概念词典 · 技术
CPO 共封装光学
把光收发器封装到 ASIC/交换芯片旁边的封装级集成方案，绕过传统可插拔光模块的功耗、带宽与信号衰减瓶颈。
别名： CPO／ 共封装光学／ 共封装／ co-packaged optics／ co-packaged-optics
是什么
传统数据中心里，光信号从光模块（pluggable transceiver）出来要走一段电信号才到 ASIC，走电信号的这段距离正在成为 AI 算力扩展的硬瓶颈：信号衰减、功耗陡升、PCB 板上挤不进更多通道。
共封装光学（Co-Packaged Optics, CPO） 把光引擎直接焊到 ASIC 旁边的封装基板上，电信号路径从厘米级缩到毫米级，单端口能耗下降 30-50%，单封装可堆叠的端口数翻倍。
为什么这是 AI 时代的关键
GB200 / B200 这一代 GPU 之后，每颗 GPU 的对外 IO 带宽要求超出传统光模块物理极限。要么继续用昂贵的电气 retimer 续命，要么换 CPO。NVIDIA 在 2024 GTC 已确认 Quantum-X800/Spectrum-X800 走 CPO 路线，AVGO/Marvell 紧跟。
Serenity 的视角
他从 2024 年开始反复强调 “CPO is the photonics chokepoint”——光模块行业表面繁荣，但真正能在 CPO 时代活下来的是少数掌握 DFB 激光器、CW 激光、磷化铟（InP）衬底的上游公司，而不是组装环节。
代表案例：
- $COHR Coherent：CEO 多次公开确认 CPO 是公司”最重要的长期增长机会”
- $SIVE Sivers Semiconductors：DFB 激光器纯粹玩家，三个月涨幅 1900% 的传奇标的
- $LITE Lumentum：传统光模块巨头，受益于 CPO 但稀释更严重
- $POET：被他列入 “Fake it Until you make it” 警示名单，不是所有 CPO 故事都是真的
时间线（行业共识）
- 2026H2: 初始 scale-out CPO 收入开始
- 2027H2: scale-up CPO 收入跟进
- 2028+: 主流化，传统 pluggable 收入开始下滑
提到此概念的推文 (453)
按时间倒序 · 滚到底自动加载更多
-
LITE财报证实CPO 27H2放量，CW激光供需失衡且中国厂商落后。
$LITE 财报电话会要点： - “我们对共封装光学(CPO)规模化部署的时间表可见度已提高。我们确信，2027年下半年我们的超高功率激光芯片需求将强劲增长。” Lumentum 确认其领先客户将于 2027 年下半年启动 CPO 规模化放量。 - “我们最近获得了首个外部光源(ELS)模块订单，将于 2027 年下半年交付。” ELS 行业放量时间表。 - “非平面光学(NPO)机会对我们完全是增量，显著增加了光学市场的总规模(TAM)。” 激光瓶颈俱乐部的 TAM 进一步扩大。 - “我们预计 1.6T 光模块的采用将从财年第一季度开始加速，并持续至 2027 年。” - “即使在连续波(CW)激光领域，我们对自己提价的能力也感到惊讶。” $SIVE 及其他 CW 激光玩家拥有积极的定价权。 - “因此，我们能够获得可观的价格溢价，并预计这种溢价将维持下去，即使这些中国厂商上线。我也提醒人们，我认为一些中国激光供应商目前并未在市场上实际交付，所以当它们抛出那些大数字时，没有追索权。迄今为止，我们在其产出方面尚未看到类似情况。” 中国厂商在激光/产能方面落后于市场预期，得到 $AAOI 和 $LITE 的证实。 - “不幸的是，我们在高功率激光的发货上远远落后。” “需求信号增加，相对于我们的供应能力，我们更加落后。” 超高功率(UHP) CW 激光产能存在极端供需失衡，EML 需求持续失衡。年底发货约 +5000 万美元 -> 1 亿美元+（“三位数季度”），进入 2027 年下半年和 2028 年更大的规模化需求。 - “所以，我认为过去三个月我们真正试图做的是确保更多的衬底供应”（因为供需失衡），“我们从 $AXTI 找到了额外的衬底帮助。他们是伟大的合作伙伴。” 磷化铟(InP)衬底瓶颈出现。 要点总结： - Lumentum 确认 CPO 2027 年下半年放量，驳斥任何“延迟”新闻为市场“噪音”。（一份延迟报告造成了巨大损害，Nvidia、Lumentum 等公司花了数周进行危机公关）。 - CW 激光需求在 2027 年下半年放量。 - 超高功率 CW 激光产能的极端供需失衡让 Lumentum 感到惊讶。 - 超高功率激光拥有极高利润率，并能对 CW 激光产能提价。 - 中国在激光/产能方面进一步落后。 基本上验证了西方 CW 激光玩家的时间表 + 需求/定价。 在主要放量开始之前，CW 激光的稀缺性已经发生……
英文原文
$LITE earnings transcript TLDRs: - "Our visibility into the timing of CPO scale-up deployments has also sharpened. We remain confident in a demand ramp for our ultra-high-power laser chips in the second half of calendar 2027" H2 2027 start of CPO scale up ramp confirmed from Lumentum for their lead customer. - "We were recently given our first external light source, or ELS, module purchase for delivery by the second half of calendar 2027" ELS industry ramp timelines - "The NPO opportunity is completely additive for us, significantly increasing the optical TAM." More TAM for laser chokepoint club - "we expect the 1.6T transceiver uptake to intensify starting in fiscal Q1 and sustain through calendar 2027." - "Even in CW, we’ve been surprised at our ability to price up" Positive pricing power for $SIVE and the other CW laser players. - "As a result, we are able to command a nice price premium that we expect to sustain, as these Chinese if these Chinese guys come online. I caution people also, I think people are some of these Chinese laser suppliers are not delivering in the market today, so they have there is no recourse when they throw out these big numbers. We have not seen anything like that to date in terms of their output." Chinese players are more behind than markets expect, corroborated by $AAOI and $LITE. - "We are way behind in our shipments, unfortunately, on high-powered lasers" "The demand signal has increased, and we are very much further behind relative to our ability to supply." Extreme demand imbalance for UHP CW laser capacity, continued demand imbalance for EML. Shipping roughly +$50M year end -> $100m+ ("triple digit quarter") into much larger H2 2027 and 2028 scale up demand,. - "So what we are trying to do, I think, really in the last three months is secure even more substrate supply" (because of demand imbalance), "we found additional substrate help from $AXTI They are a great partner." Hello InP substrate bottleneck. TLDR: - Lumentum CPO scale up H2 2027 shipments got confirmed, dismissed any "delay" news as market "noise". (one delay report caused so much damage, and it took Nvidia, Lumentum, and all these companies weeks to do damage control). - CW lasers demand ramp second half of 2027 - Extreme demand imbalance for UHP CW laser capacity surprised Lumentum - UHP lasers commands very high margins and is able to price up CW laser capacity. - China further behind in the laser/capacity angle. Basically just validated timelines + demand/pricing for Western CW laser players. CW laser scarcity is already happening before the main volume ramp even begins...
-
AAOI产能受限，优先自供激光器，CPO部署短期承压。
@Berlinergy $AAOI：“我们根本无法生产足够的[共封装光学(CPO)激光器]来满足其当前第一代[CPO]部署的需求，因为产能确实不足。我们必须优先确保为自家光模块生产激光器。”这仅适用于早期CPO相关部署的初步解读。
英文原文
@Berlinergy $AAOI: "We just can’t make enough of them [CPO lasers] to be involved in their current first-generation [CPO] deployments because there’s just not enough capacity. We have to prioritize our ability to make lasers for our own transceivers first" This is just for early CPO related deployment read through.
