供应链分析
产业链结构、上下游与瓶颈环节研究 · 共 1315 条 · 滚到底自动加载更多
-
澄清AMZN谣言，分析AXTI光学瓶颈及铜缆长期需求。
很多人误读了“无缆化”的评论，尽管主要下跌是由关于$AMZN的“color change”（色彩变化）谣言引起的。 你精准指出了$CRDO的看空逻辑，但这我们一直都知道。我的持仓观点较为微妙，因为我看到$AXTI等公司正在对规模化扩展所需的光学组件形成瓶颈。（尤其是昨天住友电工的西方供应链因出口管制而遭受重创，这 largely 是光学建设所必需的） 我们可能会看到对铜缆及铜缆寿命延长的长期需求，尽管Credo确实拥有光DSP垂直领域业务。
英文原文
Tons of people misinterpreted the "cableless comment", though the main drop was the $AMZN color change misinformation. So you've nailed the bear case for $CRDO, but we've always known this. My position is nuanced since I see $AXTI and others bottle necking optical components required for scale out. (especially now that the Western supply chain in Sumitomo got export control nuked yesterday, largely required for optical buildout) We'll likely see long tail demand for copper and copper life extension, although Credo does have optical dsp verticals.
-
AXTI垄断InP衬底，若断供将阻碍AI基建扩产。
@gemm_fc 2026日历年风险极小。我在此处发表了一些评论。 如果 $AXTI 停止出口磷化铟(InP)衬底，人工智能(AI)基础设施建设将无法扩产。 一旦竞争对手库存耗尽，AXT 将实质上成为垄断者。 https://t.co/P8LVlt2RkK
英文原文
@gemm_fc Very little risk for 2026 calendar year. Made some comments here. If $AXTI stopped exporting InP substrates, the AI buildout would not be able to ramp. AXT would effectively become a monopoly once inventory runs out for competitors https://t.co/P8LVlt2RkK
-
材料价格创历史新高，西方公司支付溢价，超大规模厂商尚未开始大规模采购。
@itsthesquonky 我会再给它几个月时间。你已经可以看到材料价格触及历史高点(ATH)，西方公司正在支付巨额溢价。 这还是在产能爬坡(pre-ramp)之前，$MSFT、$AMZN 和 $META 甚至 barely 开始。（$NVDA 预订了产能，造成了最初的冲击）
英文原文
@itsthesquonky I’d give it a few months. You can already see material prices hitting ATHs, with Western companies paying large premiums. This is pre-ramp too, $MSFT, $AMZN, and $META barely even started. ( $NVDA booked out capacity, causing the initial shock) https://t.co/rEG5X7Co81
-
美国短期依赖AXTI等解决InP衬底瓶颈。
@jvthed210725 并非如此，美国最终会绕过磷化铟(InP)衬底的瓶颈。只是对于2026-2027年，在现有架构下，他们将高度依赖 $AXTI 以及 $LITE、$COHR 等公司。
英文原文
@jvthed210725 Not really, the US will engineer around the InP substrates bottleneck eventually. It’s just that for 2026-2027 with the current architectures, they’ll be heavily reliant on $AXTI and companies like $LITE, $COHR.
-
中国出口管制致AXTI垄断InP衬底，成AI基建新瓶颈
英伟达CEO今日表示“内存瓶颈严重”。SK海力士($330B)、美光($MU, $355B)、三星($595B)。等着看当大家发现整个AI基础设施建设将被微小的$AXTI ($1B) 瓶颈化时，每个人的表情吧，尤其是在中国新的出口管制之后。
英文原文
Nvidia CEO said today that the “Memory Bottleneck is Severe” SK Hynix ($330B) $MU ($355B) Samsung ($595B). Wait and see the look on everyone’s face when they find out the entire AI buildout will be bottlenecked by the tiny $AXTI ($1B) after the China’s new export controls. https://t.co/iw6Nhitsry
-
中国出口管制重创日本InP供应链，SMTOY因业务占比小影响有限。
@thgstar2 作为股票，$SMTOY 可能表现尚可，因为它是一家体量极大的日本巨头，而磷化铟(InP)衬底生产仅占其业务的很小一部分。但中国今天通过出口管制，向日本的 InP 供应链投下了一枚“核弹”。https://t.co/Cq2wssDTA3
英文原文
@thgstar2 $SMTOY as a stock would likely be fine since it's an extremely large Japanese giant and InP substrate production is just a small part of their business. But China sent a nuke over to the InP supply chain in Japan with the export control today. https://t.co/Cq2wssDTA3
-
AI供应链因依赖单一源AXTI面临瓶颈风险，日本厂商或被迫切换产能。
这对AI基础设施建设绝对不是什么好事，因为整个AI供应链（例如$LITE、$COHR -> $GOOGL、$NVDA等）在接下来的两年里突然变得依赖于单一来源$AXTI。 一旦日本供应耗尽，像$LITE这样的公司可能会将产能从住友（Sumitomo）、JX转向$AXTI作为主要供应商（鉴于其已融资1亿美元用于扩大产能）。 不幸的是，日本没有美国那样的谈判能力来推翻这一局面。我们将看看事态如何发展。
英文原文
This is definitely not a good thing for the AI buildout because the whole AI supply chain eg. $LITE, $COHR -> $GOOGL, $NVDA others suddenly became dependent on a single source $AXTI for the next two years. Companies like $LITE will probably shift capacity from Sumitomo, JX to $AXTI as the main supplier (given they raised $100M for more capacity) once supply in Japan runs out. Unfortunately Japan doesn't have the same negotiating power as the US to overturn this. We'll see how this develops.
-
AXTI虽失日本市场，但成全球InP衬底主供。
这是一把双刃剑。$AXTI 可能无法再向日本出口，因此会失去那里的收入。然而，住友（Sumitomo）和其他衬底制造商将没有足够的材料来为全球其他地区生产足够的磷化铟(InP)衬底。因此，作为失去日本市场的代价，AXT 将成为全球主要供应商。
英文原文
It's a double edged sword. $AXTI likely cannot export to Japan anymore so it loses revenue there. However, Sumitomo and the other substrate makers would not have the materials to make enough InP substrates for the rest of the world. So AXT would be the main supplier globally for the tradeoff of losing the Japanese market.
-
中国禁运铟将切断日本InP衬底供应，重创AI供应链。
今天中国商务部宣布禁止向日本出口“两用”物品。 日本公司（如住友电气和JX日本矿业）是磷化铟(InP)衬底生产的领导者。但中国主导着生产InP衬底所需的关键原材料铟(Indium)的供应。 禁令措辞宽泛（“任何其他可能有助于增强日本军事能力的用途”）。这使得北京可以切断向日本科技巨头出口原材料铟，因为像住友这样的公司为许多军事用途做出贡献。磷化铟(InP)将符合这一定义。 日本制造商可能拥有库存，因此生产不会今天停止。但库存会在几个月内耗尽，日本将缺乏大规模制造成品InP衬底所需的原料。 同和(Dowa)等公司仍将能够提供一定数量，但没有中国，日本制造商无法创造出AI建设所需的足够InP衬底。 所以基本上，这就像是对日本供应链->西方供应链的一枚出口管制核弹，以巩固对中国的依赖。
英文原文
Today China’s Ministry of Commerce has announced a ban on the export of "dual-use" items to Japan. Japanese companies (like Sumitomo Electric and JX Nippon Mining & Metals) are the leaders in producing InP substrates. But China dominates the supply of Indium, the critical raw material required to produce InP substrates. The wording of the ban is broad ("any other uses that could contribute to strengthening Japan’s military capabilities"). This allows Beijing to cut off exports of raw Indium to Japanese tech giants because companies. eg Sumitomo contribute to many military use cases. Indium Phosphide (InP) would be under this definition. Japanese manufacturers likely have stockpiles, so production won't stop today. But stockpiles dwindle in a few months, Japan will lack the feedstock required to make the finished InP Substrates at scale. Dowa and others will still able to provide certain amounts, but without China, Japanese manufacturers cannot create enough InP substrate required for the AI buildout. So basically this was just an export control nuke on the Japanese supply chain -> Western Supply chain to solidify dependence on China.
