全部推文
共 6775 条 · 第 171 / 226 页 · 滚到底自动加载更多
-
AXTI通过掌控InP供应链实现定价权，涨价对TPU成本影响有限。
^ 澄清一下，$AXTI 掌控定价权。关于数量，指的是掌控磷化铟(InP)供应链的大部分份额。 他们拥有一定量的库存，这显然受中国许可证管制。 重点是，即使产能翻倍或翻三倍，如果你同时掌控供应链，就能掌控定价。你可以用有限的材料将价格提高100%+、500%、2000%，超大规模云服务商(hyperscalers)仍会买单，因为没有替代品。 即便涨幅如此之大，由于此前价格占比极低，对TPU集群物料清单(BOM)的影响也仅增加几个百分点。
英文原文
^ clarification, $AXTI controls the price. Regarding quantity, that was referring to controlling majority of the InP supply chain. They have a certain amount of inventory, which obviously is controlled by China for permits. Point was that can double, triple, capacity, but if you control both the supply chain, you can control pricing. You can have finite material and increase prices 100%+, 500%, 2000%, and hyperscalers would still pay because there's no alternative. Even that much increase would only be a few percent increase on TPU cluster BOM because the price was so negligible before.
-
AXTI因垄断磷化铟供应链，在AI基建中具极高价值，远超当前市值。
对于Q X P，$AXTI 在中国对日本实施出口管制后，在光子学(Photonics)领域的AI基础设施建设中，既控制了数量（受监管的垄断）也控制了价格（规模垄断）。 双头垄断(Duopoly)的消亡彻底改变了原有的定价结构。 磷化铟(InP)的潜在市场规模(TAM)仅为数亿美元，因为它是一种小众的电信产品。现在，谷歌(Google)愿意支付数十亿美元，仅为了确保其整个TPU项目不会停滞。Meta、微软(Microsoft)、亚马逊(Amazon)、英伟达(Nvidia)等也是如此。于是，TAM突然从数亿美元飙升至100亿美元以上。 再说一次，我不知道这会走向何方，因为不可能对从未发生过的极端情景（一家小盘公司控制着万亿美元AI基建的两个关键瓶颈）进行建模。 但当一家公司控制了超大规模云服务商(Hyperscaler)扩展其项目所用材料的大部分供应链以及价格结构时，我认为其价值远高于13亿美元的市场资本化。
英文原文
For Q X P, $AXTI controls both quantity (regulated monopoly) as well as price (monopoly at scale) for the AI buildout with photonics after China's export control on Japan. The death of a duopoly completely changed any former pricing structure there was. The TAM of InP was few hundred million because it's a niche telecom product. Now, Google would happily pay billions just so their entire TPU program doesn't stall out. Same goes for Meta, Microsoft, Amazon, Nvidia, and others. Then TAM suddenly goes from few hundred million to $10B+. Again I don't know where this heads, it's impossible to model an extreme scenario that's never happened before (one small cap company controlling two different critical bottlenecks for the trillion AI buildout). But when the company controls majority of the supply chain as well as price structure of a material used for every hyperscaler to scale out their programs, I place a lot more value on it than $1.3B market cap.