-
LITE财报超预期，指引强劲，看好CPO渗透及同行受益。
$LITE 财报出炉，表现非常强劲： 营收：10.1亿美元，环比+24.5%，同比+109.3% vs. 预期约9.846亿美元 调整后每股收益(Adj. EPS)：3.23美元 vs. 预期约2.95美元 毛利率(非GAAP)：50.4% / 营业利润率(非GAAP) 36.6% FY2027 Q1指引可能更重要： 营收：12.25亿–12.75亿美元 vs. 预期约11.6亿美元 调整后每股收益：4.05–4.35美元 vs. 预期约3.63美元 营业利润率(非GAAP) 39.5% - 40.5% 关于备注： 看到：“对超高功率共封装光学(CPO)激光器的需求增加，ELS模块获得初始订单” - 对其他CPO玩家如 $SIVEF、$COHR 等有很好的传导效应(read through)。(可能不包括 $AAOI，因为它们错过了第一代部署) - “NPO参与的广度是光学开始渗透机架内连接性的初步迹象，显著提升了我们的光学总可寻址市场(TAM)” - 营收预测“比目标模型提前一个多季度达成。” 正如预期的极其强劲的财报，营收从： 8.08亿 -> 10.1亿 -> 12.5亿，同时营业利润率持续增加... 这只是我的初步印象。 最重要的是即将召开的财报电话会议。
英文原文
$LITE earnings are out, very strong earnings: Revenue: $1.01B, +24.5% Q/Q, +109.3% Y/Y growth vs. ~$984.6M Adj. EPS: $3.23 vs. ~$2.95 Gross Margin (non-gaap): 50.4% / Operating margin (non-gaap) 36.6% Q1 FY2027 guidance is probably more important: Revenue: $1.225–1.275B vs. ~$1.16B Adj. EPS: $4.05–4.35 vs. ~$3.63 Operating margin (non-gaap) of 39.5% - 40.5% In terms of notes: Sees: "Increasing demand for ultra-high-power CPO lasers, an initial order for ELS modules" - Great read through on other CPO players like $SIVEF, $COHR, and others. (maybe not AAOI since they're missing out on first-gen deployments) - "Breath of NPO engagements are the first signs that optics are starting to penetrate in-rack connectivity, significantly upping our optical TAM" - Revenue projections "reaching our target model more than a quarter ahead of schedule." Extremely strong earnings as expected, you have revenue going from: $808M -> $1.01B -> $1.25B while operating margin keeps increasing... Just my first impressions. Most important thing is the earnings call coming up soon.
-
AAOI 多技术周期叠加，2028 年 CPO 产能将带来巨大增量。
给那些认为 $AAOI 只是单年周期股的人提个醒： $AAOI 预计其用于共封装光学(CPO) 的 ELSFP 产能将在 2028 年达到每月 40 万片。 40 万 * ~$400 ELS 平均售价(ASP，高盛假设) = 新增约 19.2 亿美元/年的年化营收产能，作为一条全新且独立（毛利率 >50%）的产品线。 这叠加在他们现有的 2027 年预测之上（例如，下半年起月营收 4.71 亿美元，对应年化光模块营收约 56 亿美元）。 1.6T 的潜在市场(TAM) 也在飙升，所以我预计随着更多产能上线，他们 2027 年下半年的预测值会进一步上调... 对于某些光通信标的，这是一个周期（例如 1.6T、CPO 规模化、NPO 等）叠加在另一个周期之上... 再叠加另一个... 随着 TAM 和利润率像《英雄联盟》中艾尼维亚使用 W 技能后的小精灵一样层层堆叠。 而不是仅仅依靠单一年份的“一次性”爆发。 这与某些其他板块形成鲜明对比，那些板块在经历可能仅一年的三位数同比营收增长后，增长可能会放缓。
英文原文
Just putting it out there for people that think AAOI is a one year cycle... $AAOI projects their ELSFP capacity for CPO to be 400K/units a month in 2028. 400k * ~$400 ELS ASP (GS assumptions) = + another ~$1.92B 2028 annualized revenue capacity added as a new distinct (>50% gross margin) product line. On top of their existing 2027 projections (eg. ~$5.6B annualized transceiver revenue off $471m/month entering H2). TAM for 1.6T also goes brrr so I'd expect their end of H2 2027 projections to go up as more capacity comes online... For certain optical names, it's one cycle (eg. 1.6T, CPO scale out/up, NPO etc. ), stacked on top of one another... stacked on top of another... with TAM + margins stacking like minions after Anivia uses W in line. Rather than one-and-done off of one year. This is a stark contrast to some other sectors where growth is likely to decelerate after maybe 1 year of triple digit Y/Y revenue growth.
-
OCP峰会及财报周催化光模块与机器人板块，Unitree IPO或带动机器人行情。
近期的一些事件： - 明天是OCP亚太峰会（Ayar、Lightmatter、$AMD、$NVDA），你的共封装光学(CPO)概念股正在发布公告/更新。 这应该是某些光模块玩家的催化剂。 - 财报周：$ASTS、$RKLB今天发布（太空），$LITE周二（光子学），$NBIS在本周中段（Neocloud），还有更多。 - Unitree IPO认购今天开启，预计本月上市。 如果开盘良好，应是人形机器人+机器人板块的潜在催化剂，例如 $CCXI。 接下来的一周很有趣。
英文原文
Just some near term events: - OCP APAC tomorrow (Ayar, Lightmatter, $AMD, $NVDA) and your CPO players are giving announcements/updates. This should be a catalyst for certain optical players. - Earnings week: with $ASTS, $RKLB reporting today (space), $LITE on Tuesday (photonics), $NBIS in the middle of the week (Neocloud), and more. - Unitree IPO subscriptions opened up today and Listing is expected this month. Should be a potential catalyst for the humanoid + robotics sector if it opens up well eg. $CCXI. Fun week ahead.
-
看好SIVE获代工产能及上市融资，预期2027H1收入拐点。
@wannabepanacea 鉴于目前业界普遍传闻共封装光学(CPO)中的分布式反馈激光器(DFB)短缺，我其实挺好奇 $SIVE 能否获得更多晶圆代工产能分配。因为我相当确信，一旦其资格认证进入量产爬坡阶段，产品就会售罄。除此之外： > 其他财报（如 $MTSI / $AAOI）重申的需求/供应失衡有助于佐证这一点。 > 上一份财报中提到的与可插拔(Pluggable)玩家的新进展/资格认证，可能是影响最大的因素。 > 与 $JBL、~ $AEVA、$POET、$GFS 或其他机会的进展，如果能获得初步采购订单(PO)那就太好了。 > 重申纳斯达克上市更新（我们已经得到消息，计划在未来几个季度完成）。我个人仍将此列为优先事项，因为它能为其并购(M&A)活动解锁更多资金，希望用于可插拔技术栈。 我认为收入拐点大约在 2027 年上半年开始，因此当前的收入/利润数据并不重要。 如果 $AAOI 能发表类似“我们制造的任何东西都预计在 2028 年售罄”的声明，那就太棒了。
英文原文
@wannabepanacea I’m actually kinda curious about $SIVE securing more foundry allocation given widespread CW DFB shortages now. Since I’m pretty confident it sells out once qualifications go into volume ramp. But aside from that > reaffirm demand/supply imbalance from other earnings like $MTSI / $AAOI helps > new development/qualifications mentioned in previous earnings with pluggable players, probably most material > Progress with $JBL, ~ $AEVA, $POET, $GFS or other opportunities, maybe initial POs would be nice. > Nasdaq listing update reiterated (we already got updates it’s planned completion next few quarters). I still place that on priority personally since it unlocks more funding for their M&A efforts, hopefully down the pluggable stack. I think the revenue inflection starts around H1 2027, so current revenue/profit numbers are not very material. An $AAOI statement like anything we make we expect to be sold in 2028 or something would be kinda goated.