-
AXTI因中国出口管制成InP衬底垄断者，股价大涨14%。
如果你想知道为什么 $AXTI 涨了 14%，大概是中国看到了这条推文，并向其唯一的另一家竞争对手发射了出口管制“核弹”。AXT 刚刚成为了磷化铟(InP)衬底的垄断者。https://t.co/i8FSYPscfx
英文原文
If you’re wondering why $AXTI is up 14%, China probably read this and sent an export control nuke to its only other competitor. AXT just became the monopoly of the InP substrates. https://t.co/i8FSYPscfx
-
AXTI对日营收受挫，博主将做后续分析
@AndDegen @illyquid @grok 情况很微妙，AXTI 对日本的销售收入也遭受重创。我会对此情况做后续分析。
英文原文
@AndDegen @illyquid @grok It’s nuanced, AXTI gets their revenue nuked to Japan as well. I’ll do a follow up analysis on the situation
-
解析委内瑞拉政权通过黄金与石油收入秘密积累超60万枚BTC的路径。
针对原帖被广泛转载，补充来源与澄清（基于此评论）： 威尔逊中心(The Wilson Center)与路透社(Reuters)的历史数据证实，马杜罗政权在2018年清算了约73吨黄金，当时价值约27亿美元。 - Whale Hunting的情报报告（引用HUMINT来源）评估，这批黄金资本立即转换为比特币以规避美国财政部冻结。 - 基于2018年交易价格（3k–10k美元），这一特定批次约占40万枚BTC。 - 剩余余额（使总估计值接近60万+ BTC）归因于石油收入。截至2025年底，报告显示该政权约80%的石油出口以USDT (Tether)结算，并将这些资金“清洗”进入比特币以避免冻结。 澄清： - 这是**推测金额**，因为马杜罗的比特币数据被刻意隐藏以规避制裁。 - 该图表显示的是**潜在**可扣押的比特币数量，而非实际已扣押。链上数据无法确认，因为委内瑞拉一直在积极规避制裁。 来源： Reuters: 2019年2月8日 – 独家报道确认2018年向土耳其和阿联酋出售73吨黄金。 The Wilson Center: 2021年6月 – 发布题为《委内瑞拉的威权盟友》的全报告（章节：“土耳其与委内瑞拉：便利联盟”），法证分析了2018年贸易数据。 Whale Hunting: 2026年1月3日 – 题为《600亿美元的问题》的报告，引用政权崩溃后收集的最新HUMINT（人力情报）。 Reuters: 2024年4月22日 – 独家报道详述PDVSA开始要求50%的USDT (Tether)预付款以避开冻结的银行账户。 Chainalysis / Binance Research: 2025年底 – 行业分析报告确认委内瑞拉石油中间商的USDT大量流入混币服务（“清洗”为BTC）。 Binance Research / Binance News: “委内瑞拉稳定币使用量预计因经济不稳定而上升” Chainalysis: “Chainalysis 2025加密货币犯罪报告” 1. 黄金清算（约27亿美元/73吨），Reuters: 文章《独家：委内瑞拉去年向土耳其、阿联酋出售73吨黄金》，日期：2019年2月8日 2. “入门资本”转换：文章《600亿美元的问题：委内瑞拉是否秘密成为比特币超级大国？》，日期：2026年1月3日 3. 石油换USDT转向（积累阶段）：文章《委内瑞拉国家石油公司寻求Tether (USDT)以绕过美国制裁》，日期：2024年4月22日
英文原文
Follow up sources + clarifications for the original post since this is being re-posted everywhere (piggybacking off this comment): Historical data from The Wilson Center and Reuters confirms the Maduro regime liquidated ~73 tons of gold in 2018, valued at approximately $2.7 billion at the time. - Intelligence reporting from Whale Hunting (citing HUMINT sources) assesses that this gold capital was immediately converted into Bitcoin to evade US Treasury freezes. - Based on 2018 trading prices ($3k–$10k), this specific tranche would account for roughly 400,000 BTC. - The remaining balance (pushing the total estimate toward 600,000+ BTC) is attributed to oil revenues. By late 2025, reports indicate the regime settled ~80% of oil exports in USDT (Tether) and "washed" these funds into Bitcoin to avoid freezing. Clarifications ) - This is the **speculated amount** since Maduro's Bitcoin figures were actively being hidden to avoid sanctions. - This chart figure demonstrates the **potential** amount of Bitcoin to be seized, not actively seized. On-chain data cannot confirm this as Venezuela has been actively avoiding sanctions. _ Reuters: February 8, 2019 – Breaking report confirming 73 tons were sold to Turkey and UAE in 2018. The Wilson Center: June 2021 – Full report released titled "Venezuela’s Authoritarian Allies" (Chapter: "Turkey and Venezuela: An Alliance of Convenience" by Imdat Oner), which forensically analyzed the 2018 trade data. Whale Hunting: January 3, 2026 – Report titled "The $60 Billion Question" referencing fresh HUMINT (Human Intelligence) gathered after the regime's collapse. Reuters: April 22, 2024 – Exclusive report detailing how PDVSA began requiring 50% prepayment in USDT (Tether) to avoid frozen bank accounts. Chainalysis / Binance Research: Late 2025 – Industry analysis reports confirming massive flows of USDT from Venezuelan oil intermediaries into mixing services (the "wash" into BTC). "Venezuela's Stablecoin Usage Predicted to Rise Amid Economic Instability"- Binance Research / Binance News "The Chainalysis 2025 Crypto Crime Report" - Chainalysis _ 1. The Gold Liquidation (~$2.7B / 73 Tons), Reuters: Article: "Exclusive: Venezuela sold 73 tonnes of gold to Turkey, UAE last year" Date: February 8, 2019 2. The "Entry Capital" Conversion: Article: "The $60 Billion Question: Is Venezuela Secretly a Bitcoin Superpower?" Date: January 3, 2026 3. The Oil-for-USDT Pivot (The Accumulation) Article: "Venezuela state oil firm looks to Tether (USDT) to bypass US sanctions" Date: April 22, 2024
-
梳理委内瑞拉从黄金变现到石油换USDT再转BTC的资本路径。
路透社：2019年2月8日——突发报道证实2018年向土耳其和阿联酋出售了73吨黄金。 威尔逊中心：2021年6月——发布题为《委内瑞拉的威权盟友》（章节：“土耳其与委内瑞拉：权宜之盟”作者Imdat Oner）的完整报告，对2018年贸易数据进行了法证分析。 Whale Hunting：2026年1月3日——题为《600亿美元之问》的报告，引用政权倒台后收集的最新人力情报(HUMINT)。 路透社：2024年4月22日——独家报道详述委内瑞拉国家石油公司(PDVSA)如何开始要求50%的USDT (Tether) 预付款，以避免银行账户被冻结。 Chainalysis / Binance Research：2025年底——行业分析报告确认大量USDT从委内瑞拉石油中间商流入混币服务（即“清洗”为BTC）。 “委内瑞拉稳定币使用量预计因经济不稳定而上升” - Binance Research / Binance News 《Chainalysis 2025加密犯罪报告》 - Chainalysis _ 1. 黄金变现（约27亿美元/73吨），路透社： 文章：“独家：委内瑞拉去年向土耳其、阿联酋出售73吨黄金” 日期：2019年2月8日 2. “入场资本”转换： 文章：“《600亿美元之问：委内瑞拉是否秘密成为比特币超级大国？》” 日期：2026年1月3日 3. 石油换USDT转向（积累阶段） 文章：“委内瑞拉国家石油公司寻求Tether (USDT) 以规避美国制裁” 日期：2024年4月22日
英文原文
