公司。 -> $SIVE 很可能从去年开始就和 Win 有激光产能爬坡协议。 $NVDA 很可能还没把这些激光产能完全分配完，所以像 $JBL、$AMD 以及其他公司会把溢出订单给 Sivers。 因为 $LITE 已经暗示 2028 年的产能基本满了。 我可能也有错，但基于公开信息来看，情况就是这样。 至于为什么我认为它是个好长线： -> Sivers 基本上过去没怎么暴露在 800G 或更早的代际里。 -> 欧洲市场按过去 12 个月营收定价……所以之前估值一直被压得很低。 -> 它们拿到了 $NVDA 造成的超额需求。 但它们也正好站在 CPO 和甚至下一代 1.6T（$JBL LRO）扩产的最前沿，明年就会开始。 然后到 2027 年下半年或 2028 年，还能扩到硅光等相邻领域，比如可能的 $AAPL 消费设备。 或者 FMCW 4D AI 公司，比如 $AEVA。 这么多年开发，终于在明年要开花结果了。 我个人觉得市场漏掉了一个大东西，公开信息会随着时间推移慢慢被大家通过供应链关系、网站挖掘或者演示材料发现。 超大规模云厂商供应商不会无缘无故选一家 10 亿美元市值的瑞典激光公司。 和 $JBL 的直接合同就是最大的信号。 这也是我为什么高 conviction 做多。
英文原文
I'm happy Japanese communities started positions in $SIVE after doing research! A stronger international shareholder base is always positive. As for some thoughts, my read on the market looks like: 1. $NVDA bought out allocation from $LITE / $COHR 2. $AMD CPO went with $GFS + $SIVE / Win for remaining laser supply maybe $LITE if there’s still allocation. 3. And… $MRVL CPO will need lasers regardless. $SIVE looks like one of the last remaining pure play merchant laser suppliers. So Marvell will go with $SIVE (fits Celestial specs already) directly with multi-source down the road (maybe $MTSI). After they vertically integrate away interposer packaging process IP that feeds into Celestial. Just some interesting things to back that up: -> Ayar removed $MTSI and $LITE from their website and went with $SIVE as primary. Ayar’s connected to AlChip/GUC and others. -> If look at the $GFS slide there's only two public players with $SIVE and $LITE after $AMD went with Globalfoundries for their CPO program. -> $SIVE likely has agreements with Win since last year for laser capacity scaling. $NVDA likely hasn't fully allocated that laser supply, so the remaining companies like $JBL, $AMD, and others go to Sivers for overflow. Since $LITE signaled they were already fully allocated for 2028. I could be wrong, but just based on public information that’s what it looks like. As for why I think it's a good long: -> Sivers also basically had no exposure to 800G or previous generations. -> European markets price in previous 12 months revenue... hence previous depressed valuations -> they get all the hyperscaler overflow created by market panic from $NVDA But they also happen to be in the bleeding edge of CPO and even for gen-2 1.6T ( $JBL LRO) scaling next year in 2027. Then for H2 2027 or 2028, they scale in adjacent areas like Silicon Photonics for likely $AAPL consumer devices. Or FMCW 4D AI companies like $AEVA. Many many years of development, finally coming to fruition next year. I personally think markets are missing something big here, that the public uncovers over time with mapping hyperscaler relationships, website digging, or presentation slides. Hyperscalers suppliers don't randomly choose a $1B Swedish laser company for no reason. The direct contract with $JBL was the biggest signal of that. And it’s my high conviction long moving forward.