-
InP激光器短缺利好SIVE，可套用AAOI产能模型。
我认为现在是重新审视 $SIVE 的时机，其年化产能收入的中位预测约为 4.27 亿美元。 鉴于近期的行业财报： > $AAOI 指出光模块（transceivers）需求存在 20-40% 的失衡（激光器是瓶颈）。 > $MTSI 表示许多“客户因磷化铟（InP）DFB 激光器普遍短缺而紧急联系我们”（在其他产品线中）。 > 美国提议禁止新的中国光模块（中际旭创/新易盛）。 > AOI 因激光器被分配给光模块，未实质性参与第一代共封装光学（CPO）部署。 解读是：可以将适用于 $AAOI 的产能-需求吸收模型，同样应用于 Sivers 等其他激光器厂商。 鉴于 InP CW DFB 产能的广泛短缺和需求失衡。
英文原文
I think it’s a good time to revisit $SIVE ~$427m/yr midpoint capacity revenue modeled projections. Given recent industry earnings: > $AAOI cited +20-40% demand imbalance for transceivers (with lasers as the bottleneck). > $MTSI saying many "Customers are coming to us with urgency due to the general supply shortage of indium phosphide DFB lasers" (in other product lines) > US proposal ban on new Chinese optical transceivers (Innolight/Eoptolink) > AOI no meaningful participation in first-gen CPO deployments due to lasers being allocated toward transceivers. The read through is that it’s possible to apply the same capacity-demand absorption modeling applying to $AAOI, but for other laser players like Sivers too. Given the widespread shortage and demand imbalance for InP CW DFB capacity.
-
激光器及InP衬底未来2-3年或现短缺，CPO扩产加剧瓶颈。
@SL2056 @Sofigoodboy 好问题！这是逐年变化的，很难从2026年预测到2030年的短缺，但这绝对是一个结构性的多年需求周期。 我想指出的是，激光器（lasers）在特定领域很可能在未来2-3年面临极端短缺，尤其是在共封装光学(CPO)规模化初期至少如此。 例如，CPO激光器市场所需的产能“比今天大8到10倍”，并补充说：“我认为……完全没影响，因为需求远大于全球产能。” 至于磷化铟(InP)衬底，$AAOI 表示2027年没问题，但由于其关键性希望组建合资企业。 $AXTI 看起来只是逐年倍增产能，具体取决于你问谁。 一些台湾外延片(epiwafer)厂商称InP衬底面临极端短缺，而它们正是激光器厂商的上游。 你的激光器厂商如 $AAOI 和 $MTSI 表示他们没问题。 无论如何，我们看到这转化为 $AXTI 的收入拐点（2027年约1.3亿美元/季度），但我预计鉴于其市场地位，平均售价(ASP)一路上涨。
英文原文
@SL2056 @Sofigoodboy Good question! This something year-by-year, can't really predict shortages into 2030 from 2026, but it's definitely a structural multi-year demand cycle. Something that I want to flag is that it looks like lasers in specific would likely have extreme shortages for the next 2-3 years, especially during CPO scale up at the very least. eg. CPO laser market would require capacity "8 to 10x bigger compared to today," and added, "I think… no effect at all because the demand is much bigger than the worldwide capacity." As for InP substrates, $AAOI said they were fine for 2027 calendar year but wanted to form JVs because of likely how critical they are since. $AXTI looked like they were just multiplying capacity year after year and depends on who you ask.. Some of the Taiwan epiwafer players say it's extreme shortages for InP substrates, when they feed into your laser players. Your laser players like $AAOI and $MTSI say they're fine. Regardless, we see that convert into revenue inflection points with AXTI anyway (~$130m/quarter in 2027), but I'm expecting ASP to be hiked along the way given their positioning.
-
Rosenblatt 光通信调研精准，AAOI 财报证实中国 CPO 激光器落后。
我通常喜欢调侃卖方分析师，但 Rosenblatt 对光通信标的的渠道调研（Channel Checks）确实非常出色，他们的报告是我少数愿意讨论的。$AAOI、$LITE、$SIVE 以及激光板块已开始在其预估的时间窗口内恢复。例如，上个月他们提到：“多位空头表示他们可能会在 7 月底和 8 月初平仓”。他们可能误判了 $AMD 会是 $AAOI 的首个共封装光学(CPO) 客户……因为 AAOI 表示他们将缺席第一代 CPO（也许是未来几代？）。但他们关于中国 CPO 激光器定位（例如落后 2-3 年）的许多观点，今天被 AAOI 的财报逐字证实了。
英文原文
I usually make fun of sellside, but Rosenblatt has pretty goated channel checks on optical names, and their reports are one of the few I like talking about. $AAOI, $LITE, $SIVE, and the laser party has started to recover around the est. timing windows. eg. last month: "Multiple short sellers told them they will likely close their positions... late July and early August". They might have missed that they believed $AMD would be $AAOI's first CPO customer... since AAOI said they would be missing from first-gen CPO (maybe future gen?) But lot of their talking points about China CPO laser positioning (eg. 2-3 years behind) got corroborated word-for-word from AAOI earnings today.
-
博主兴奋于CPO激光器格局及CW产能瓶颈研究获验证。
@Financewisdom01 我个人真的非常兴奋，因为这些细枝末节的线索验证了我几个月前做的很多研究。关于共封装光学(CPO)激光器玩家格局以及连续波(CW)激光器产能瓶颈。
英文原文
@Financewisdom01 I’m just really excited personally since these little breadcrumbs validated a lot of the research I did months ago. With CPO laser player dynamics and CW laser capacity bottlenecks.
-
@pjtesus 这是因为 $AAOI 的 800G/1.6T 光模块（optical transceivers）已经存在巨大的 20-40% 供需失衡…… 他们无法生产足够的产品来满足需求。这是一件好事。 但无法满足所有需求，意味着较小的共封装光学(CPO) 激光器玩家存在市场机会。
英文原文
@pjtesus It's because $AAOI already has a massive 20-40% demand imbalance for their 800G/1.6T optical transceivers... They can't make enough to satisfy demand. That's a positive thing. But not being able to fulfill all demand, means there's market opportunities for smaller CPO laser players.
-
AAOI缺席CPO部署揭示产能瓶颈，暗示其他CPO玩家价值被低估。
@00m00oo 我对 $AAOI 极度看好，它是我的核心持仓之一？ 但当 AOI 表示因产能不足而不参与第一代共封装光学(CPO)部署时，背后有无数问题需要拆解。 1. 独立连续波分布反馈激光器(CW DFB)产能，比市场预期的更少。 2. 合格的 CPO 激光器玩家，比市场预期的更少。 3. 我记得 Rosenblatt 曾认为 $AAOI 会是 $AMD 的第一个 CPO 客户，但情况可能已变... 而且由于 $LITE 和 $COHR 站在 $NVDA 阵营且未来两年订单已售罄... 其他一些 CPO 玩家可能比人们认为的更有价值。
英文原文
@00m00oo I'm extremely bullish on $AAOI, it's one of my top holdings? But there's an unholy amount of things to unpack when AOI said they're not participating in first-gen CPO deployments because there's not enough capacity. 1. Independent CW DFB capacity, even less than markets thought. 2. Qualified CPO laser players, even less than markets thought. 3. I remember Rosenblatt thought $AAOI would be $AMD's first CPO customer, but might not be the case anymore... And since $LITE and $COHR are in camp $NVDA and sold out for next 2 years... Some other CPO players might be more valuable than others think.
