Ace Informatic

$MU: Our financial outlook has strengthened since our last earnings call. We're on track for another substantial record free cash flow in fiscal Q3. We expect tightness for HBM, Dram, and Nand to continue well beyond calendar year 2026."

Leopold lost his aura That’s my honest opinion Got Iran-pilled and loaded up on semi puts before one of the biggest rallies in history Exited his huge $INTC call position at $44 before the $100+ rally - > Sold $HUT before it went up 105% - > Sold $COHR before it went up 52% - > Sold $CIFR before it went up 49% - > Sold $TSEM before it went up 41% - > Sold $LITE before it went up 35% - > Trimmed 40% of $BE before it went up 90% - > Trimmed $CORZ before it went up 59% - > Trimmed $SEI before it went up 40% Exited calls and missed a ton of upside Portfolio getting lower quality, in my opinion, and messier

Trade plan 5/18 $ASTS break over 85, 100c Jun 18--to target: $91.65 $INOD break over 96.7, 115c Jun 18--to target: $103 $OSS break over 17.03, 17.5c Jun 18--to target: $18 $DOCN break over 158.61, 175c Jun 18--to target: $164 $SATS break over 139, 150c Jun 18--to target: $145 Like❤️this if you want more trade plans!

@Remzztrades TTP MF!

Chat should we tell them?

Here are the robotics stocks grouped by segment Robot Automation $ROK – Rockwell Automation $ZBRA – Zebra Technologies $CGNX – Cognex $PATH – UiPath $PEGA – Pegasystems Medical Robotics $MDT – Medtronic $PRCT – PROCEPT BioRobotics $OMCL – Omnicell $SYK – Stryker $ISRG – Intuitive Surgical Industrial Robotics $TSLA – Tesla $HON – Honeywell $TER – Teradyne $LECO – Lincoln Electric Professional Robotics $OII – Oceaneering International $FARO – FARO Technologies Service Robotics $RR – Richtech Robotics Robot Software / AI $NVDA – NVIDIA $PTC – PTC Inc $PDYN – Palladyne AI $QCOM – Qualcomm Logistics Robotics $AMZN – Amazon $SYM – Symbotic $SERV – Servr Robotics $ATS – ATS Corp Defense Robotics $AVAV – AeroVironment $LMT – Lockheed Martin $BA – Boeing $TDY – Teledyne Technologies $TXT – Textron $ESLT – Elbit Systems $KTOS – Kratos Defense $NOC – Northrop Grumman $GD – General Dynamics $RTX – RTX Corp $LHX – L3Harris Technologies

@vulturetrades Trade

I’ve only seen this setup 1 time in 6 years 👀 I’m putting $250k in this SINGLE stock Very similar to $PENG that made millionaires $100 → $100,000 overnight $300 → $300,000 in a single day This is the third time. Comment “Trade” and I’ll send it. (Must be following to DM)

People are bidding $WYFI as a derivative play on the $CRBRS IPO. Seems the cheaper way to create the that exposure may be $BTBT. BTBT owns ~27M WYFI shares, worth ~$706M at today’s price, while BTBT’s entire market cap is only ~$643M. So the WYFI stake alone is worth more than the whole company, before giving any credit for ~$355M of ETH. Holdco dilution risk is certainly worth considering, but if comfortable holding or buying $WYFI (not saying I am), then holding $BTBT seems like the smarter play. @matthew_sigel am I missing anything? Of course, this is not financial advice.