-
所以我就直说了：市值 13.5 亿美元的 Towa（6315）…… 它是 HBM4（压缩成型）上很罕见、活生生的垄断定义。 虽然今年到现在几乎没怎么涨，但像 $MU、SK 海力士、三星这些存储厂商都是他们的客户。而且每一家存储厂商的财报都在指向大幅增加 capex。 就连从 $TSM 的财报也能看出来，所有主要半导体公司都在经历一轮巨额资本开支周期，以满足 AI 需求。 三大存储厂商也都从 HBM3E 和 NAND 里赚得很猛……所以下一轮 capex 周期大概不会像上一轮那样了（意思就是花得会更多）。 像 $ASML 一样，这种东西周期性极强，但我想在未来几个月的 HBM4 capex 爆发里拿到敞口。 我想把这个名字和 $LPK（玻璃核心基板）一起放到大家视野里，因为它们都是功能性垄断。 但我确实觉得现在这个时间点刚刚好，毕竟所有机器供应商都在大幅重估，而它相对还是平的。 这也不算什么新发现，因为之前还有别的分析师和一些随机关注者提过这个名字，但希望我这次把时点抓对了。 （免责声明：我确实持仓，这只是我自己想法的 TLDR，请不要跟单）
英文原文
So just putting it out there: Towa (6315), at $1.35B... Is a rare, living definition of monopoly over HBM4 (compression molding). It's been kinda flat YTD, but every memory company like $MU, Sk Hynix, Samsung are their customers. And each of the memory company earnings signaled massive capex increases. Even as seen with $TSM earnings, every major semi is going through a massive capex cycle to meet AI demand. And all three memory makers have printed from hbm3e and nand... so the next capex cycle is probably not like the last (meaning a lot more spend). Like $ASML, this is hyper-cyclical but I wanted exposure to the upcoming HBM4 capex ramp over these next few months. Thought I'd put this name on people's radar alongside $LPK (glass core substrates) as a functional monopoly. But I do feel like this timing is about right while every machine supplier is having a massive re-rating yet this was relatively flat. Not exactly a new find, since a few other analysts + random followers had this name but hope I get the timing right. (Disclaimer: I do hold positions, this just TLDR of my own thoughts, please don’t copy trade)
-
存储链条继续起飞
从 $MU 到 $EWY（SK 海力士、三星）再到 $SNDK，存储相关名字最近都在疯狂起飞。 除了三大厂之外，还有 NAND 或台湾的名字，比如南亚科。 现在也许是看看 Towa（6315）的好时机，用来捕捉 HBM4 上游垄断的 capex 周期？ https://t.co/VZPwzB0SFK
英文原文
Memory names from $MU to $EWY (SK Hynix, Samsung) and $SNDK are going brrrr lately. Aside from the big three, then your NAND or TW names like Nanya. Might be a good time to look at names like Towa (6315) to capture upstream monopoly capex cycles for hbm4? https://t.co/VZPwzB0SFK
-
供应链不同层级不能混为一谈
@soscoder 它们处在供应链的不同环节。就像问 $CMG 的墨西哥卷饼能不能当成 AI 领域的 $NVDA GPU 替代品一样。
英文原文
@soscoder They're different parts of the supply chain. It's like asking if $CMG burritos be an $NVDA GPU alternative to AI.
-
MTSI 拟投 IQE 4500 万美元
刚刚，$MTSI 计划向 $IQE 投资 4500 万美元，并进入董事会。 我说什么来着，匿名网友？？Landmark 早就产能过剩了…… + IQE 在光子供应链里本来就是结构性卡点，市值才 1 亿美元时就已经是这样了。 这笔交易也是为了锁定长期外延晶圆供应。 https://t.co/lS2TGFI6v6
英文原文
Just now, $MTSI looking to invest $45 Million into $IQE and sit on the board. What did I say anon??? Landmark was overcapacity... + IQE was a structural chokepoint in photonics the supply chain back at $100M MC. This deal is also to secure long term epiwafer supply. https://t.co/lS2TGFI6v6
-
AAOI 也可能是潜在客户
谢谢提醒。 $AAOI 也有可能是其中之一，因为他们在 OFC 上宣布了一个 ELSFP。 他们通常都是自家代工，然后卖成品光模块。他们有这种可选性。 但我不认为这和 $SIVE 设计进 Celestial 的架构是一样的？我得再研究一下。
英文原文
Thanks for reminding me. It's possible for $AAOI could to be one since they announced a ELSFP at OFC. They usually just do captive and sell the finished optical transceivers. They have the optionality to. But I don't think it matches the same architecture that $SIVE designed into Celestial? Will need to look into it more.