-
AAOI财报揭示CPO激光器产能瓶颈，利好少数西方厂商，中国技术落后数年。
哇，$AAOI 的财报中对于 $SIVE 及其他激光器玩家来说，利好一个接一个。 1. $AAOI 并未实质性参与第一代共封装光学(CPO)部署。 “我们根本无法生产足够的[CPO激光器]来参与他们当前的一代[CPO]部署，因为产能根本不够。我们必须优先保障自有光模块的激光器供应。” 因此，第一代CPO激光器的部署范围可能会进一步缩小，仅限于$SIVE、$LITE、$AVGO和$COHR等西方厂商。 如果$AAOI不得不拒绝更多客户，独立的CPO连续波(CW)激光器产能将变得更有价值。 还要记得Trendforce曾提到$AMD签署了CW长期协议(LTA)？$COHR/$LITE已与$NVDA签署了产能协议…… 我原以为是$AAOI，但他们可能出局了。$MTSI在早期CPO参与中也不显眼（除了NPO外财报未提及，且从Ayar网站移除）。 所以我在想$AMD在Helios项目中会与谁合作（例如Sivers+Ayar现在更可能是候选者）？ 顺便说一句，这对$AAOI并非看空，因为他们光模块业务需求旺盛。只是对少数有产能的合格CPO激光器厂商更加看多。 2. 磷化铟(InP)激光器/光模块的需求失衡与瓶颈。 AOI CEO的话某种程度上证实了这一点：“客户需求比扩建后的产能高出20%-40%。” “我们几乎每周都从几家大客户那里收到这种需求。激光器目前是光模块业务最大的瓶颈。” 3. 中国在CPO波分复用(DWDM)规格激光器方面落后数年。 我之前在渠道调研CPO竞争时曾覆盖过这一点。 但$AAOI确认中国“至少落后两三年甚至更久”才能拥有CPO激光器，这极大地印证了西方激光器在CPO激光器卡脖子环节上的防御性地位。 总结： - CW激光器和光模块需求严重失衡。 - $AAOI因产能限制未进入第一代CPO（对$AAOI不算看空，因其光模块需求旺盛，但对$SIVE等少数有产能的独立厂商更是利好）。 - 中国在CPO激光器方面落后数年。
英文原文
Wow, there's gem after gem in $AAOI earnings for $SIVE + other laser player readthrough. 1. AAOI not meaningfully participating in CPO first-gen deployments. "We just can’t make enough of them [CPO lasers] to be involved in their current first-generation [CPO] deployments because there’s just not enough capacity. We have to prioritize our ability to make lasers for our own transceivers first" So first-gen CPO laser deployments, might even be narrowed down further for Western players like $SIVE, $LITE, $AVGO, and $COHR. And independent CPO CW laser capacity became more valuable if $AAOI had to turn away more customers... Also remember Trendforce was talking about $AMD singing CW LTAs? $COHR / $LITE have capacity signed with $NVDA ... I thought it would be $AAOI, but they might be out of the equation. $MTSI also doesn't look meaningful with early CPO participation (eg. no mention in ER aside from NPO, and removal from Ayar website). So I wonder who players like $AMD is going to go with for Helios (eg. Sivers + Ayar more likely candidate now)? Btw, this is not bearish AAOI because they have too much demand for optical transceiver business. Just more bullish on the existing few qualified CPO laser names that have capacity. 2. Demand imbalance and bottleneck for InP lasers / optical transceivers. AOI's CEO stated kinda supported that when they said: "The customer demand is 20%-40% higher" than expanded built out capacity. "We are getting this kind of demand from several big customers almost every week. Lasers are the biggest bottleneck right now for the transceiver business" 3. China being years away from having CPO DWDM specification lasers. I covered this earlier when looking at CPO competition from channel checks. But AAOI confirming that China is "easily at least two, three years or even longer from having CPO lasers is incredible tread through on defensibility Western laser positioning in the CPO laser chokepoint. TLDR: - High demand imbalance for CW lasers and optical transceivers. - $AAOI not in first-gen CPO due to capacity constraints (not exactly bearish AAOI because they have too much demand for their optical transciver business, but even better news for the few independent players with capacity coming online like $SIVE) - China years behind in CPO lasers.
-
AAOI财报验证光模块需求远超供应，产能爬坡中，看好板块传导。
以下是关于 $AAOI 财报的简要要点： - 预计超大规模客户将在未来几周内完成 1.6T 产品的全面认证（有助于收入爬坡 #2，时间线） - 继续认为 AOI 将拥有美国最大的 AI 数据中心光收发器产能（在竞争对手可能被禁时期重申雄心） - 预计磷化铟(InP)产能设施将于 2027 年初投产（进一步爬坡的时间线参考） - 总产能接近每月 20 万颗/单位，年底 800G/1.6T 达 65 万颗/单位。2027 年底达 93 万颗/单位（这是我想要看到的爬坡） - “增加外部光源(ELSFP)的制造产能。这是用于共封装光学(CPO)的”。我们预计今年晚些时候及 2027 年扩大生产，最终在 2028 年达到每月约 40 万件（需要时间将其建模为收入） - “如前所述，我们多年来一直内部制造激光器。这使我们避免了影响行业其他公司的短缺”（CW/EML 激光器短缺期间的垂直整合看多逻辑） - 我们相信未来 CPO 将继续推动对高功率激光器的需求增加（CPO 板块论点验证） - “长期目标是将非公认会计准则毛利率恢复到 40% 左右” - “第二季度末现金等价物总计 5.088 亿美元” 我需要再次确认 ATM 是否完成 - “我们交付收入的能力，特别是 800G 产品，目前受限于我们的生产能力”“如果我们能生产更多，我们现在就能发货更多” 需求 > 供应验证。 - “大部分增加的产能将在美国。即便如此，让我说，正如我一直强调的，这还不足以应对客户需求。客户需求高出 20%-40%。” 整个板块的光子学需求验证惊人，阅读 $LITE, $SIVE / $JBL 等公司的传导效应令人惊叹。 - “在未来两三年内不会，尤其是需求如此之大。好吧？即使结合 AI，$LITE, $COHR, $AVGO 加在一起，未来几年满足客户需求仍然非常困难” 更多光通信板块需求验证。 - CPO 时间线：“如果你说的是 CPO 市场的高产量制造商，我会说是明年第三季度末左右”以及“我们已与至少五家客户密切合作” 如果你关心当前财报（我并没有仔细看） 营收：1.91 亿美元 vs 1.90 亿美元 每股收益：0.06 美元 vs 0.02 美元 简而言之：激光器 + 光通信板块需求传导非凡。某种程度上支持 Lumentum CEO 关于激光器短缺比内存短缺更严重的说法。 2027 年产能爬坡按计划进行。要映射拐点期与时间线，大约在明年年初，正如他们之前的财报电话会议所述。 AAOI 现在拥有客户。限制在于制造足够的激光器和收发器。
英文原文
Just some TLDR notes on $AAOI earnings: - Expects full qualification of 1.6T products by their hyperscaler customer within next couple of weeks (helpful revenue ramp #2, timelines) - Continues to believe that AOI will have the largest AI DC transceiver production capacity in the US (reiterating ambitions during a time when their competitors might get banned) - Expects facilities toward InP capacity to come online in early 2027 (timeline FYI in terms of further ramp) - Total capacity is approaching 200,000/u per month, EOY 650,000/u per month of 800g/1.6t. EOY 2027, 930,000/u (this is the ramp i want to see) - "increase our manufacturing capacity for our external light source or ELSFP. That's for co-packaged optics or CPO". we anticipate ramping production later this year and into 2027, ultimately culminating in about 400,000 pieces per month in 2028 (need some time to model this into revenue) - "As we have mentioned before, we've been manufacturing lasers internally for many years. This has allowed us to avoid some of the shortages that have affected others in the industry" (vertical integration bull case during CW/EML laser shortages) - We believe that in the future, CPO will continue to drive increased demand for high-power lasers (thesis validation on CPO sector) - "to our long-term objective of returning non-GAAP gross margins to around 40%" - "We ended the second quarter with $508.8 million in total cash equivalents" I need to double check if the ATM finished or not - "our ability to deliver revenue in general, and specifically when it comes to 800G products, is limited by our production capacity right now" "If we could produce more, we could ship more right now" Demand > Supply validation. - "Most of the increased capacity will be in U.S. Even so, let me say that, like I keep emphasizing, that is not good enough for the customer demand. The customer demand is 20%-40% higher." Unholy photonics demand validation across the whole sector, read through for $LITE, $SIVE / $JBL, and others is amazing. - "Not in the next two, three years, especially the demand is so big. Okay? Even combined AI, $LITE, $COHR, $AVGO all together, it's still very tough to meet the customer demand in the next few years" More optical sector demand validation. - CPO Timelines: "If you're talking about really high volume manufacturer [for CPO market], I would say more like the late Q3 next year" and "We have been working very close with at least five customer" If you care about current earnings (which I'm not really looking at closely) Revenue: $191M vs. $190m EPS: $.06 vs. $.02 TLDR: Extraordinary demand across the laser + optical sector read through. Kinda supporting Lumentum CEO statement that laser shortage is worse than memory shortages. 2027 capacity ramp on track. To map inflection period with timelines, would be around early next year, as stated in their previous earnings call. AAOI has the customers now. Limitation is making enough lasers and transceivers.