$PENG 9/9 bullish confirms (weekly). Been doing some research... It's the company you call when you have a billion dollars worth of $NVDA GPUs sitting in boxes and you have no idea how to turn them into a working AI cluster... They design it, build it, deploy it, and manage it for you. End to end. That sounds boring until you realize the alternative is hiring a 50 person team for 18 months and probably still getting it wrong. Here's the actual moat... They have already deployed and managed over 85,000 GPUs in production environments. That is not a number you can buy or copy. It's institutional knowledge built over 25 years of HPC experience and every cluster they ship makes the next one easier. The customer list tells you everything. Shell runs HPC on Penguin infrastructure. Sandia National Labs deploys NextSilicon accelerators through them. Georgia Tech AI Makerspace was built by them. One of Korea's largest GPU clusters with over 1,000 Blackwell GPUs running in a single cluster is a Penguin build. They just got chosen by Deepgram, who picked them over everyone else to architect the production AI infrastructure behind their voice agents, sitting on Dell PowerEdge with Blackwell GPUs. The OriginAI platform scales from hundreds to 16,000 GPU clusters and integrates their ClusterWare software, which is the actual sticky part of the business. Once your cluster is running on their software, you don't rip it out. They're already in the memory squeeze. Memory chip prices doubled in Q1 and the company guided fiscal 2026 net sales up 12% with their integrated memory segment riding the same wave that's about to take $MU higher. Trading at roughly 14x forward earnings.

$PENG quotes and notes from the Needham Conference (I just watched the whole conference): - "There's a lot more demand for COUs and memory beyond just HBM. Memory is becoming a key requirement in scaling these workloads." - "Lots more need for memory in inference to make decisions. Having CXL KV Cache allows the LLM to respond much faster and easier." - "Large enterprises and governments are using inference and agentic AI on premise. We built one of the largest sovereign AI Factories in South Korea and working with other sovereign clients." - "Nobody else in the market has a product like MemoryAI which helps our customers and helps us win the deals." - The AI business is >$1B run rate and growing >50% - CEO ran Dell's AI HPC business and competed against Penguin Computing. The opportunity in memory and AI infrastructure got him quite excited about the opportunity. - "We have a deep relationship with SK Hynix. SK is an investor in the company and based on our relationships, we are able to have access." - "Very strong bookings and outlook for Advanced Compute. We are landing new logos and it takes longer to get to the state where we can recognize the revenue. Business is doing really well, and the challenge is time to book revenue is increasing." - Just a timing issue. - "Bookings are a leading indicator and are strong." - "AI adoption will significantly increase in the next 1 year. Some companies are embracing AI sooner than others, but for those adopters, the technology will show more returns."

A $2.0 billion company building the first commercial photonic memory product on the market in partnership with $MRVL and nobody is talking about it. $PENG Penguin Solutions. Trading at $40. 15x forward earnings. Zero photonics revenue priced in. Penguin was an early investor in Celestial AI. $MRVL acquired Celestial in February 2026 for billions. $PENG got a $27.5 million cash payout from that acquisition and kept the most important thing. The partnership. They are actively developing the Photonic Memory Appliance with $MRVL right now and the CEO confirmed it explicitly on the last earnings call. The PMA is the first commercial photonic memory disaggregation product being built for large scale AI environments. Photonic connectivity dramatically increases the capacity to share memory across workloads at speeds conventional interconnects cannot match. As inference scales this is a critical bottleneck solution. $PENG builds the box. $MRVL guided $500 million run rate by FY28 and $1 billion by FY29 on the Photonic Fabric platform. $PENG is in the critical path of that entire buildout. The market is pricing none of it. Then the $AMD partnership just dropped. $PENG and $AMD signed a collaboration with Shell to enhance AI data center efficiency. Shell’s Houston facility already runs 864 Penguin built servers loaded with dual 96 core EPYC processors. $AMD takes the chip revenue. $PENG takes the systems revenue. Every enterprise AI deployment on EPYC is a $PENG revenue event Full year revenue growth guidance just raised from 6% to 12% on the back of memory segment strength. Enterprise and sovereign AI pipeline growing 50%. 15x forward earnings. Photonics optionality at zero. $AMD CPU supercycle tailwind. $MRVL partnership confirmed. The re-rate will happen soon.