Reuters: February 8, 2019 – Breaking report confirming 73 tons were sold to Turkey and UAE in 2018. The Wilson Center: June 2021 – Full report released titled "Venezuela’s Authoritarian Allies" (Chapter: "Turkey and Venezuela: An Alliance of Convenience" by Imdat Oner), which forensically analyzed the 2018 trade data. Whale Hunting: January 3, 2026 – Report titled "The $60 Billion Question" referencing fresh HUMINT (Human Intelligence) gathered after the regime's collapse. Reuters: April 22, 2024 – Exclusive report detailing how PDVSA began requiring 50% prepayment in USDT (Tether) to avoid frozen bank accounts. Chainalysis / Binance Research: Late 2025 – Industry analysis reports confirming massive flows of USDT from Venezuelan oil intermediaries into mixing services (the "wash" into BTC). "Venezuela's Stablecoin Usage Predicted to Rise Amid Economic Instability" - Binance Research / Binance News "The Chainalysis 2025 Crypto Crime Report" - Chainalysis _ 1. The Gold Liquidation (~$2.7B / 73 Tons), Reuters: Article: "Exclusive: Venezuela sold 73 tonnes of gold to Turkey, UAE last year" Date: February 8, 2019 2. The "Entry Capital" Conversion: Article: "The $60 Billion Question: Is Venezuela Secretly a Bitcoin Superpower?" Date: January 3, 2026 3. The Oil-for-USDT Pivot (The Accumulation) Article: "Venezuela state oil firm looks to Tether (USDT) to bypass US sanctions" Date: April 22, 2024
-
分析马杜罗政权黄金与石油收入转比特币以避制裁的推测数据。
关于60万美元（注：原文$600k在此语境下疑为笔误，结合后文应指60万枚BTC）数据的来源： - 威尔逊中心(The Wilson Center)和路透社(Reuters)的历史数据证实，马杜罗政权在2018年清算了约73吨黄金，当时价值约27亿美元。 - Whale Hunting的情报报告（引用HUMINT来源）评估，这批黄金资本立即被转换为比特币以逃避美国财政部冻结。基于2018年的交易价格（3000美元–10000美元），这一特定部分约占40万枚BTC。 - 剩余余额（使总估计值接近60万+ BTC）归因于石油收入。截至2025年底，报告显示该政权约80%的石油出口以USDT（Tether）结算，并将这些资金“清洗”为比特币以避免冻结。 这是推测的金额，因为马杜罗的比特币数据被刻意隐藏以避免制裁。链上数据无法确认这一点，因为这正是为了规避美国制裁。
英文原文
Regarding sources for the $600k figure: - Historical data from The Wilson Center and Reuters confirms the Maduro regime liquidated ~73 tons of gold in 2018, valued at approximately $2.7 billion at the time. - Intelligence reporting from Whale Hunting (citing HUMINT sources) assesses that this gold capital was immediately converted into Bitcoin to evade US Treasury freezes. Based on 2018 trading prices ($3k–$10k), this specific tranche would account for roughly 400,000 BTC. - The remaining balance (pushing the total estimate toward 600,000+ BTC) is attributed to oil revenues. By late 2025, reports indicate the regime settled ~80% of oil exports in USDT (Tether) and "washed" these funds into Bitcoin to avoid freezing. This is the speculated amount since Maduro's Bitcoin figures were actively being hidden to avoid sactions. On-chain data cannot confirm this as again, this was done to avoid US sanctions.
-
软件仍复杂，但AI瓶颈已下移至材料与先进封装层。
我不同意，对于SaaS软件服务(SaaS)确实如此，但用于运行数据中心(数据中心)及其他垂直领域的软件（例如针对$CRWV和$NBIS的GPU编排(GPU Orchestration)）仍然极其复杂。不过，大多数瓶颈已下移至材料/基础设施层面，如磷化铟(InP)衬底、集成电路(IC)衬底、高端PCB材料/原料，以及更高层级的HBM和高密度互连封装(CoWoS)。
英文原文
I disagree, for SaaS sure, but software (eg. GPU orchestration for $CRWV, $NBIS) for running DCs to other verticals is still incredibly complex. But most bottlenecks moved down to materials/infra like InP substrates, IC substrates, high-end pcb materials/feedstock, then a level higher with HBM and CoWoS
-
委内瑞拉政权更迭带来资源与不良资产投资机遇。
国家重建是投资者极少见且最佳的机会之一。随着政权更迭，它现在由美国政府掌控。委内瑞拉是一座资源金矿。像 $ASHM 持有的债券等大量不良资产，由于鲜有机构抢先布局（因为这是突袭式的军事行动），通过发现合适的公司存在巨大的潜在上行空间。这与像 $CVX 这样资产曾被锁定如今已解锁的公司，或在那里运营的银行（参考阿根廷银行/股票的情况）类似。
英文原文
Nation building one of the rarest and best opportunities as an investor. With the regime change, it's now run by the US gov. Venezuela is a resource gold mine. There's a lot of distressed assets like bonds that $ASHM have, and given little institutions frontran these events (since it was a surprise military action), tons of potential upside by discovering the right companies. Same with companies like $CVX that had assets trapped and are now unlocked or banks that operated there (look at what happened to Argentina banks/stocks)
-
解析委内瑞拉政权更迭后，从债务重组到能源基建的重建投资逻辑。
美国现已推翻委内瑞拉马杜罗政权。 每个人的第一反应都是： 我该如何从委内瑞拉的“国家重建”中获利？ 方法如下： 1. 困境债务与债券 ($ASHM, $HLI, $LAZ) Ashmore ($ASHM) 拥有“纯题材”困境委内瑞拉债务，预计可获得2-3倍收益。他们是委内瑞拉债务最大的机构持有者。 债券交易价格在10-20美分区间（取决于制裁波动）。在政权更迭情景下，分析师（花旗、安联）估计每美元可回收30-55美分。相对于账面资产，这约为2-3倍收益。 $HLI - 其投行部门是委内瑞拉债权人委员会的主要财务顾问。在主权债务重组中，顾问收取“成功费”，是重组领域的“卖铲人”。 $LAZ - 主权债务重组（曾顾问希腊、乌克兰）。该公司受益于交易的复杂性，而委内瑞拉的债务结构 arguably 是历史上最复杂的。 2. 重质原油移植 (巴黎: TE, $GHM) Technip - 委内瑞拉关键基础设施的历史架构师。新政府可能会向原始设备制造商(OEM)授予“免标”或“单一来源”服务合同以加快修复，因为引入新公司逆向工程工厂需要数年。 $GHM - 这家小型工业公司制造炼油厂和升级装置中使用的真空喷射系统。要升级委内瑞拉的重质原油，必须在真空中蒸馏以防止其变成固体焦炭。 3. 稀释剂 ($TRGP) 在委内瑞拉大量出口石油之前，必须大量进口稀释剂（石脑油或天然汽油）以使重质原油通过管道流动。 $TRGP 运营加拉纳帕克海洋终端，这是向委内瑞拉发送稀释剂的主要枢纽。 回归美国供应意味着 Targa 的加拉纳帕克海洋终端（休斯顿主要的LPG/石脑油出口枢纽）将立即看到巨大的量增，以取代伊朗供应。 4. 银行板块 (巴拿马: MVZ.A / MVZ.B) Mercantil 是一个独特的异常值，一家在巴拿马上市并在美国有业务（Amerant已分拆，但Mercantil保留）的委内瑞拉银行控股公司。 它是美元化资金流、汇款和援助资金从美国/迈阿密回流重建中的加拉加斯最合理的“桥梁”。 5. 能源板块 ($CVX, $VLO, $PSX) 政权更迭和国家重建最明显的受益者是雪佛龙 ($CVX)。与其他离开的美国巨头不同，雪佛龙在委内瑞拉保持了存在。他们拥有人员、许可证（通过OFAC）和油田（Petroboscan, Petropiar），可以立即扩大产能。 墨西哥湾沿岸炼油商 $VLO 和 $PSX 也将受益，因为他们在德克萨斯州和路易斯安那州的炼油厂是专门为了处理委内瑞拉的重质、高硫原油而建的。自制裁实施以来，他们不得不从其他地方购买更昂贵的重质原油。委内瑞拉石油的涌入将大幅降低他们的原料成本，扩大利润空间。 人工智能是2025年最盈利的交易之一，并延续至2026年。 鉴于政权快速更迭的意外转折，投资于从银行到石油加工的国家重建可能成为2026年最盈利的交易。
英文原文
US has now toppled Maduro's regime in Venezuela. Everyone's first thought is: How do I profit off Nation Building in Venezuela? Here's how: 1. Distressed Debt + Bonds ( $ASHM, $HLI, $LAZ ) 2x–3x multiple from "pure play" distressed Venezuelan debt owned by Ashmore $ASHM. They are the largest institutional holders of Venezuelan debt. Bonds trade in the 10–20 cent range (depending on sanctions flux). In a regime change scenario, analysts (Citi, Allianz) est a recovery of 30–55 cents on the dollar. ~2x–3x multiple on book of assets that are marked down. $HLI - IB is the primary financial advisor to the Venezuela Creditor Committee. In sovereign restructurings, the advisors are paid "success fees", and are the "picks and shovels" play for restructuring. $LAZ - sovereign debt restructuring (advised Greece, Ukraine). This firm benefit from the complexity of the deal, and Venezuela's debt stack is arguably the most complex in history. 