-
Aeva切入CPO/NPO光源领域，利好上游激光器供应商Sivers。
本周的意外之事……是 $AEVA 成为共封装光学(CPO)/近封装光学(NPO) 光源玩家。 背景：$SIVE 被视为向 Aeva（以激光雷达闻名）供应连续波分布反馈(CW DFB) 激光器的高确定性供应商。 今天，Aeva 盘后大涨 18.26%，回应了与一家光学引擎厂商签署联合开发协议(JDA) 的消息，该协议将在 2027 年下半年部署于某大型超大规模数据中心，并于 2028 年量产爬坡。 “光互联将利用这一基础，结合 Aeva 现有的大规模制造和晶圆代工供应链”用于： - ELSFP（类似 O-Net/Enablence + $SIVE），Sivers 可能是 CW 激光器及潜在放大器供应商 - 片上集成光源（架构/集成方面可能是 $SIVE DFB + Aeva PIC） 简而言之：Aeva 在 CPO/NPO 领域获得新总可服务市场(TAM) 和超大规模数据中心项目，对 $SIVE 等上游供应商构成利好传导。
英文原文
The unexpected thing of the week... Was $AEVA becoming a CPO/NPO optical source player. For background: $SIVE is viewed as the high confidence CW DFB laser supplier to Aeva (which is known for lidar). Today, Aeva reacted with +18.26% AH on news that they signed a JDA with an optical engine player for deployment at a major hyperscaler in H2 2027, production ramp 2028. "Optical Connectivity will leverage that foundation along with Aeva’s EXISTING high-volume manufacturing and foundry supply chain" for: - ELSFP (similar to O-Net/Enablence + $SIVE), with Sivers as the likely CW laser + potential amplifier supplier - On-chip integrated light source (likely $SIVE DFB + Aeva PIC on the architecture/integration side) TLDR: New TAM and a hyperscaler programs for Aeva for CPO/NPO, positive read through for upstream suppliers like $SIVE.
-
VIAV 澄清 CPO 无延迟，测试收入已落地，时间表符合预期。
今年最有趣的辟谣来自 $VIAV 关于共封装光学(CPO) 时间表的言论： “嗯，业界有很多关于延迟的讨论，比如‘哦，因为良率会较慢’。 那都是胡扯。CPO 和所有相关事物都在推进。” 他们表示，虽然始终存在问题，但流程正在改善，情况正在好转。 对此补充，CEO 表示： - Viavi 拥有 CPO 测试活动的采购订单(PO) - “我们本季度已经开始获得 [CPO 收入]，可能 12 月会开始加速。” 顺便一提，他们还表示 1.6T 正在“快速爬坡”，明年应与 800G 持平，并暗示 800G 需求有更长的尾部效应。 简而言之：看到 CPO 测试收入正从 $VIAV 流向 $FORM 等公司。英伟达(Nvidia) 也表示用于扩展的交换机侧 CPO 已提前进入量产。 至于垂直扩展(scale up) 时间表，部分厂商的早期生产应在 2027 年下半年，而其他 CPO 厂商 Ayar 表示 2028 年是放量拐点。 基本上，所有时间表似乎与之前相同，收入正开始首先惠及测试厂商。
英文原文
Funniest callout of the year from $VIAV on CPO timelines: "Well, there's been a lot of industry talk [about delays], like, 'Oh, because the yield is going to be slower.' That's all nonsense. CPO and all that thing is moving forward" They stated that Yes there's always issues, but the process is being improved and things are getting better. To follow up on that, the CEO stated: - Viavi has POs for CPO-testing activity - “We’re already getting [CPO revenue] this quarter, and probably in December it will start accelerating.” On a side note, they also stated, 1.6T is "ramping very quickly", should reach parity with 800g next year, and signaled longer tail for 800g demand. TLDR: Seeing CPO test revenues hit for $VIAV to $FORM around now. Nvidia also stated switch-side CPO for scale out and scale across has entered mass production earlier. As for scale up timelines, early production should be H2 2027 for some players, with other CPO players Ayar stating 2028 is the volume inflection period. Basically, all the timelines seem to be the same as before and revenue is starting to hit testing players first.
-
AI激光器供应危机加剧，验证AAOI等标的瓶颈逻辑。
$LITE CEO Michael Hurlston 在 RAISE 峰会上警告称，AI 数据中心（AI DCs）用磷化铟(InP)激光器的供应缺口： 正面临比存储芯片更严重的供应链危机。 凭借 Lumentum 的 5 座 InP 晶圆厂，其出货量将比客户需求低 30% 以上。 这在目前的电吸收调制激光器(EMLs)上尤为明显，但随着共封装光学(CPO)的推进，已扩展至连续波(CW)激光器。 我一直看好 $AAOI、$SIVE、$LITE 和 $COHR 带来的激光器卡脖子+瓶颈逻辑。很高兴这一论点开始得到验证。
英文原文
$LITE CEO Michael Hurlston at the RAISE Summit warned that the supply gap for InP lasers for AI DCs: Is facing a more severe supply chain crisis than memory. And with Lumentum's 5 InP fabs, shipments would be more than 30%+ below customer demand. This is especially visible with EMLs today but is already expanding to CW, especially as CPO ramps. I've always been a fan of the laser chokepoint + bottleneck from $AAOI, $SIVE, $LITE, and $COHR. And glad this thesis is starting to see validation.
-
博主感叹深度研究常被媒体误读为迷因股，故对他人遭遇有同理心。
@Jacktecau990 我并不惊讶。我可以去参加会议进行共封装光学(CPO)渠道调研，发现独特的超大规模云服务商供应链联系并随后得到证实，进行晶圆收入建模，构思通过并购实现收入增长的策略，通过相关员工关系(ER)电话会议记录/SEC文件持续检查出货量时间表，通过机构研报进行竞争对手分析，进行总可寻址市场(TAM)分析。然后免费发布这些论点。然而媒体却忽略所有底层工作，将这一想法标签化为单纯的“迷因股”。如果它出现回调，即使处于早期阶段，我也会在X上遭受大量人身攻击。所以我对他人有更多的同理心。
英文原文
@Jacktecau990 I'm not surprised. I could go to conferences to do CPO channel checks, find unique hyperscaler supply chain links that later get confirmed, do wafer revenue modeling, think of a strategy for revenue ramping through M&A, constantly check volume timelines through related ER transcripts/SEC filings, competitor analysis through institutional notes, TAM analysis. Then post that thesis all for free over time. Only for media media ignore all the underlying work then label the idea as just a meme stock. And if it ever corrects, I get a bunch of personal attacks on X, even if it's early in the timeline. So I have more empathy than others.