-
HBM 瓶颈导致传统建模失效，磷化铟成新瓶颈，$AXTI 因卡位 AI 建设具极高价值。
当出现像 $MU 或 SK 海力士的 HBM（高带宽内存）这样极不常规的瓶颈时，诸如 NTM（下一财年）这样的标准建模就会失效。目前 HBM 的需求极度缺乏弹性，买家不再询问价格，即使在价格飙升数百%后，他们仍在寻求配额分配。在这种情况下，基于 $NVDA 2025 年上半年 EML（边缘机器学习）的现有短缺，以及 $MSFT 等公司仅其项目就需要全球两位数的磷化铟(InP)产量（针对 2026 年下半年至 2027 年），$AXTI 很可能在接下来几个月成为下一个瓶颈。因此，当 $GOOGL 等超大规模云服务商为避免 TPU 项目停滞而迫切寻求配额时，原本数亿 TAM（总可寻址市场）可能迅速膨胀至数十亿甚至上百亿，这使得对瓶颈的建模变得不可能。这本质上是买入“比特分配”、原料控制和瓶颈环节。一家市值 13 亿美元的公司卡住了整个 AI 建设的脖子，在我看来很便宜。我不知道这会走向何方，但我预计会出现极大的价格挤压（我们在 SMM 上已看到迹象），这将增加 $AXTI 的底线利润。许可证波动一直是一个风险因素。但同样，任何延迟都会因为这是单点故障而卡住整个 AI 建设，前提是如果中国继续对日本实施出口管制。
英文原文
Standard modeling like NTM breaks when there's extremely unconventional bottlenecks like HBM with $MU or Sk Hynix. HBM demand is so inelastic rn that buyers aren't asking for price, they're asking for allocation even after hundreds of percent increases. And in this case InP and $AXTI will likely be the next bottleneck in the upcoming months, just based on existing shortages from $NVDA EML H1 2025 and how others like $MSFT require the world's double digit inp output just for their program for H2 2026 into 2027. So it's impossible to model bottlenecks where few hundred million TAM previously might squeeze to few billion or tens of billions when hyperscalers like $GOOGL are desperate for allocation so their TPU program doesn't stall out. This is buying into bit-allocation, feedstock control, and bottlenecks. A $1.3B company bottlenecking the entire AI buildout looks cheap to me. I don't know where this is heading, but I do expect an extremely large price squeeze (we're seeing it now on SMM), which just increases the bottom line of $AXTI. License volatility was always a risk factor. But again, any delays would just bottleneck the entire AI buildout since this is single point of failure if China continues their export controls on Japan.
-
博主调侃自己在 X 上与名人互动
@jukan05 现在我可以吹牛说我在 X 上和名人聊过天了
英文原文
@jukan05 now i can brag I've talked to someone famous on X
-
质疑AXTI估值逻辑，指出AH股价波动源于低流动性。
如果市场试图将 $AXTI 2024年签约积压订单的收入确认时机计入当前结构性需求，我仍觉得有些好笑。但 $AH 的许多波动仅仅是因为流动性低（例如 $2 的买卖价差，仅100股的卖出就会导致价格下跌10%），让我们看看接下来几周会发生什么。
英文原文
I'm still chuckling if the market wants to price in $AXTI revenue recognition timing from 2024 contracted backlog over current structural demand. But a lot of the AH movement is just low liquidity (eg. $2 spread, one 100 share sale would drop the price 10%), but we'll see what happens in the upcoming weeks.
-
AXTI因InP衬底垄断地位及全产业链优势受益，股价上涨逻辑多维。
$AXTI 从 Northland 获得 1 亿美元用于资本支出扩张，推迟的自由现金流（FCF）源自 2024 年或 2025 年初的 500 万美元合同收入，这并无实质性影响。 股价上涨是多方面的： -> 认识到磷化铟(InP)衬底将成为光子学领域 AI 行业部署的瓶颈 -> 认识到 InP 衬底生产是双寡头垄断（住友、$AXTI），还有 JX 等较小玩家 -> 认识到 InP 原料是 InP 衬底生产商（$AXTI、Vital）的另一个瓶颈，还有 $DOWA 等较小玩家 -> 来自 $COHR 等财报的供应短缺通知 -> 认识到西方超大规模云服务商在 12 月 24 日左右在伦敦金属交易所(SMM)价格之上支付溢价 -> 认识到两天前中国对日本实施出口管制后，AXT 实际上成为垄断者（这将限制其主要竞争对手的大规模衬底生产）。估计有 3-6 个月的库存。 7N 级产品价格已达历史新高，$AXTI 拥有从矿山到衬底生产的全产业链，因此西方超大规模云服务商与 SMM 之间的多次涨价和溢价分化将增加其底线利润。 我个人认为，在整个供应瓶颈出现之前，合同积压订单和推迟一个季度的合同收入（追溯至 2024 年或 2025 年初）并不构成实质性影响。但解读因人而异。
英文原文
$AXTI secured $100m from Northland for capex expansion, deferred FCF from $5m of contracted revenue from 2024 or early 2025 does not make a material difference. Stock runup was multifaceted: -> Recognition that InP substrates would bottleneck the AI industry deployment with photonics -> Recognition that InP substrate production was a duopoly (Sumitomo, $AXTI), with some smaller players like JX ->Recognition that InP feedstock was another bottleneck for the InP substrate producers ( $AXTI, Vital) with some smaller players like $DOWA. -> Supply shortage notice from earnings reports like $COHR -> Recognition that Western hypescalers are paying markups on top of SMM prices ~Dec 24th -> Recognition that AXT effectively became a monopoly after China's export controls on Japan 2 days ago (which would limit substrate production at scale from AXT's major competitor). Est 3-6 months inventory. 7n prices have hit ATHs, $AXTI owns the mines all the way to substrate production so many price increases and bifurcation in premiums between western hyperscalers + SMM would increase their bottom line. I personally don't consider contracted backlog and 1 quarter deferred contracted revenue all the way back from 2024 or early 2025 to be material before the entire supply chokepoint started. But to each their own interpretation.