-
MRVL 可能直接向 SIVE 采购
@mpunaskar 这件事挺微妙的，不算完全正面，但也可能最终更正面。 我猜 $MRVL 很可能会直接向 $SIVE 采购。 直接对接超大规模云厂商，比如 Sivers + $JBL，要比中间再加一家公司正面得多。
英文原文
@mpunaskar It's nuanced, it's not positive but could end up more positive. It's very likely $MRVL just buys directly from $SIVE is my guess. Going direct to hyperscaler like Sivers + $JBL is much more positive than adding in another company in the middle.
-
材料和封装要分开看
它们都处在供应链的不同层级。$AXTI 生产的是光子相关一切所需的材料。我手里还是有 AXT 仓位。 如果超大规模云厂商垂直整合带来一些细节变化，封装层的玩家可能会受影响。 但直接卖掉任何激光或上游玩家都没什么道理，对一些人来说反而是不错的买点。
英文原文
They're all different parts of the supply chains. $AXTI makes the materials needed for anything to do with photonics. I still have my AXT positions. Packaging layer players might be impacted given nuances of hyperscaler vertical integration. But selling off any laser or upstream player doesn't make much sense and it's a good buying opportunity for some.
-
激光和封装是两码事
这显然不是一回事。激光和封装在光学供应链里是两个实质上不同的环节。 如果我的 thesis 失效了，我会像当初对 $XLU 和降息那样坦白说出来。 但我对 $SIVE 仍然非常看多，因为我还是认为 $MRVL 会为了 Celestial 直接从他们那里买，然后把 $POET 的封装流程垂直整合掉。
英文原文
It's clearly not. Laser and packaging are two materially different parts of optical supply chains. I would be transparent if the thesis breaks like I did with $XLU and rate cuts. But I'm incredibly bullish on $SIVE since I still think $MRVL just buys directly from them for Celestial and vertically integrates the $POET packaging process.
-
LITE/COHR 产能大概率已给 NVDA
$LITE 和 $COHR 可能已经把大部分甚至全部产能分配给了 $NVDA，正如 Nvidia 去年对 EML 做的那样。 由于 Win 是 merchant foundry，想对它这么做更难，所以 $MRVL 大概率会直接去找 $SIVE，因为他们本来就符合 Celestial 的规格。
英文原文
$LITE and $COHR are probably majority/fully allocated to $NVDA after their deals. Since Nvidia did the same thing with EML last year. It's harder to do that with Win since they're a merchant foundry, so it's likely $MRVL just goes directly to $SIVE since they already fit Celestial specifications.
-
MRVL 长期受益于垂直整合
@O_bhaina $MRVL 长期会从把封装流程垂直整合中受益。 我猜他们很可能只是达成一笔交易，然后直接从 $SIVE、$MTSI，也许还有 $LITE 那里买激光器。
英文原文
@O_bhaina $MRVL benefits long term by vertically integrating the packaging process. They will likely just make a deal and buy lasers directly from $SIVE, $MTSI, and maybe $LITE is my guess.
-
LITE 更可能供给 NVDA
@adrian_jega $LITE 的大部分产能大概率是分给了 $NVDA，而不是 $MRVL。 Ayar 把 $LITE 从官网上移除了，而且在这之前他们主要是在 $SIVE、$MTSI 和 $LITE 之间多来源采购，这其实就是一个提示。 所以他们会需要别的激光合作伙伴。
英文原文
@adrian_jega $LITE likely has majority capacity allocated to $NVDA, not $MRVL. Ayar removed $LITE from their website and before they primarily multi-sourced between $SIVE, $MTSI, and $LITE which was a hint to that. So they'll need to find other laser partners.