-
博主澄清CPO延迟谣言，认为亚洲股错杀，看好MSSCorp及早期受益股。
@Erick_P23 核心逻辑更偏向于共封装光学(CPO)的规模化，预计其复杂性将使MSSCorp成为检测良率方面的垄断型瓶颈（他们声称目标市场份额为90%）。尽管这些技术也可用于其他领域。 $FORM 的产能爬坡早了很多，所以下次财报中营收可能会开始显现？ 无论如何，亚洲概念股因CPO时间表延迟的报道而大幅抛售，但我通过$AMD 大会上的渠道核实发现该报道不实。 我认为我的观点没有改变，更多的是在等待2027年。 像Shunsin这样拥有更多确认的规模化交换机出货量，以及像 $AAOI / $SIVE 这样的激光概念股，其产能爬坡要早得多。在我看来，这只是时间线的问题，以及市场何时愿意对此进行定价。
英文原文
@Erick_P23 Thesis is more for CPO scale up expecting complexity making make MSSCorp a monopoly-type chokepoint over inspection yields (they were claiming 90% market share target). Although they can be used for other segments. $FORM ramped a lot earlier, so maybe will revenue would start to appear in next earnings? Regardless, Asian names sold off extremely hard on CPO timeline delay report that I found to be incorrect with my channel checks at the $AMD conference. Don't think my views have changed, it's moreso waiting for 2027. Names with more confirmed scale-out switch volumes like Shunsin + laser names like $AAOI / $SIVE ramp a lot earlier. It's just timelines in my view and when markets want to price that in.
-
博主分享近期因财报或市场波动对多只AI供应链个股的加仓操作及资金调配逻辑。
@RockyKh49376975 嗯，我个人在40多美元时加仓了$AMKR，因为我认为在$NVDA 15亿美元交易+$TSM 10年协议后，财报后跌这么多令人惊讶。不过它更多是2028年的标的，尽管2027年有独立的爬坡计划。- 在Foci跌至400时做了成本平均（$TSM COUPE共封装光学+ $NVDA的主要FAU供应商，其他财报如$FORM显示COUPE现在开始爬坡）。财报后也加仓了一些$XFAB，因为他们验证了CPO共封装光学+800V曝光，不过可能需要等待一段时间。在13 EU末日暴跌时加仓了$LPK，GCS不会消失，我认为由于Absolics延迟，出货量爬坡仅推迟了一个季度，我愿意等3个月。CoPoS、三星等应该在2027年推动更多需求。在OE Solutions跌至约14,000韩元时加仓，认为韩国领先的EML公司且拥有独立InP晶圆厂市值仅约1.1亿美元很奇怪。- Win Semi等名字对我非常有吸引力，但可惜没有足够资金在不使用高杠杆的情况下加仓。可能想逐步建仓$SIVE和$AAOI，因为那是我的两个最爱。鉴于Unitree IPO还有3周左右，正在关注$CCXI的机会，该上市应该像$SPCX的光环效应一样提振人形机器人板块。TLDR：只是资金调配，我个人的风险管理。标的与之前相同。
英文原文
@RockyKh49376975 Hmm, I personally added $AMKR in the $40's since I thought it was surprising it dropped that much after ER, even after $NVDA $1.5B deal + $TSM 10Y agreement. It's more of a 2028 name though, though they have separate ramps in 2027. - Did some cost averaging on names like Foci on drop to 400 (leading FAU supplier to $TSM COUPE + $NVDA, COUPE shows signs of ramp now from other ERs like $FORM). Added some $XFAB too after earnings since they verified CPO + 800V exposure, might be a bit of waiting though. Added $LPK on the 13 EU doomdrop, GCS isn't going anywhere, think volume ramp just got delayed a quarter because of Absolics delay and I'm fine waiting 3 months. CoPoS, Samsung, and others should drive demand more in 2027. And added OE Solutions on drop to ~14,000 KRW, thought it was strange Korea's leading EML company with independent InP fabs was only worth ~$110M MC. - Names like Win Semi + others are really really compelling to me but just dont have enough capital to add sadly without going on heavy margin. Probably want to scale into $SIVE and $AAOI more since those are my two favorites. And looking at $CCXI opportunities given Unitree IPO is around 3 weeks, that listing should boost humanoid/robotics sector a bit similar to the $SPCX halo effect. TLDR: Just moving capital around, my own personal risk management. Same names as before.
-
四大云厂商上调Capex指引，看好上游半导体及AI供应链瓶颈环节重估。
作为业绩复盘：$AMZN、$META、$GOOGL 和 $MSFT 将 2026 年联合资本支出(Capex)指引上调至约 7200-7450 亿美元。此前指引为 6950-7250 亿美元。亚马逊：2200 亿；谷歌：1950-2050 亿；Meta：1300-1450 亿；微软：1750 亿。我们已看到显著的降杠杆和零售/机构保证金清算（可能还会持续，谁知道呢）。但中长期来看，我不太明白为何有人看空上游半导体或新云厂商。鉴于每家超大规模云服务商都指出了算力短缺、云需求上升+定价权增强，或增加芯片/网络支出。我对许多上游半导体供应链的“瓶颈”论点是：当数万亿美元资本流入磷化铟(InP)衬底或存储（此前均被视为廉价大宗商品）甚至能源领域时。许多此前在电信周期中被视为无用甚至“卖马桶”的当前 AI 概念股。当拐点到来且资本支出流入其资产负债表时，估值将重估。正如过去几年 $NVDA GPU、今年存储、未来几年 CPU/MLCC、2027 年共封装光学(CPO)、2027 年玻璃衬底、2027 年 800V 等所见。
英文原文
As a earnings recap: $AMZN, $META, $GOOGL, and $MSFT guided a 2026 combined capex forecast to ~$720-$745 Billion. Up from $695-$725B Billion previously. Amazon: $220B Google: $195B-$205B Meta: $130B-$145B Microsoft: $175B We've already seen significant deleveraging and retail/institutional margin liquidations (maybe it continues for more time, who knows). But medium-long term, I'm not quite sure how anyone can be bearish the upstream semis or neoclouds. Given each hyperscaler has flagged either compute shortages, rising cloud demand + pricing power, or increased spending for chips/networking. My "bottleneck" thesis with many of these upstream semi supply chains is that when trillions in capital flows into things from InP substrates or memory (which were both treated as cheap commodities) or even energy. Lot of these current AI names that were treated as useless before in telecom cycles or even toilet sellers. Gets rerated when their inflection period hits and capex flows into their balance sheets. As seen with $NVDA GPUs past few years, memory this year, CPUs/MLCCs next few years, CPO in 2027, Glass Substrates in 2027, 800V in 2027, and so on.
-
POET与Sivers合作加速，2027H1订单落地，验证CPO等新技术拐点临近。
仔细想想，$POET 的光学中介层（optical interposer）业务与 $SIVE 的关联比我想象的更重要。起初措辞是2026年底的“准备就绪”，但 $POET 将其升级为与 Sivers 激光器合作的“量产”，预计2026年底实现。因此，从 $JBL 到 $POET 的批量订单应在2027年上半年落地。这很合理，因为 Sivers 已完成超认购的机构轮融资以支持未来大规模生产。顺便一提，难以置信已是八月，距2027年仅剩5个月……这是我预判玻璃基板、共封装光学(CPO)、800V 及新架构开始放量的拐点窗口。
英文原文
Now that I think of it more, $POET optical interposer piece is actually a little more material than I thought with $SIVE. Originally the wording was "readiness" for EOY 2026. But $POET upgraded that wording, by stating "production" with Sivers lasers projected EOY 2026. So a volume orders from across $JBL to $POET, should hit in H1 2027. Which makes a bit of sense, since Sivers did raise an institutional oversubscribed round for future mass production. On a side note, can't believe it's already August and 5 months left until 2027... Which was my inflection window for glass substrates, CPO, 800v, and new architectures to ramp.