-
欢迎RJCcapital加入讨论
@RJCcapital 欢迎加入派对
英文原文
@RJCcapital Welcome to the party
-
当然，我不是来试图说服你们的，我只是发布我自己关于 $AXTI 的投资论点。 我要指出的细微差别是，该公司受益于 AI 基础设施建设，此前该领域遭遇了铜互连瓶颈，现在正转向光子学以及新型互连和 ASIC 架构。 直到 2025 年后期，超大规模云服务商才真正开始确保外调制激光器(EML)的产能。 磷化铟(InP)衬底是光子学 AI 基础设施建设所需的关键材料，而此前的铜代或旧架构并不需要。
英文原文
Sure, I’m not here to convince you I’m just posting my own thesis about $AXTI. The nuance I’ll point out is that this is levered to the AI buildout that hit the copper wall and is shifting to photonics with newer interconnects and ASIC architectures. Only later on in 2025 did hyperscalers really start securing EML capacity. InP substrates is a critical material required for the AI buildout with photonics, not previous copper generations or former architectures.
-
指出某标的流动性低且买卖价差大。
@straightcash385 流动性很低，买卖价差在 19 美元和 21 美元之间，哈哈 https://t.co/W1cQJYo6Hn
英文原文
@straightcash385 It’s low liquidity, the spread is $19 and $21 lol https://t.co/W1cQJYo6Hn
-
AXTI股价波动属旧账延期，核心应关注衬底瓶颈与InP涨价。
$AXTI 简而言之：人们正对一年前签约收入因延期一个季度而过度反应。此处并未报告前瞻性收入指引。相比于2024年的签约收入，我更关注实时的衬底(Substrate)瓶颈以及磷化铟(InP)价格的上涨。
英文原文
$AXTI TLDR: people are overreacting to contracted revenue from over a year ago being deferred by a quarter. There was no forward revenue guidance reported here. I care more about the real-time substrate bottleneck and increasing InP prices than 2024 contracted revenue.
-
切断关键材料将引发中美全面贸易战，双方受损。
@valuedrift @4drant 英伟达(Nvidia) GPU与美国贸易。如果他们切断关键材料流向美国，将引发一场两败俱伤的全面贸易战。
英文原文
@valuedrift @4drant Nvidia GPUs and American trade. There would be a full blown trade war that hurts both sides if they stop the flow of critical materials to US
-
澄清无前瞻指引，仅为旧合同及收入递延反应
@4drant 没有前瞻性指引。这纯粹是对一年前签订的合同以及一个季度收入递延的反应。
英文原文
@4drant There’s no forward guidance. This is solely a reaction to the contracts made a year ago with one quarter revenue deferral.
-
预计磷化铟短缺带来博弈论配置机会，算法或先猎杀止损。
@dl123abc @william_R2Rclub 嗯，如果我是算法，我会进行止损猎杀(stop loss hunting)——首先触发任何散户的短期恐慌性抛售。 我预计目前会出现严重的磷化铟(InP)短缺，并在未来2-3个季度内出现博弈论类型的资金配置，因此这提供了一个相当不错的机会。
英文原文
@dl123abc @william_R2Rclub Hmm, if I were an algorithm, I’d go stop loss hunting -> trigger any retail short term panic sellers first. Im expecting a major InP shortage around now and game theory type allocations over the next 2-3 quarters, so this presents a pretty good opportunity.