2. Heavy Crude Transplant ( Paris: TE, $GHM) Technip - Historical architect of Venezuela’s critical infrastructure. The new gov will likely award "no-bid" or "sole-source" service contracts to the OEM to expedite repairs, as bringing in a new firm to reverse-engineer the plants would take years. $GHM - The small-cap industrial firm manufactures the vacuum ejector systems used in refineries and upgraders. To upgrade Venezuela's heavy oil, you must distill it under a vacuum to prevent it from turning into solid coke. 3. Dilutents ( $TRGP ) Before Venezuela can export high volumes of oil, it must import high volumes of diluent (naphtha or natural gasoline) to make the heavy crude flow through pipelines. $TRGP operates the Galena Park Marine Terminal, the primary hub for sending diluents to Venezuela. Reverting to US supplies means Targa’s Galena Park Marine Terminal (a major LPG/Naphtha export hub in Houston) would see an immediate massive spike in volume to displace the Iranian supply. 4. Banking Plays ( Panama: MVZ.A / MVZ.B) Mercantil is a unique anomaly, a Venezuelan bank holding company that listed in Panama and has a US presence (Amerant was spun off, but Mercantil remains). It is the most logical "bridge" for dollarized flows, remittances, and aid money moving from the US/Miami back into a reconstructed Caracas. 5. Energy Sector ( $CVX, $VLO, $PSX) The most obvious beneficiary of regime change and nation building in Venezuela is Chevron $CVX. Unlike other US majors that left, Chevron has maintained a presence in Venezuela. They have the staff, the licenses (via OFAC), and the fields (Petroboscan, Petropiar) ready to ramp up immediately. Gulf Coast Refiners $VLO and $PSX would stand to benefit as well as their Texas and Louisiana refineries were specifically built to process Venezuela's heavy, sour crude. Since sanctions hit, they have had to buy more expensive heavy crude from elsewhere. A flood of Venezuelan oil would drastically lower their feedstock costs, widening their profit margins. Artificial intelligence was one of the most profitable trades in 2025 and moving forward to 2026. Given the unexpected turn of events with a fast regime change, investing in Nation Building from banks to oil processing might become the most profitable trade in 2026.
-
分析师对二阶效应分析精准，市场已现抢跑现象。
@Citrini7 你们的分析师对二阶效应(second-order effects)的分析非常精准。我们已经看到了抢跑(frontrunning)的现象 https://t.co/29HglUyQEE
英文原文
@Citrini7 Your analysts are spot on about second-order effects. We're already seeing frontrunning https://t.co/29HglUyQEE
-
分析美侵委局势下，重油、氮肥及海军装备供应链的受益标的。
美国现已入侵委内瑞拉。 大家可能都在想同一个问题： 如何从这一局势中获利？ 1. 重质原油(Heavy Sour)、氨(Ammonia)和氮肥(Nitrogen Fertilizers)供应中断（$CF, $CVE）。 这些是委内瑞拉最大的出口产品。 大多数人会购买泛石油ETF或轻质甜原油(light sweet crude)生产商。这效率低下，因为轻质原油在复杂炼厂中并非重质原油的完美替代品。如果加勒比海的氨供应受阻，全球氮价将飙升。最大受益者将是使用廉价美国天然气且不依赖加勒比海运航线的美国本土生产商。 2. 劣质原油加工(Dirty Crude Processing)（$VLO）——如果竞争对手缺乏委内瑞拉原油，Valero从多元化来源获取重质原油的能力（及其对柴油利润率的杠杆作用）使其具备韧性。 3. 海军作战（$LDOS）——当散户投资者购买洛克希德·马丁(F-35)时，加勒比海地区的行动侧重于海上监视、作战和自主巡逻，以在不危及美国人员的情况下执行封锁。像Leidos这样的公司提供此类海军技术。 4. 从$AVAV到$HII和$LHX的国防和航空航天领域也受益。 - $AVAV最近推出了专为海上行动设计的Red Dragon和更新版Switchblade 600变体 - $LHX提供将无人机($AVAV)与舰船($HII)及喷气式飞机($BA)连接的传感器和通信设备。 - 封锁需要大量的海上监视和海军资产，这有利于造船商($HII) 5. 近期军事行动的直接供应商： - 来自$BA的F/A-18E/F超级大黄蜂（对加拉加斯进行精确打击） - 来自$BA的B-1B枪骑兵 - UAS（无人机），MQ-9收割者 - $RTX（MTS-B传感器），$HON霍尼韦尔提供发动机 - 战斧巡航导弹(TLAM)，$RTX 目前为止： $AVAV - 5.91%+ $BA - 4.91% $LHX - 3.72% $CF - 3.61%+ $CVE - 3.61%+ $HII - +2.85% $RTX - 2.1% $VLO - 1.55%+ $LDOS - 1.7%+ $HON - .4%+
英文原文
US has now invaded Venezuela. Everyone is probably wondering the same thing: How do you profit off the situation? 1. Heavy Sour, Ammonia, and Nitrogen Fertilizers disruption ( $CF , $CVE). These are Venezuela's biggest exports. Most people will buy generic oil ETFs or light sweet crude producers. This is inefficient because light oil is not a perfect substitute for heavy oil in complex refineries. If Caribbean ammonia is stranded, the global price of nitrogen spikes. The biggest beneficiary is a US-domestic producer that uses cheap US natural gas and doesn't rely on Caribbean shipping lanes 2. Dirty Crude Processing ( $VLO ) - If competitors are starved of Venezuelan oil, Valero’s ability to source heavy crude from diverse locations (and its leverage to diesel margins) makes it resilient. 3. Naval Warfare ( $LDOS) - While retail investors buy Lockheed Martin (F-35s), the operations in the Caribbean focuses on maritime surveillance, warfare, and autonomous patrolling to enforce blockades without risking US personnel. Companies like Leidos provide these tpyes of naval tech. 4. Defense and aerospace from $AVAV to $HII and $LHX also benefit. - $AVAV recently unveiled the Red Dragon and updated Switchblade 600 variants specifically for maritime operations - $LHX provides the sensors and communications gear that link the drones ($AVAV) to the ships ($HII) and the jets ($BA). - A blockade requires significant maritime surveillance and naval assets, which benefits shipbuilders ( $HII ) 5. Direct Suppliers of recent military operation: - F/A-18E/F Super Hornet from $BA (Precision strikes on Caracas) - B-1B Lancer from $BA - UAS (Drone), MQ-9 Reaper - $RTX (MTS-B Sensors), $HON Honeywell for the Engine - Tomahawk (TLAM), $RTX So far: $AVAV - 5.91%+ $BA - 4.91% $LHX - 3.72% $CF - 3.61%+ $CVE - 3.61%+ $HII - +2.85% $RTX - 2.1% $VLO - 1.55%+ $LDOS - 1.7%+ $HON - .4%+
-
磷化铟许可瓶颈难改，镓锗政策稳定则供应可续。
@Nichola30641241 所以磷化铟(InP)一直是一个许可瓶颈，我不认为这会改变。镓/锗出口政策只是半导体材料板块的一个代理指标。因此，如果该政策在一年内保持稳定，磷化铟(InP)很可能将继续通过许可流程流通。
英文原文
@Nichola30641241 So InP was always a permit bottleneck, don't see that changing. The gallium/germanium export policy is just a proxy of the semi materials sector. So if that is fine for a year, the InP will likely continue to flow through permitting.