-
看好$XFAB作为早期$TSEM潜力股，CPO预期未计入模型。
@jim29223471 是的，我的想法是 $XFAB 可能成为早期的 $TSEM，作为一家实际的碳化硅(SiC)/微机电系统(MEMS)晶圆代工厂，在他们发展光子学部门的过程中起到桥梁作用。他们重申了对共封装光学(CPO)的敞口，且2028年的时间表尚未被计入任何营收模型中（这一点我很欣赏）。可能还稍微有点早，但我们正接近2026年下半年的节点，且管理层指引(ER)非常有帮助。
英文原文
@jim29223471 Yeah my thought was $XFAB could be an early $TSEM with actual SiC/MEMs foundry kinda bridging the gap while they develop their photonics arm. They did reiterate exposure to CPO and 2028 timeline isn't priced into any revenue models yet (which I liked about them). Might be a tad early, but we're getting closer in H2 2026 and the ER statements were very helpful.
-
XFAB确认CPO与800V布局，验证投资逻辑，预期2027-2028年估值重估。
关于 $XFAB 投资者关系(ER)电话会议纪要： 简单回顾一下，我对 $XFAB 晶圆代工增长机会的观点源于碳化硅(SiC)在800V架构中的应用，以及光子工厂(MTP)在共封装光学(CPO)领域的适用性。 - $XFAB：“我们在共封装光学方面正在取得进展”。 - “向800V架构的过渡……预计将推动对碳化硅(SiC)和氮化镓(GaN)的需求……对 $XFAB 而言”。 所以，那些自信地唱空 $XFAB、声称其“在800V、光子学、CPO方面零敞口”的分析师/人士，结果被证明完全错了。 在我分享观点时，行业材料显示日月光(ASE)等公司将光子工厂+ $XFAB 列为欧洲即将崛起的CPO玩家。 filings显示他们普遍涉及如 $NVTS / $POWI 等名称。 很高兴 $XFAB 出面明确澄清，表示他们正在开发CPO/800V相关技术。 回到ER纪要： - “我们继续预计光子学量产将于2028年开始”。公司目前有三个硅光子项目正在与关键合作伙伴开发中。 我原本以为会在2027年下半年爬坡，但可能早估了几个月。 - 多个新的数据中心设计订单，例如3个新的碳化硅设计订单（与800V部署启动后2027年普遍爬坡一致）。 因此，一旦这些在2027-2028年爬坡（市场往往会提前定价），我预计 $XFAB 将基于增长机会获得估值重估。 总结： - 当前营收数据并不惊艳。 - $XFAB 确认直接涉足CPO和800V转型（这是投资逻辑的最大胜利） - 800V敞口带来的碳化硅机会（应在2027年） - 光子学量产时间表为2028年 很高兴CPO/800V敞口的投资逻辑在此次ER中得到确认。营收爬坡可能需要一点等待时间。
英文原文
For $XFAB ER transcript: Just to recap, my idea on $XFAB foundry growth opportunities stemmed from SiC for 800v and photonixfab/MTP for CPO applicability. - $XFAB: "We are making progress in co-packaged optics". - "The transition to 800v... is expected to support demand for both SiC and GaN... for XFAB" So those analysts/people that were confidently bearposting about $XFAB that they had "0 exposure to 800V, photonics, CPO" just turned out plainly wrong. At the time I shared my idea, industry materials showed ASE and others cited photonixfab + XFAB as an upcoming CPO player in EU. And filings showed they had exposure names like $NVTS / $POWI in general. So glad $XFAB went out to explicitly clear things up to state they were developing in CPO/800v. Okay back to the ER: - "We continue to expect photonics volume production to start in 2028". Company currently has three silicon photonics projects in development with key partners. I thought it would ramp sometime in H2 2027, but was probably early by a few months. - Several new data center design wins, eg. 3 new SiC design wins (consistent with general 2027 ramp as 800v deployment takes off). So once those ramp in 2027-2028, (markets tend to price things ahead of time), I'd expect $XFAB to be rerated based on growth opportunities. TLDR: - Current revenue numbers not as impressive. - $XFAB confirmed direct exposure to CPO and 800V shifts (biggest win for thesis) - 800V exposure opportunities for SiC (should be 2027) - Timeline is 2028 ramp for photonics volume production Happy exposure thesis to CPO/800V was confirmed from this ER. Might be a bit of waiting for revenue ramp.
-
亚马逊资本支出上调及算力短缺证实，AXTI 产能扩张且供不应求，看多 AI 半导体及光通信供应链。
几条财报摘要，重点是我看好的 $AXTI 和 $AMZN： 亚马逊： - 将 2026 年资本支出上调至 2200 亿美元（此前为 2000 亿美元），部分原因是内存成本上升，这对 $MU 和三星是利好。 - 即使达到 2200 亿美元，亚马逊的产能仍无法满足 2026 年的所有需求；Jassy 预计 2027 年情况相同。 - 大多数 2027 年的新增产能已被预订，2028 年的大量产能也已预订。 主要浏览了上游半导体板块。$GOOGL、$META、$AMZN 和 $MSFT 均指出算力短缺。 几周前关于 Meta 等公司“算力过剩”以及超大规模云厂商削减支出的所有叙事……全是胡扯。 亚马逊财报对 AI 半导体交易非常看多。 AXTI： - AXT 计划在 2026 年将磷化铟(InP)产能翻倍，2027 年再翻倍。 - 预计这将使 AXT 成为“迄今为止全球最大的磷化铟生产商”——冲鸭。 - 第二季度营收为 4760 万美元，创 AXT 历史单季新高（来自数据中心应用的 InP 营收创纪录达 3070 万美元）。 - 营收环比增长 77%，同比增长 164%。 - InP 营收-产能目标：2026 年底每季度 6000 万美元。 - 2027 年底每季度 1.3 亿美元。 由于管理层表示他们将“寻找任何方式增加产能扩张”，1.3 亿美元的目标可能会进一步提高。 管理层称报告中的积压订单仍远高于 1 亿美元，但该数字已不能反映所有可用需求。 “无论我们增加产能的速度多快，客户需求持续超过供应。” 800G/1.6T 正在驱动当前周期，而 NPO/CPO 将其延伸至 2027 年之后。 此外，他们的目标是 50% 以上的毛利率：“我们绝对应该瞄准一个以五开头的数字。” 中国需求翻倍有余，且与 Casela、$COHR 和 $LITE 的协议并未实质性推动第二季度业绩。 这甚至还没有考虑我预测的随着 InP 衬底瓶颈加剧带来的巨大平均售价(ASP)上涨。 摘要： - 亚马逊算力需求过大，需要资本支出来满足，因此上游半导体起飞。 - AXT 是全球最大的 InP 衬底供应商，高毛利率，产能扩张，且供应跟不上需求。 看多来自亚马逊的需求端+资本支出。看多上游光通信供应链，因为需求过大。
英文原文
Few earnings TLDRs with my favorite $AXTI and $AMZN: Amazon: - Raised 2026 capex to $220B vs. prior $200B (partly due to higher memory costs, which is bullish on $MU to Samsung) - Even at $220B, Amazon will not have enough capacity to meet all 2026 demand; Jassy expects the same in 2027. - Most incoming 2027 capacity is already reserved, with substantial 2028 capacity also reserved. Mostly read through on upstream semis. $GOOGL, $META, $AMZN, and $MSFT all identified compute shortage. All the narratives a few weeks earlier was "excess compute" from Meta and others + hyperscalers cutting back on spend... All BS. Amazon earnings was very bullish on AI semi trade. AXTI: - AXT to double InP capacity during 2026. Then double again in 2027. This is expected to make AXT "by far the largest indium phosphide producer in the world." - LFG - Q2 rev was $47.6M, the highest quarterly revenue in AXT history (InP revenue reached a record $30.7M, from DC applications.) - Revenue increased 77% sequentially and 164% year over year. - InP revenue-capacity targets: $60M per quarter exiting 2026. - $130M per quarter exiting 2027 That $130m target could be hiked too since management stated they find "whatever ways to increase that capacity expansion" - Management said the reported backlog remains well above $100M, but that number no longer reflects all available demand. "Customer demand continues to outpace supply, no matter how fast we add capacity." 800G/1.6T is driving the current cycle, while NPO/CPO extends it beyond 2027. Also they're targeting 50%-plus gross margin: "We should definitely be targeting a number that begins with a five." China demand more than doubled, and their agreements with Casela, $COHR and $LITE did not materially drive Q2. This is not even considering my projected massive ASP hikes yet as InP substrates get more bottlenecked. TLDR: - Amazon too much compute demands, needs capex to fufill it, so upstream semis go brrr. - AXT world largest InP substrate supplier, high gross margins, expansion, and supply can't keep up with demand. Bullish on demand side from Amazon + capex. Bullish on upstream optical supply chains from too much demand.