-
AXTI基板瓶颈缓解，SMM价格改善，定价权增强。
@itsdsha 自发布关于当前基板(substrate)瓶颈的帖子以来，已经过去了整整13天。我观察到 $AXTI 的硅微机械(SMM)价格有所不同，且未来的定价能力增强，但人们可以自由卖出，如果他们对2024年合同产生的递延收入做出反应的话。 https://t.co/Sxuw0B3rez
英文原文
@itsdsha It’s been a solid 13 days since posting about the substrate bottleneck happening now. I’m seeing something different with SMM prices and increased future pricing power with $AXTI, but people are free to sell if they want to react to deferred revenue from contracts made in 2024. https://t.co/Sxuw0B3rez
-
AXTI因短期合同被错杀，当前短缺期远期估算更关键。
@william_R2Rclub 是的，我笑得很开心，$AXTI 因为 2024-2025 年初的合同加上第一季度递延被抛售，而实际上当前的短缺/价格挤压期间，远期估算才是最重要的。此外，小盘股盘后流动性不足也会造成很大差异。
英文原文
@william_R2Rclub Yeah I was laughing a lot that $AXTI got sold off on 2024-early 2025 contracts + 1Q deferral when forward estimates matter the most during the actual shortage/price squeeze happening now. After hour lack of liquidity on a small cap makes quite a difference too
-
指出市场忽视InP瓶颈，基于旧合同抛售存在机会。
@TraderSmooth 市场对磷化铟(InP)当前及未来瓶颈缺乏理解，这提供了一个机会，如果人们是基于2024年或2025年前三季度的递延收入合同而抛售的话。
英文原文
@TraderSmooth The lack of understanding current/future bottlenecks around InP presents an opportunity if people are selling based on deferred revenue from 2024 or q1-3 2025 contracts.
-
材料出口管制致AXTI垄断，下游产能受阻，错杀带来机会
他们确实可以，这一瓶颈向下游传导，损害了 $COHR 到 $LITE 向 $GOOGL 和 $MSFT 的产能爬坡。特别是鉴于住友/JX 面临材料出口管制导致的短缺，$AXTI 将在几个季度后处于垄断地位。人们非理性地对 2024 年合同的前置收益和递延收入确认做出反应，而忽视了当前状况，在我看来这提供了一个绝佳的机会。
英文原文
They can, and this bottleneck flows downstream and hurts $COHR to $LITE capacity ramp into $GOOGL to $MSFT. Especially given $AXTI will be monopoly status in a few quarter once Sumitomo/JX face shortages from material export controls. People illogically reacting to former earnings from 2024 contracts and deferred revenue recognition, while ignoring the situation now presents a great opportunity imo
-
市场误读旧合同数据，忽视当前AI供应链的真实瓶颈。
情况没有变化。这是对2024年及去年第一至第三季度签订的老合同，以及针对一个季度收益递延(Deferred Revenue)的滑稽反应。真正的供应短缺和价格飙升(Squeeze)正在此刻发生。市场通常会提前行动，看着大家还在担心供应短缺全面爆发前的旧季度数据，真是有趣。
英文原文
状況に変化はありません。これは2024年や昨年の第1〜第3四半期に締結された古い契約と、1四半期分の収益繰延べ（デファラル）に対する滑稽な反応です。 主要な供給不足や価格高騰（スクイーズ）が起きているのは、まさに今です。市場は通常、先を見越して動くものですから、供給不足が本格化する前の過去の四半期を気にしている様子は、見ていて面白いですね。
-
指出市场忽视InP短缺挤压，仍纠结于旧合同收入确认。
@Paulyy_K 这个决定由你自行判断。我只是觉得有点好笑，因为目前磷化铟(InP)的价格和短缺情况正面临巨大的挤压。然而，人们却在反应2024年或去年第一季度至第三季度（其中第一季度确认递延）的收入确认和合同。 https://t.co/5ScwrUYBDG
英文原文
@Paulyy_K That decision is up to you. All I’m saying is I’m laughing a bit because what we’re seeing with InP prices and shortages is a huge squeeze. Yet people are reacting to former revenue recognition and contracts made in 2024 or q1-q3 last year (with 1q recognition deferred) https://t.co/5ScwrUYBDG
-
AXTI盘后大跌，博主认为不影响基于需求瓶颈的核心投资逻辑。
$AXTI 盘后大跌 29.54%，原因如下： - 下调 2025 年第四季度营收预期至 2300 万美元中值，此前为 2700 万-3000 万美元。 我的看法：这对瓶颈论点无关紧要，因为价格冲击刚刚发生在 12 月 27 日左右至 1 月（现在）。这是过去的财报，我们关注的是拥有定价权从而带来高毛利的未来盈利预测。 - AXTI 收到出口管制延迟通知，预计将在 2026 年第一季度确认收入。 这只是确认了上一季度的积压订单。这不是收入损失。出口管制带来的延迟和风险是已知的。 简而言之：过去的财报对论点无关紧要，这些只是已知的延迟，收入将在下一季度确认。 更应关注从 12 月 27 日左右开始的需求挤压拐点，而非磷化铟(InP)热潮之前的积压营收。