$NOK news this morning

TSMC 亚利桑那厂已经悄悄失败——400 亿美元烧完，良率不到日本厂的一半 主流叙事是"美国半导体回流成功了"。但真实数据完全相反——TSMC 亚利桑那 3 年投了 400 亿美元，2026 Q1 良率仅 52%，对比日本熊本厂 95%、台湾本土 98%，差距大到不能称作工厂。 问题 1：人才训练根本没完成。 TSMC 从台湾调了 1500 名工程师轮班支援，但当地招聘的 6000 名本地员工平均培训周期需要 4-6 年——美国大学根本没有"半导体产线工程师"这个专业方向。 问题 2：供应链没跟上。 高纯度光刻胶、特种气体、超净化学品——80% 在日本、20% 在台湾。亚利桑那厂每片晶圆的综合成本，是台南厂的 1.6 倍。 问题 3：客户开始用脚投票。 苹果、AMD 原本承诺把高端订单转到亚利桑那，2026 Q2 实际只下了 15% 的计划订单——其余全部还在台湾生产。"政治正确"敌不过经济账。 TSMC 内部怎么说？ 张忠谋早在 2023 年就警告："亚利桑那是个浪费。"现任 CEO 魏哲家最近第一次承认"美国厂进展不如预期"——这是台积电高管 30 年来最直白的自我否定。 半导体回流不是美国想做就能做的，这是 50 年产业积累的结果，不是 4000 亿补贴能买到的。

@StockMKTNewz SK and $PENG Leadership team making the partnership to tackle the memory wall together. Your not Bullish enough!

$PENG 我现在最关注的不是短期涨跌，而是 SK 体系后面会不会继续加深持股关系。 SK Telecom 已经对 Penguin 完成 2 亿美元战略投资，拿的是可转优先股。 随后 Penguin 又和 SK Telecom、SK hynix 签了 AI 数据中心合作协议，SK hynix 还会和 Penguin 旗下 SMART Modular 合作 AI 加速计算相关内存方案。 这就不是简单供应商关系了。 资本进来了，业务合作也进来了，AI 数据中心这条线也对上了。 如果未来 SK 体系继续增持 $PENG，或者进一步绑定双方合作，我觉得并不奇怪。 一个二十亿美元左右市值的小公司，站在 SK AI 基础设施生态里，这个故事还远没有讲完。 看涨 $PENG。

everybody is chasing down AI bottlenecks but what if it was as simple as unlocking more memory capacity per server? $PENG makes CXLs which, per node, unlocks 11TB of additional memory capacity per server. how significant is 11tb? a maxed out server generally has 2-3tb. OVER 4X THE MEMORY CAPACITY PER SERVER! this is inference’s saving grace penguin has the first production ready CXL KV Cache server ready to push inference with unmatched memory potential. im top blasting this one

$MRAM is one of the most interesting technology monopolies on the market. No position. Not planning one until the execution matches the monopoly. Everspin Technologies is the world’s only manufacturer of Toggle MRAM at scale. Memory that stores data using magnetic states instead of electric charge. Retains information when power goes out. Immune to radiation. Works in extreme temperatures. The US military runs it in missiles, submarines, aerospace systems, and classified platforms. No substitute exists. No competitor operates at this level. The defense contract they just signed has the most interesting clause in any contract. If Everspin ever decides to stop making MRAM the US government requires them to hand over all process technology, IP, and manufacturing know-how to ensure continuity of domestic supply. The Pentagon literally built a legal mechanism to keep this technology alive because they cannot afford for it to disappear. That is not a customer. That is a national security dependency. $40 million contract with the US Department of War just signed. New foundry partnership with Microchip for domestic production in Oregon. The US defense industrial base is onshoring MRAM supply and Everspin is the only game in town. Now the problem. The execution has been frustrating. Revenue growth has been inconsistent. Margins have been all over the place. For a company with this level of monopoly positioning the financial results should look much better than they do. The technology and the contracts are real. The execution is not where it needs to be. $MRAM

@TrinityTrading_ TTP MF! $MRAM 🐏 $MU $SNDK

$DELL new highs as Trump positive on company says "Go and Buy a Dell"

In the last month and a half: $CRWV: • Multi-year and multi-gigawatt deals with META and Anthropic $NBIS: • $27 billion expanded deal with META $CIFR: • Signed third hyperscaler lease agreement $HUT: • New 15yr/$9.8 billion lease with hyperscaler tenant $IREN: • Energized 1.4 GW & acquired Mirantis The race for compute is accelerating

⚡ Nvidia and Corning ink major optical fiber deal $GLW shares jump; deal targets AI data center scale. Could reshape interconnect costs industry-wide.