-
FORM财报显示CPO业务加速，处于早期爆发前夜，看好硅光子产业链。
以下是 $FORM 财报电话会议中关于共封装光学(CPO)的一些有趣要点： 由于“共封装光学(CPO)的加速增长”，系统收入几乎翻倍。 2026年Q1：“我们现在预计2026年CPO收入将达到1000万至2000万美元区间的高位。” 2026年Q2（当前）：“预计第三季度末将超过该区间，全年整体将显著超过2000万美元水平。” 管理层表示：“我们看到这项[CPO]业务正在显著加速。” 问：CPO采用前景。我们是否预期时间表会略微提前？ 答：“我想说的是短期内确实有加速。”但管理层表示时间表没有显著提前。 引用生产基础设施（不仅是研发）用于： - 今年晚些计划量产的CPO芯片日益增长的出货量 + 用于扩展和扩展交换机的测试插入。 - 似乎CPO部分通过 $TSM 的 COUPE 流程 [可能对应 $NVDA 的CPO产品] - “CPO部分确实通过那10%的客户流向。” 因此，鉴于管理层提到CPO芯片出货量增长，测试环节正在爬坡。 “我们CPO业务的近期快速增长令人兴奋，我们认为这代表了硅光子技术在更广泛半导体行业中广泛采用的非常早期阶段。” 简而言之：从根本上说，CPO作为一个主题仍处于非常早期阶段，但由于它现在开始出现在财报中，应该很快就会爆发（设备/测试玩家通常会在光引擎/激光器出货量爬坡之前先出现）。 个人而言，一旦所有去杠杆过程停止，我认为市场将开始更加关注这一点。
英文原文
Just some interesting takeaways from $FORM earnings call on CPO: Since they attributed systems revenue nearly doubled due to: "accelerating growth in co-packaged optics or CPO". Q1 2026: "we now expect 2026 CPO revenues to come in at the high end of the $10 million to $20 million range" Q2 2026 (now): "expect to exceed that range by the end of the third quarter, and to significantly exceed the $20 million level for the year overall." Management stated "We’re seeing some significant acceleration in this [CPO] business". Q: CPO adoption outlook. Are we perhaps expecting the timeline to accelerate a little bit? A: "What I would say is there’s acceleration here in the very short term." But management stated there's not a significant pull-in on timelines. Citing production infrastructure (not just R&D) for: - The growing volumes of CPO chips planned for later this year + test insertion for scale-up and scale-out switches. - Seems like that CPO piece flows through $TSM COUPE [ likely maps to $NVDA CPO products] - "The CPO piece does flow through that 10% customer" So now, test is ramping given management cited growing volumes of CPO chips. "The rapid recent growth of our CPO business is an exciting development, which we believe represents the very early stages of widespread adoption of silicon photonics in the broader semiconductor industry" TLDR: Fundamentally, CPO as a theme is very early, and should start to go brrr soon since it's starting to show up in earnings now (equipment/test players usually appear up first before optical engines / laser volumes ramp). Personally once all the deleveraging stops, I think markets will start to care more.
-
美政府入股GFS推CPO，利好西方供应链玩家。
@Yeshuaisking11 对于关键矿产，我原本就预料到西方需要尽可能多的补贴来与中国竞争。 美国政府入股 $GFS 以支持共封装光学(CPO)并不在我的预期之内。但这确实表明 CPO 已成为算力供应链的优先事项，且他们已将格芯(Globalfoundries)打造为本土冠军。 我认为这对所有参与美国 CPO 供应链的西方玩家如 $LITE / $SIVE 都是利好。但对世界其他地区的相关企业则不然。
英文原文
@Yeshuaisking11 I kinda expected it for critical minerals since Western needs as much subsidies as possible to compete with China. US Gov taking a stake in $GFS for CPO was not on my calendar. But it does go to show CPO becoming a compute supply chain priority and they made Globalfoundries a domestic champion. Which my thoughts are that it's positive for all the Western players this US CPO supply chains like $LITE / $SIVE. Not so much for other parts of the world.
-
GFS获政府入股支持CPO却股价下跌，对比INTC令人惊讶。
@4intheflames 我并未持有 $GFS，但令我个人惊讶的是，在美国政府为支持其推进共封装光学(CPO)技术而直接入股的消息公布后，其股价仍在下跌。考虑到大家都记得 $INTC 是最近一家受美国政府青睐的晶圆代工厂，这确实令人惊讶。
英文原文
@4intheflames I don't own $GFS, but I'm personally it's surprised it's still down on the news US Gov is taking a direct stake in them for advancing CPO. Surprising to say the least, given everyone remembers $INTC as the latest foundry the US Gov liked.
-
美政府注资GFS加速CPO研发，利好硅光子供应链标的SIVE和LITE。
美国政府将持有 $GFS 1% 的股份，并授予其 3 亿美元用于美国《芯片法案》(CHIPS ACT)。 鉴于该《芯片法案》专门旨在推进共封装光学(CPO) + 硅光子技术(Silicon Photonics)，这实际上是对 $SIVE / $LITE 的强烈利好映射(read through)。 （参考：Sivers 和 Lumentum 是 GFS 演示文稿中提到的唯二两家公开激光供应商。Sivers 激光阵列最近被指定为 Globalfoundries SCALE 用于 CPO 的参考设计。） 根据美国政府公告：“GlobalFoundries 将获得高达 3 亿美元，以将国内 CPO 研发加速两到三年”（NIST） 从未想过我们会看到美国政府 / $INTC 晶圆代工模式在特定 CPO 领域的应用...
英文原文
US Gov to take 1% stake in $GFS, and award them $300m for the US CHIPS ACT. This is actually a strong read through on $SIVE / $LITE, given this CHIPS ACT is specifically aimed at advancing CPO + Silicon Photonics. (For reference, Sivers and Lumentum were the only two public laser suppliers named in GFS presentation slides. Sivers laser arrays was named recently as a reference design in Globalfoundries SCALE for CPO). Per US Gov announcement: "GlobalFoundries will receive up to $300 million to accelerate domestic CPO R&D by two to three years" (NIST) Never thought we'd see the US Gov / $INTC foundry playbook for CPO in specific...
-
博主暂不建仓，静待CPO标的逻辑兑现，后续将分享个股思路。
@twisterchessur 就我个人而言，目前只想静待既有的做多逻辑逐步兑现，尤其是近期大盘回调之后的共封装光学(CPO)相关标的。当前暂不打算建新仓，不过稍后我会再分享几个个股思路。
英文原文
@twisterchessur 就我个人而言 目前只想静待既有的做多逻辑逐步兑现，尤其是近期大盘回调之后的 CPO 相关标的。当前暂不打算建新仓 不过稍后我会再分享几个个股思路