英文原文
$AXTI drops 29.54% AHs on: - Lowered Q4 2025 revenue expectations to a range of $23M midpoint, from $27m-$30m. Thoughts: this is immaterial to the bottleneck issue since price shock just came in end of ~December 27th into January (now). This was former earnings, we care about future earning projections with inflated margins from pricing power. - AXTI received export control delays and expect to recognize the revenue Q1 2026. This just realizes the former backlog Q1. This is not lost revenue. Export controls have known delays and risks. TLDR: former earnings are immaterial to thesis and these are just known delays that recognizes the same revenue next quarter. Care more about what happens at the inflection point of demand squeeze starting ~Dec 27th onwards, not former backlogged earnings before the InP craze.
-
POET 非当前 InP 瓶颈解法，或是 2028 年降 InP 用量的工程方案。
$POET 对于 $AXTI 在 2026-2027 年因磷化铟(InP)瓶颈以及当前专用集成电路(ASIC)/图形处理器(GPU)架构而言，基本不存在。但它是 2028 年以后降低 InP 用量的一种工程变通方案（尽管仍需使用 InP）。例如，富士康 H2 订单良率低、缺乏超大规模数据中心供应商认证，且未纳入 TPU Ironwood 等当前架构设计，目前一切均依赖 InP。但 2028 年推出的未来设计可能会采用它。
英文原文
$POET is basically nonexistent for the $AXTI 1-2 year InP bottleneck for 2026-2027 and current asic/GPU architectures but it’s one of the engineering workarounds to lower InP usage 2028 onward (still requires InP though) Eg. Foxconn h2 orders, low yield, lack of hyperscaler qualification, and not in current architecture designs like TPU ironwood, and everything is currently dependent on InP. But future designs might use it going out in 2028
-
中国新出口管制影响日企前体供应，需二次审视。
@Citrini7 研究信息的“圣杯”，感谢免费分享。此外，由于味之素(Ajinomoto)、日东纺(Nittobo)和三菱瓦斯化学(Mitsubishi Gas Chemical)等公司使用中国原材料前体，其中一些受中国前几天实施的新双重出口管制影响，因此需要进行二次审视。
英文原文
@Citrini7 Holy grail of research info, thanks for posting it free. Also some of these are affected by China’s new dual export controls the other day since Ajinomoto, Nittobo, and Mitsubishi Gas Chemical and others use Chinese raw precursors, so it does require a secondary look
-
机构正涌入$AXTI，看好其InP垄断地位及产能爬坡带来的涨价预期。
这很可能全是机构资金在动，许多对冲基金正在意识到 $AXTI 在磷化铟(InP)领域的垄断地位，并试图在北陆(Northland)获得1亿美元注资后入场交易。 这也是一个完全新颖的发展，正在实时上演，99%以上的人甚至还不知道这个瓶颈/公司（我的X粉丝群是个信息茧泡）。 真正的乐趣将在几个月后开始，届时InP价格将飙升得更高，因为产能爬坡才刚刚开始1个半星期。
英文原文
This is likely all institutional movement, and a lot of hedge funds realizing $AXTI's InP monopoly position + trying to enter the trade following Northland $100m funding. This is completely novel development too playing out real time, and 99%+ of people aren't even aware of this bottleneck/company yet (my follower group on X is a bubble) The real fun starts in a few months when InP prices shoot up even higher, since it only stated to ramp 1 1/2 weeks ago.