-
解析AXTI作为光子AI供应链底层材料商的关键地位。
如果 $GOOGL 想建造更多 TPU，就需要 $LITE 的光子共封装(OCS)。$LITE 需要 $AXTI 的衬底（据估占独立市场30%，CEO称占磷化铟(InP)供应链40%）。$AXTI 需要由 AXT 拥有的化学品和材料。 $AXTI 是整个光子学 AI 供应链的底层，因为他们拥有10多家生产高纯度化学品的材料公司。 基本上现在就像一个材料对冲基金，生产输出衬底，因此对关键的 InP 瓶颈拥有如此大的控制力和市场份额。
英文原文
If $GOOGL wants to build more TPUs, they need $LITE’s for OCS. $LITE needs $AXTI substrates (est. 30% of merchant market, CEO says 40% of InP supply chain). $AXTI needs the chemicals and materials from that happened to be owned by AXT. $AXTI is the bottom of the entire AI supply chain for photonics since they own 10+ materials companies that produce the high-purity chemicals. Basically a materials hedge fund at this point that produces the output substrate, hence why it's has so much control and market share over the critical inp bottleneck.
-
分析SKC和SHMD在CPO及半导体供应链中的机会。
谢谢！虽然还有很多遗漏，但我仍在深入研究它们。 $SKC/Absolics（市值约25亿美元）作为共封装光学(CPO)领域的首发者，且美国政府深度参与，对于$NVDA、$AVGO/$MVRL而言颇具看点。 $SHMD（市值约2.5亿美元）也很有趣，因为$AVGO可能是其客户，且他们可能处于$INTC和三星的路线图之中。 但总体而言，市场提供了大量有趣的机会。
英文原文
Thanks! There's a lot I'm missing but still doing a lot of research into them. $SKC/Absolics (~$2.5B MC) was interesting for $NVDA, $AVGO / $MVRL CPO play as the first-to-market with US Government heavily involved, $SHMD (~$250m) was also pretty interesting too since $AVGO was a likely customer and they're probably on $INTC, Samsung roadmaps. But generally market presents a ton of interesting opportunities.
-
AXTI和SHMD的瓶颈论点同样适用于上市公司。
@retail_mourinho 谢谢！希望你度过了一个愉快的新年。 目前很多讨论集中在私募市场，但关于 $AXTI 或 $SHMD 的瓶颈论点同样高度适用于上市公司。
英文原文
@retail_mourinho Thanks! Hope you’ve had a great new years Lot of the discussion in around private markets but the same thesis around bottlenecks with $AXTI or $SHMD is highly relevant to public companies
-
AI与机器人领域迎来拐点，软体机器人商业化加速，应用前景广阔。
AI 的情况也是如此，我大约在 8 年前发表过几篇关于深度学习(Deep Learning)的基础论文，但当时并未预期很快能看到商业化落地。直到 ChatGPT 横空出世。机器人领域也是如此，我的一位同事在斯坦福大学攻读该领域的博士学位，随后获得 DARPA 资助，致力于研发聚合物肌肉纤维(Polymer Muscle Fibers)和模仿生物运动的机器人。据我所知，作为行业先驱之一，他们已在此领域耕耘了 6-7 年，但直到去年才真正加速发展。总体而言，过去一年是 AI/机器人领域的拐点，随着大语言模型(LLMs)赋予其“大脑”，软体机器人(Soft Robotics)（如 Optimus 等）正逐渐变为可商业化的现实。因此，赋予机器人更多类人动作/灵活性是下一步自然演进。其在国防及其他垂直领域的应用前景广阔。
英文原文
Same with AI ig, I published a few of the few fundamental deep learning papers like 8y ago, but didn't really expect to see anything commercialization for awhile. Then ChatGPT popped up out of nowhere. I'd say same with robotics, one of my colleagues did his stanford phd in this area and went out to get darpa funding to build out polymer muscle fibers and robots that mimic human creature movements. Pretty sure they've been building it for the past 6-7 years or so as one of the pioneers of the industry but it only really picked up steam this past year. But in general past year has really been the inflection point for ai/robotics and we're finally seeing soft robotics become more of a commercializable reality (eg. extension of optimus and others) with LLMs giving it a brain. So the next natural evolution is giving robotics more life-like movements/flexibility. It's really an open book for applications to defense and other verticals.
-
看好能取代人力、实现端到端自主运营的AI，而非仅改善利润率的辅助型AI。
我认为 $LMND 式的 AI 只是公司试图改善利润率(Margins)的副产品。我更关注基础层面的东西，即 $PLTR 代表的下一次进化，它能真正取代整个人类劳动力。以编程为例（Cursor, Windsurf），它们很棒但仍仅作为辅助工具。我们尚未实现向零监督(Zero-supervision) AI 的飞跃。例如在商务拓展(BD)中，每家公司都需要大量的外联和沟通循环。技术已存在，但依然过于碎片化。目标是建立一个可扩展的自主系统，无需人工介入，端到端地处理从外联到客户成交的整个流程。如果有私营或上市公司能做到这一点，我认为其价值将不可估量。
英文原文
I see $LMND-style AI as a byproduct of companies just trying to improve their margins. I'm more interested in the foundational stuff, next evolution $PLTR, that actually replaces the entire human workforce. For coding, as an example (Cursor, Windsurf), they're great but still used as assistance. We're not quite at the jump to zero-supervision AI. For example, in BD, you have a massive loop of outreach and communication that every company needs. The tech is there, but it’s still too fragmented. The goal is a scalable, autonomous system that handles the entire pipeline from outreach, to customer closing across every segment, end-to-end without a human in the loop. And if a private company/public company can make that happen, then I see that being incredibly valuable.
-
对比JTNC与AbsoLics，看好后者在CPO及玻璃基板领域的先发优势。
是 $JTNC 吗？很难想到多少纯概念股，但它们做的是 TGV，而非用于 $NVDA 共封装光学(CPO) 的光学方案，如康宁/AbsoLics。我原本更关注 CPO 方面，但还有其他机会。AbsoLics 似乎已是明确的首选，因为其市值仅 25 亿美元，获得美国 CHIPS 法案政府资助，且率先上市。尽管背负沉重包袱，但鉴于其体量较小，其玻璃基板垂直整合业务应能强力带动整体表现。
英文原文
Is it JTNC? Can't really think of many pure-play, but they're for TGV, not optical used for $NVDA CPO like Corning/Absolics. I was more focused on CPO aspect but there's other opportunities. Absolics seems like the clear favorite already since it's $2.5B MC, funded by US Gov from CHIPS, and first-to market. It does carry heavy baggage, but its glass substrate vertical should just hard-carry the rest given the small size.