-
出口禁令导致日本基板材料溢价飙升，西方市场定价更高。
这是本地定价。西方市场的定价会有很大不同（会加上溢价），而且现在溢价可能会更加极端，例如由于出口禁令，甚至无法以3000美元的价格为日本基板生产采购材料。此外，在去年12月底禁令实施前，价格就开始飙升了。
英文原文
this is local pricing. Western pricing would be a lot different (premiums applied), and now there'd probably be even more extreme premiums eg. can't even secure materials for $3k in for japanese substrate prod because of the export ban. it also started spiking late Dec last year before the ban.
-
确认拐点已至，好奇后续走势是否线性上涨。
@zephyr_z9 是的，看起来终于到达了拐点。好奇接下来几个月是否会呈直线上涨。
英文原文
@zephyr_z9 Yeah looks like it's finally hit the inflection point. Curious if it's just a straight line up for the next few months.
-
超大规模云服务商竞价推高利基电信产品供应价格
~12月27日只是价格指数级增长的起点，我预计这种上涨趋势将在今年持续。特别是当超大规模云服务商(hyperscalers)试图相互竞价以争夺即时供应时，而该产品的成本本身非常低廉（最初是利基电信产品）。另外，第一张照片是一个基准测试，这张图表更具代表性。
英文原文
~Dec 27th was just just the start of the exponential increase in price, and I expect this to continue increase over the course of this year. Especially as hyperscalers try and outbid each other for immediate supply when the cost itself is so cheap (originally niche telecom product). Also first photo was a benchmark, this graph is more representative.
-
磷化铟供应瓶颈加剧，AXTI成AI基建关键节点。
你正在实时目睹磷化铟(InP)“瓶颈”的演变，而超大规模云服务商正为争夺材料而忙乱。用于光子学的7N级铟(7-N Indium)在SMM上呈抛物线式暴涨，2026年屡创历史新高。这是由$AXTI控制的材料引发的实时“博弈论”场景。
英文原文
You're watching the InP "bottleneck" play out real time while hyperscalers scramble for material. 7-N Indium (required for photonics) is now parabolically exploding on SMM, reaching daily ATHs in 2026. This is a live "Game Theory" scenario for materials controlled by $AXTI. https://t.co/3YDMVfjGTx
-
看好$AVAV，因获大额国防合同及无人机实战成功，小市值具上行潜力。
虽然算不上10倍股，但我真的很看好 $AVAV。去年12月他们赢得了一份8亿美元以上的合同，随后制造了成功瘫痪委内瑞拉防空系统的自杀式Switchblade无人机。鉴于其成功以及1.5万亿美元以上的国防开支，他们很可能成为巨大的赢家，且鉴于其市值较小，上行空间存在。
英文原文
It's not 10x but I really like $AVAV. They won a $800m+ contract earlier last Dec, and then made the kamikaze switchblade drones that successfully disabled Venezuelan air defenses. Given the success and $1.5T+ in defense spending, they'll likely be a huge winner, and upside is there given its small MC.
-
AIRO估值远低于同行，重估至AVAV水平仍有3倍空间，继续持仓。
在我看来，$AIRO 还有很长的路要走，这也是我继续持仓的原因。其交易估值约为3倍远期市销率(Forward P/S)，而其他无人机股票则超过30倍。当然，由于存在来自其他垂直领域的混合收入，情况更为复杂。但即使重新估值至保守的航空航天公司如 $AVAV，其价格也将是当前价格的3倍。
英文原文
$AIRO has a long way to go in my opinion, which is why I’m still holding. Trading at ~3 frwd p/s while other drone stocks are 30+. Of course this is more nuanced since there’s blended revenue from other verticals. But even rerating to conservative aerospace companies like $AVAV would be 3x current prices.
-
博主分享持仓 $AVAV 和 $UMAC，并解释发帖偏好。
@LibertadEterno 我一直在持有 $AVAV 和 $UMAC。不过并非每只股票都值得专门发帖，上面提到的那些是我的最爱。
英文原文
@LibertadEterno I’ve been riding $AVAV and $UMAC. Not every one deserves a post though, the ones above were my favorites.