-
2026年十大主题投资：聚焦AI供应链瓶颈、软体机器人及支付颠覆。
2026年通讯。 主题投资：演进、颠覆与瓶颈 1. 软体机器人 - 向 $TSLA、$ONDS、波士顿动力演进。 2. 硅光子(SiPh) - 磷化铟(InP)瓶颈 | $AXTI、$LITE、$GOOGL 3. 玻璃基板 - 瓶颈 | $NVDA、$INTC、$TSM 4. 资金流动 - 对 $V、Stripe、$BOA 的颠覆 5. AI云层级 - 瓶颈 | $NBIS、$IREN、$HUT 6. LLM网络安全 - 向 $CRWD、$CSCO、$MSFT 演进 7. 低轨(LEO)太空基础设施 | 向 $RKLB、SpaceX、$ASTS 演进 8. 消费者代理工作流(50步) - 对消费者劳动力的颠覆，来自Manus、$PATH Cognition 9. 分布式计算延迟 - 瓶颈 | $TSLA、$AMZN、$GOOGL 10. 铜互连寿命延长 - 瓶颈 | $NVDA (LPU/Groq)、$AMD、$INTC _这是我对从公开信息综合及瓶颈二/三阶效应来看最感兴趣的主题投资的简要概述！_ 1. 软体机器人：向机器人的演进 传统机器人（Optimus、波士顿动力）依赖逆运动学控制刚性关节。软体机器人改变了数学模型。 我们已到达硬件（Optimus、波士顿动力、Figure）与LLM（Gemini、Grok、Opus）相遇的节点，正处于大规模商业化的开端。 通过使用受章鱼触手和人类皮肤启发的材料，机器人正从齿轮转向流体性，以处理极其精细的任务，如像人手一样处理农产品，或为 $ONDS/Andruil 无人机添加类章鱼延伸以拾取极重表面。 这种演进在于跳出思维定势思考机器人能做什么。我记得7年前曾与该领域的斯坦福博士合作，AI在多年研究后开始商业化，因此该领域也是如此。 将类生物流体性添加到刚性机器人中的可能性是无限的，这只是自然演进。 大多数可能是私人公司。 2. 硅光子 - AI基础设施的瓶颈“磷化铟(InP)卡脖子点” 从Blackwell Ultra集群到Google TPU已触及上限，需要光子互连 | 共封装光学(OCS)以实现扩展。 基板：$AXTI（通过Tongmei）和住友（日本）控制全球约60-70%的InP基板市场。 材料：Vital Materials（中国）和AXT等公司控制原材料铟本身的精炼（78%+的供应链）。 如果你是美国科技巨头，你2026年的整个“AI增长故事”取决于由地缘政治对手控制的材料。 唯一可扩展的解决方案是工程绕行，要么实现芯片上光传输，同时减少90%的铟使用，要么使用微小的磷化铟薄片代替大型昂贵晶圆。 瓶颈本身有机会，如AXT、住友。或帮助解决它的公司如 $POET。 3. 玻璃基板 - 解决从 $NVDA 到其他公司的共封装光学(CPO)瓶颈 向玻璃基板的转变本质上是半导体行业对当前材料物理极限的回答。 当前芯片位于有机材料（本质上是专用塑料）制成的基板上。随着芯片变大，如Nvidia巨大的GPU封装，塑料基板会翘曲。 因此，玻璃基板正成为共封装光学(CPO)的行业标准，因为它们解决了光子学中最大的对齐问题。 美国政府已视其为必要，我们看到巨额补贴流向这些公司。 $INTC、三星电子、Absolics（SKC子公司）、DNP等是主要受益者，尤其是随着MRVL和 $AVGO（推动光学开关的玻璃）推进CPO革命。 4. 资金流动 - 对卡网络、银行、交易所和支付的颠覆 几十年来，资金转移一直是“收费公路”业务。每次刷卡，2%到3%的钱流入卡网络（Visa/Mastercard）和发卡行的口袋。 或者从交易所买卖加密货币是0.2-1%。这是历史上最有利可图、“不可杀死”的商业模式。 直到现在。2025年的“天才法案”刚刚将金钱传输许可证或银行特许状交给像 $XRP 这样的公司，赋予了他们王国钥匙。 对我来说并非理论。我恰好正在自己的初创公司与创建V / $PYPL 实时支付网络的人一起从事这项工作。 但基本上，拥有现有MTL或追求银行特许状并利用天才法案及其他技术的公司，现在可以通过在美联储和区块链之上进行结算来绕过传统百分比费用，有效地将基于百分比的费用转化为几美分。 99%的公司会这样做吗？可能不会，因为支付行业的所有利润率都将归零。但我乐意看到。 但基本上，Stripe以11亿美元收购Bridge本应是对现有公司的红色警报，表明1天ACH、 interchange模式、25美元国际转账的日子即将结束。 这扩展到许多其他相邻领域，从低费用颠覆者如 $HOOD、Mercury，一直到稳定币新银行，或制作自己稳定币的公司如 $SOFI。 5. AI云层级 - 超大规模计算瓶颈的解决方案 当超大规模云厂商被困在3-5年的电网互连队列中时，像WULF和IREN这样的矿工今天就拥有即插即用的GW级算力。 这是千载难逢的机会，超大规模云厂商将其现金牛云收入流向小公司。 这里有不同的层级，从Fluidstack、Poolside、Fireworks在GPU编排层，到IREN等公司构建的裸金属层。 然后有成为超大规模云厂商本身，如NBIS拥有物理位置、GPU、软件编排，然后为推理提供简单接口。 这是少数小公司在未来一两年成为AWS或Azure，或被收购（例如GOOGL以47亿美元收购Intersect）的机会。 像NBIS、IREN、CRWV这样的新云，以及像CIFR、WULF、HUT这样的colo玩家（以及私人部门->能源）将受益。 6. LLM网络安全 - 向现代安全和漏洞防御的演进 最近的报告（例如来自Anthropic红队）显示，高级模型如Opus（及未来版本）可以自主扫描开源智能合约，并在几分钟内识别价值数百万美元的“零日”漏洞。 含义：如果AI能在不可变的区块链合约中找到逻辑缺陷，它也能在银行的SWIFT API或电网控制软件中找到缺陷。 同样适用于KYC/AML。像Gemini Nano Banana这样的模型能够创建逼真的图像/视频，人们能够绕过许多程序。 这个领域有很多不性感但具Alpha潜力的事情，如LLM自动化SOC2/PCI DSS合规，代理坐在服务器上，持续监控日志，并自动生成审计所需的证据。 7. 低轨(LEO)太空基础设施 | 向拓展最后疆域的演进 太空是下一个大事情。这并不新鲜。（希望你懂这个笑话）。但从像 $RKLB、SpaceX这样的公司，到解决轨道拥堵或发射节奏瓶颈的公司，再到像ASTS或Starlink这样商业化基础设施的公司，在未来一年呈现许多机会。 因此，像Impulse、Blue Origin、$ASOZF到RKLB、$ASTS这样的公司将受益于整个链条。 8. 消费者代理工作流(50步) - 对消费者劳动力的颠覆，来自Manus、PATH Cognition 这一点很简单，无需解释。但在对就业+成本节约的潜在影响上显而易见。 你如何自动化商务拓展？如何自动化营销？如何自动化软件工程师？ 这超越了ChatGPT的几步回答，直接进入现实世界，AI代理可以在X上漫游，找到合适的人，发送DM，继续对话，并在一个工作流中导致销售电话。 这是“聊天机器人”时代的结束和“行动”时代的开始，取代公司以前需要的所有人。 我尚未看到任何公司大规模做到这一点。拥有这些的公共公司如META并没有呈现最佳敞口。也许是 $PATH 在公共空间。 9. 分布式计算延迟 - 解决AI计算容量紧张的瓶颈 像GOOGL Cloud、MSFT Azure这样的超大规模云厂商已达最大容量。 Elon Musk已经提出分布式计算作为解决此问题的未来（例如，拥有 $TSLA 网络为LLM推理提供计算）。 “Tesla计算云”论点很迷人，但我识别出的最大物理障碍是：推理延迟。 要生成“Token B”，模型必须先生成“Token A”。它不能同时做两者。如果你将一个巨大模型（如Grok-3）拆分到5辆不同的汽车中以适应内存，你必须为每个生成的Token在这些汽车之间发送数据。 因此，如果汽车之间的网络延迟甚至是20ms（5G的乐观估计），而你生成50个Token，你刚刚在计算时间之上添加了1秒的纯“等待时间”（延迟）。在使用NVLink的数据中心中，该等待时间以纳秒计。 同样适用于零售用户拥有的任何备用计算机、GPU等。有数十亿消费级GPU（Teslas、iPhones、游戏PC）90%的时间闲置。 解决推理的“分布式延迟”问题呈现了计算史上最大的套利机会之一。 尚未看到任何公司大规模完成此任务。也许NVIDIA Dynamo、$AKAM、TSLA正在接近。 10. 铜互连寿命延长 - 解决Nvidia和其他公司的瓶颈 既然我们不能拥有无限的InP，我们必须用现有的东西（例如铜）进行工程绕行，所以铜电缆可以做物理上说它不应该做的事，如在不损失信号的情况下跨机架传输224G信号。 行业在InP上遇到了硬性停止，美国在物理上无法开采和精炼足够的InP将数据中心中的每个链接变成光纤。 如果有任何帮助，那就是好事。例如，NVDA对Groq团队和IP的200亿美元“收购雇佣”。LPU更多是关于推理延迟/架构，但它作为副产品解决了铜寿命延长。Groq的整个架构在延迟上击败了Nvidia，因为它拒绝了光学。Groq使用“确定性”网格，依赖芯片之间的直接电气（铜）连接，避免光学开关的“抖动”和转换时间。 像 $ALAB、$CRDO、Groq，或任何能找到用铜绕过光学瓶颈方法的公司将是赢家。 _这里有从私人部门投资到公共部门的众多交易！只是今天即兴写下了我的想法，但乐意稍后详细阐述。 无论如何，我相信这些主题投资中的许多： 从投资InQ瓶颈绕行（$POET）或瓶颈本身（$AXTI）到公共部门的颠覆者（$CRCL）。 到投资铜扩展瓶颈修复（Groq）、银行特许状颠覆者（Mercury）到私人部门的演进公司（Lightmatter、Festo）。 在2026年呈现不对称上行空间。 新年快乐！
英文原文
2026 Newsletter. Thematic Investments: Evolution, Disruption, and Bottlenecks 1. Soft Robotics - Evolution to $TSLA, $ONDS, Boston Dynamics. 2. SiPh - InP Bottleneck | $AXTI, $LITE, $GOOGL 3. Glass Substrates - Bottleneck | $NVDA, $INTC, $TSM 4. Money Movement - Disruption to $V, Stripe, $BOA 5. AI Cloud Layers - Bottleneck | $NBIS, $IREN, $HUT. 6. LLM Cybersecuirty - Evolution to $CRWD, $CSCO, $MSFT 7. LEO Space Infrastructure | Evolution to $RKLB, SpaceX, $ASTS 8. Consumer Agentic Workflows (50 Step) - Disruption to the Consumer Workforce, from Manus, $PATH Cognition 9. Distributed Computing Latency - Bottleneck | $TSLA, $AMZN, $GOOGL, 10. Copper Interconnect Life Extension - Bottleneck | $NVDA (LPU/Groq), $AMD, $INTC _ This is an light overview of thematic investments I find the most interesting from a public-information synthesis perspective + second/third-order effects from bottlenecks! _ 1. Soft Robotics: The Evolution to Robotics Traditional robotics (Optimus, Boston Dynamics) relies on Inverse Kinematics to rigid joints. Soft robotics changes the math. We've met the point where hardware (Optimus, Boston Dynamics, Figure) met LLMs (Gemini, Grok, Opus), and we're at the beginning of possible widespread commercialization. By using materials inspired by octopus tentacles and human skin, robots are moving away from gears and toward fluidity to handle extremely delicate tasks like handling produce like the human hand, to picking up extremely heavy surfaces adding Octopus-like extensions to $ONDS/Andruil Drones. The evolution is thinking outside the box in terms of what robotics can do. I remember working with some Stanford PHds in this field like 7 years ago, and it just so happens AI is starting to be commercialized after many years of research. So expected, this field to be as well. Possibilities are limitless adding organism-like fluidity to rigid robotics, this is just the natural evolution. Most of these are prob private companies. _ 2. Silicon Photonics - Bottleneck of the AI Infrastructure "InP Chokepoint" Blackwell Ultra Clusters to Google TPUs have hit the upper wall and requires photonics for interconnects | OCS to scale up. The Substrates: $AXTI (via Tongmei) and Sumitomo (Japan) control roughly 60-70% of the world's InP substrate market. The Materials: Companies like Vital Materials (China) and AXT control the refining of the raw Indium itself (78%+ of supply chain). If you are a US tech giant, your entire "AI Growth Story" for 2026 depends on materials controlled by geopolitical rivals. The only scalable solution is engineering around it, either by delivering light-on-chip, while using 90% less InP or companies that use tiny slivers of Indium Phosphide instead of large, expensive wafers. There's opportunities with the bottleneck itself like AXT, Sumitomo. Or companies that help address it like $POET. _ 3. Glass Substrates - Fixing the Bottleneck for CPOs from $NVDA to others. The shift toward glass substrates is essentially the semiconductor industry’s answer to a physical wall they are hitting with current materials. Current chips sit on a substrate made of organic materials (essentially specialized plastic). As chips get larger, like Nvidia's massive GPU packages, plastic substrates warps. So, glass substrates is becoming the industry standard for Co-Packaged Optics (CPO) because they solve the single biggest problem in photonics with alignment. US Government already sees this as a necessity and we've seen huge subsidies funneling down to some of these companies. Companies like $INTC, Samsung Electronics, Absolics (SKC Subsidiary), DNP, and others are the main beneficiaries, especially as MRVL and $AVGO (driving glass for optical switches) move forward with CPO revolution. _ 4. Money Movement - The Disruption to Card Networks, Banking, Exchange, and Payments For decades, moving money has been a "toll road" business. Every time you swiped a card, 2% to 3% of that money vanished into the pockets of the Card Networks (Visa/Mastercard) and Issuing Banks. Or buying/selling crypto from an exchange would be .2-1%. It was the most profitable, "un-killable" business model in history. Until now. The "Genius Act" of 2025 just handed companies like $XRP with Money Transmitter Licenses or Banking Charters the keys to the kingdom. Not really theoretical for me. I happen to be working on this myself at my own startup with some folks who created V / $PYPL's real-time payment networks. But basically companies with existing MTLs or pursuing banking charters leveraging the Genius Act and some other tech can now bypass legacy % fees by doing settlement on top of the Federal Reserve and blockchains, effectively converting percentage-based fees into a few cents. Would 99% companies do it? Probably not since every single margin from across the payment industry would just go to 0. I'd be happy though. But basically Bridge's $1.1B acquisition by Stripe should have been a red-alarm to existing companies that days of 1-Day ACH, interchange models, $25 international transfers, are soon to be over. This extends to many other adjacents from low fee disruptions like $HOOD, Mercury all the way to Stablecoin Neobanks, or companies making their own stablecoins like $SOFI. _ 5. AI Cloud Layers - The Solution to HyperScaler compute Bottleneck While Hyperscalers are stuck in 3-5 year grid interconnection queues, miners like WULF and IREN are sitting on plug-ready GWs today This is the opportunity of a lifetime as hyperscaler funnel their cash cow Cloud revenues down to tiny companies. There's many different layers to this from Fluidstack, Poolside, Fireworks on the GPU orchestration layer, to the bare metal layer that companies like IREN are building. Then there's becoming the hyperscaler themselves like NBIS owning the physical locations, the GPU, software orchestration, and then providing simple interfaces for inference. This is the opportunity for a few small companies to become Amazon Web Service or Microsoft Azure over the next year or two, or get acquired (eg. GOOGL buying Intersect for $4.7B) Neoclouds like NBIS, IREN, CRWV, down to colo plays like CIFR, WULF, HUT (and private sectors -> Energy) stand to benefit. _ 6. LLM Cybersecurity - The Evolution to Modern Security and Vulnerability Defense Recent reports (e.g., from Anthropic's Red Team) showed that advanced models like Opus (and future iterations) could autonomously scan open-source smart contracts and identify "Zero-Day" exploits worth millions of dollars in minutes. The Implication: If an AI can find a logic flaw in a immutable Blockchain contract, it can find a flaw in a bank's SWIFT API or a power grid's control software. Same with KYC/AML. Models like Gemini Nano Banana are able to create realistic images/videos of people and people are able to get past a lot of programs. There's tons of things as an unsexy alpha in this field like LLMs automating away SOC2/PCI dss compliance to agents sitting on a server, continuously monitor logs, and auto-generate the evidence needed for auditors. 7. LEO Space Infrastructure | The Evolution to Expanding into the final frontier. Space is the next big thing. This is not anything new. (hope you got the joke). But anywhere from companies like $RKLB, SpaceX. Companies that fix orbital congestion or launch cadence bottlenecks. To companies that commercialize the infrastructure like ASTS or Starlink present many opportunities over the next year. So companies like Impulse, Blue Origin, $ASOZF to RKLB, $ASTS stand to benefit across the entire chain. 8. Consumer Agentic Workflows (50 Step) - Disruption to the Consumer Workforce, from Manus, PATH Cognition This one is simple and needs no explanation. But largely obvious in potential impact on employment + cost saving. How do you automate away business development? How do you automate away marketing? How do you automate away software engineers? This is going past few step ChatGPT answers and directly in to the real world where an AI agent can roam X, find the right people, DM someone, continue conversations, and lead to a sales call in just one workflow. This is the end of the "Chatbot" era and the beginning of the "Action" era replacing everyone previously required in a company. I haven't quite seen this done at scale yet with any company. Public companies like META that own these, don't really present the best exposure. Maybe $PATH for public space. 9. Distributed Computing Latency - Fixing the Bottleneck for AI Compute Capacity Strains Hyperscalers like GOOGL Cloud, MSFT Azure at max capacity. Elon Musk already floated distributed computing as the future of solving this issue (eg. having networks of $TSLA's providing compute for LLMs for inference). The "Tesla Compute Cloud" thesis is fascinating, but the single biggest physical barrier I've identified is: Inference Latency. Too generate "Token B," the model must first finish generating "Token A." It cannot do both at the same time. If you split a massive model (like Grok-3) across 5 different cars to fit it in memory, you have to send data between those cars for every single token generated. So, if your network latency between cars is even 20ms (optimistic for 5G), and you are generating 50 tokens, you just added 1 full second of pure "waiting time" (latency) on top of the compute time. In a data center using NVLink, that wait time is measured in nanoseconds. Same applies to any spare computer, GPU, and others owned by retail users. And there's billions of consumer GPUs (Teslas, iPhones, Gaming PCs) that sit idle 90% of the time. Solving the "distributed latency" problem for inference presents one of the single greatest arbitrage opportunity in the history of computing. Haven't really seen any companies that accomplished this at scale yet. Maybe NVIDIA Dynamo, $AKAM, TSLA, getting a little closer. 10. Copper Interconnect Life Extension - Addressing the Bottlenecks of Nvidia and Others Since we can't have infinite InP, we have to engineer around it with what we have (eg. Copper), so copper cables can do things that physics said it shouldnt like carrying 224G signals across a rack without signal loss. The industry is hitting a hard stop on InP where, US cannot physically cannot mine and refine enough InP to turn every link in a data center into fiber optics. If anything helps, then it's good. EG. NVDA's $20B "Acqui-hire" of Groq's team and IP. LPU is more about inference latency/architecture but it addresses copper life extension as a byproduct. Groq’s entire architecture beat Nvidia on latency because it rejected optics. Groq uses a "deterministic" mesh that relies on direct electrical (copper) connections between chips, avoiding the "jitter" and conversion time of optical switches. Companies like $ALAB, $CRDO, Groq, or anyone who can find ways to engineer around the optical bottleneck with copper will be a winner. _ There are tons of trades from both private sector investments to public! Just wrote up my thoughts on the fly today, but happy to elaborate later. Regardless I believe a lot of these thematic investments from: Investing in InQ Bottleneck Workarounds ( $POET ) or the bottleneck itself ( $AXTI ) to Disruptors ( $CRCL ) in the public sector. To Investing in copper extension bottleneck fixes (Groq), bank charter disruptors (Mercury) to evolutionary companies (Lightmatter, Festo) in the private sector. Present asymmetrical upside in 2026. Happy New Year!
-
Benchmark 悄悄改口亚马逊芯片供应商，媒体误导严重。
SemiAnalysis 那就错了。Benchmark 最初声称：“我们现在有高度把握认为，该公司已将亚马逊的 Trainium3 和 4 设计输给了其台湾竞争对手 Alchip。”大量媒体对此进行了报道。但在股价下跌 14% 后，Benchmark 在无人关注时悄悄收回了该说法，基本改口称：亚马逊增加了 Alchip 作为第二设计支持供应商。媒体的框架设定糟透了。
英文原文
SemiAnalysis would be wrong then. Benchmark originally claimed: "We now have a high degree of conviction that the company has lost both Amazon's Trainium3 and 4 designs to its Taiwanese competitor, Alchip." Tons of media reported on that. But Benchmark silently walked that back after stock dropped 14%, when nobody was paying attention and basically said: Amazon added Alchip as second design support vendor. The media framing is atrocious.
-
询问SKC与三星电机在CPO玻璃基板领域的看法。
@jukan05 说到三星，我很好奇你对 SKC/Absolics 和三星电机在共封装光学(CPO) 玻璃基板领域的看法，因为它们都是韩国公司。
英文原文
@jukan05 Speaking of Samsung, I’m curious what your opinion is on SKC/Absolics and Samsung Electro-mechanics on the glass substrates space for CPO, since they’re both Korean companies
-
新云板块处于验证期，当前下跌是建立护城河前的黄金买入机会。
新云(Neoclouds)正处于“证明实力”阶段。 $NBIS 的年经常性收入(ARR)增长率高达700%+，达到70-90亿美元；$IREN 的数据也超过34亿美元。 许多公司拥有多年的收入可见性，并得到七大科技巨头(Mag7)超大规模云服务商的背书。 然而，市场表示： “这不是一个无限刷钱的漏洞”。 即股价上涨 -> 可转换票据/稀释 -> ARR增加 -> 循环往复。 它们现在都有资金，$NBIS 持有48亿美元以上的现金储备，$IREN 持有来自 $MSFT 预付款/票据的数十亿美元以完成建设并转化为自由现金流(FCF)。 尽管有《加速推进电气化法案(SPEED Act)》通过、OpenAI融资（降低交易对手风险）以及降息等顺风因素，但近期的下跌趋势似乎是年底税务收割、做空、主动式自动取款机(ATM)增发以及主要是： 等待证明这些公司能否从 $MSFT 与 IREN 的交易到 Nebius 的杠杆内部收益率(IRR)预测中，实现规模化的利润率。这似乎归结为执行力和等待下一次财报。 许多散户投资者在此期间似乎已经投降，但机构持股比例仅上升（$NBIS 从30%多升至50%以上）。 但非对称性就在这里： 如果 Nebius 管理层能实现20-30%的息税前利润(EBIT)利润率并达到70-90亿美元的 ARR 目标，那么在保守分析师给出200美元以上目标价(PT)的情况下，85美元的价格显得极其、极其低估。特别是考虑到 Clickhouse 的扩张/IPO 以及 Avride 与 Uber 合作的自动驾驶出租车规模化。 我在主题上也特别看好 $IREN 在微软交易上的杠杆 IRR 预测，以及来自 $CIFR / $WULF 与 $GOOGL 交易的新云数据中心(Colo)玩家。 简而言之：如果新云能在其预测的利润率、规模上执行到位，并在这2-3年的窗口期内建立自己的护城河，那么整个板块的抛售似乎是一个黄金机会。
英文原文
Neoclouds are in the "Prove It" phase. You have absurd 700%+ ARR growth rates to $7-9B from $NBIS and $3.4B+ figures from $IREN. Many have multiple year revenue visibility and backstop from Mag7 Hyperscalers. However, markets have said: "This is not an infinite money glitch". Where a stock goes up -> convertible note/dilution -> ARR increases -> repeat. They all have funding now, $NBIS sitting on a $4.8B+ cash stack, $IREN sitting on billions from $MSFT prepayment/notes to finish their buildout and turn that into FCF. Despite tailwinds from the initial SPEED act passing, OpenAI fundraising (for less counterpaty risk), and rate cuts, the recent downtrend seems to combine EoY tax harvesting with short selling, active ATMs, and mainly: Waiting for proof that these companies can deliver margins at scale from levered IRR projections on $MSFT's IREN deal to Nebius. It seems to comes down to execution and waiting for their next earnings report. Many retail investors seem to have capitulated during this time but institutional ownership has only gone up (30's from $NBIS to 50's+ now) But here's where the asymmetry comes in: If Nebius management scales to 20-30% EBIT margins with their $7-9B ARR target, this seems incredibly, incredibly off at $85 when conservative analysts are throwing out $200+ PTs. Especially considering possible Clickhouse ramp/IPO and Avride robotaxi scaling with Uber. I'm especially bullish thematically too with $IREN levered IRR projections on the Microsoft deal and Neocloud colo players from $CIFR / $WULF $GOOGL deals. TLDR: If Neoclouds can execute with their projected margins, scale, and create their own moats during this 2-3 year window, then the whole sector selloff seems like a golden opportunity.