# ![@valuechainkoala Avatar](https://lunarcrush.com/gi/w:26/cr:twitter::1938970612732698629.png) @valuechainkoala Value Chain Koala

Value Chain Koala posts on X about ai, $lrcx, dram, $amat the most. They currently have [--] followers and [--] posts still getting attention that total [-----] engagements in the last [--] hours.

### Engagements: [-----] [#](/creator/twitter::1938970612732698629/interactions)
![Engagements Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::1938970612732698629/c:line/m:interactions.svg)

- [--] Week [------] +74%
- [--] Month [------] +43,651%

### Mentions: [--] [#](/creator/twitter::1938970612732698629/posts_active)
![Mentions Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::1938970612732698629/c:line/m:posts_active.svg)

- [--] Week [--] +34%
- [--] Month [--] +4,400%

### Followers: [--] [#](/creator/twitter::1938970612732698629/followers)
![Followers Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::1938970612732698629/c:line/m:followers.svg)

- [--] Week [--] +74%
- [--] Month [--] +5,800%

### CreatorRank: [-------] [#](/creator/twitter::1938970612732698629/influencer_rank)
![CreatorRank Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::1938970612732698629/c:line/m:influencer_rank.svg)

### Social Influence

**Social category influence**
[stocks](/list/stocks)  57.61% [technology brands](/list/technology-brands)  32.61% [finance](/list/finance)  23.91% [cryptocurrencies](/list/cryptocurrencies)  15.22% [currencies](/list/currencies)  11.96% [automotive brands](/list/automotive-brands)  4.35% [countries](/list/countries)  3.26% [fashion brands](/list/fashion-brands)  1.09%

**Social topic influence**
[ai](/topic/ai) 58.7%, [$lrcx](/topic/$lrcx) 11.96%, [dram](/topic/dram) #112, [$amat](/topic/$amat) #53, [$mu](/topic/$mu) 9.78%, [samsung](/topic/samsung) 8.7%, [$be](/topic/$be) #13, [$nvda](/topic/$nvda) 7.61%, [$lite](/topic/$lite) 5.43%, [trend](/topic/trend) 5.43%

**Top accounts mentioned or mentioned by**
[@geoffreykarren](/creator/undefined) [@realjgbanks](/creator/undefined) [@anandchokshi19](/creator/undefined) [@jukan05](/creator/undefined) [@arkkdaily](/creator/undefined) [@curiousfounder](/creator/undefined) [@optionscjp](/creator/undefined) [@thexcapitalist](/creator/undefined) [@deitaone](/creator/undefined) [@kresooptions](/creator/undefined) [@ypiotr](/creator/undefined) [@yntelligence](/creator/undefined) [@saaspocalypse](/creator/undefined)

**Top assets mentioned**
[Lam Research Corporation (LRCX)](/topic/$lrcx) [Applied Materials, Inc. (AMAT)](/topic/$amat) [Micron Technology, Inc. (MU)](/topic/$mu) [Bloom Energy Corporation (BE)](/topic/$be) [NVIDIA Corp. (NVDA)](/topic/$nvda) [Lumentum Holdings Inc. Common Stock (LITE)](/topic/$lite) [Astera Labs, Inc. (ALAB)](/topic/$alab) [Broadcom, Inc. (AVGO)](/topic/$avgo) [KLA Corp (KLAC)](/topic/$klac) [Advanced Micro Devices (AMD)](/topic/$amd) [Coherent Inc (COHR)](/topic/$cohr) [Centrus Energy Corp. (LEU)](/topic/$leu) [Bitcoin (BTC)](/topic/bitcoin) [Rocket Lab USA, Inc. Common Stock (RKLB)](/topic/$rklb) [Microsoft Corp. (MSFT)](/topic/microsoft) [Tesla, Inc. (TSLA)](/topic/$tsla) [Axon Enterprise, Inc. (AXON)](/topic/$axon) [Valero Energy Corporation (VLO)](/topic/$vlo) [Kratos Defense & Security Solutions, Inc. (KTOS)](/topic/$ktos) [Palantir Technologies Inc. (PLTR)](/topic/$pltr) [Iris Energy Limited (IREN)](/topic/$iren) [Raytheon Technologies Corp (RTX)](/topic/$rtx) [Oklo Inc. (OKLO)](/topic/$oklo) [Duke Energy Corp (DUK)](/topic/$duk) [NextEra Energy Inc (NEE)](/topic/$nee) [GE Vernova Inc. (GEV)](/topic/$gev) [NuScale Power Corporation (SMR)](/topic/$smr) [Modine Manufacturing Co (MOD)](/topic/$mod) [Adobe, Inc. (ADBE)](/topic/$adbe) [Salesforce Inc (CRM)](/topic/$crm) [Oracle Corporation (ORCL)](/topic/$orcl) [Super Micro Computer, Inc. Common Stock (SMCI)](/topic/$smci) [Arista Networks Inc (ANET)](/topic/$anet) [Cisco Systems Inc (CSCO)](/topic/$csco) [Corning Incorporated (GLW)](/topic/$glw) [Dell Technologies, Inc. (DELL)](/topic/dell) [Western Digital Corp (WDC)](/topic/$wdc) [Stacks (STX)](/topic/$stx) [Energy Fuels Inc. (UUUU)](/topic/$uuuu) [Comfort Systems USA, Inc. (FIX)](/topic/$fix) [Amazon.com, Inc. (AMZN)](/topic/$amzn) [Taiwan Semiconductor (TSM)](/topic/$tsm) [Alphabet Inc Class A (GOOGL)](/topic/$goog) [MP Materials Corp. (MP)](/topic/$mp) [Limitus (LMT)](/topic/$lmt) [General Dynamics Corporation (GD)](/topic/$gd) [Northrop Grumman Corporation (NOC)](/topic/$noc) [American Electric Power Company Inc (AEP)](/topic/$aep) [Cummins Inc. (CMI)](/topic/$cmi) [Plug Power, Inc. (PLUG)](/topic/$plug) [Heroes of NFT (HON)](/topic/$hon) [Fluence Energy, Inc. (FLNC)](/topic/$flnc)
### Top Social Posts
Top posts by engagements in the last [--] hours

"The [--] Strongest Power Industries for AI Data Centers (Beyond the Grid) Co-locate with power plants $DUK $NEE $D $AEP On-site gas or diesel microgrids $GEV $SIEGY $CMI Fuel Cells and CHP $BE $PLUG $HON Renewables plus storage $FLNC $FSLR $ENPH $TSLA SMRs and nuclear microgrids $SMR $GE $OKLO $WEC Behind the meter hookups $PWR $MTZ $ABB $ET Whats the best AI power trade in this stack https://twitter.com/i/web/status/2015484814208729356 https://twitter.com/i/web/status/2015484814208729356"  
[X Link](https://x.com/valuechainkoala/status/2015484814208729356)  2026-01-25T18:00Z [--] followers, [---] engagements


"$LITE and $COHR are sitting on the same structural connectivity wall just from the optics side. Lumentum and Coherent supply the highspeed optical components and modules (400G/800G/1.6T and beyond) that let AI clusters actually move data between GPUs/TPUs and between data centers at scale. As networks become the hard bottleneck every step up in lane speed and density pushes more value into the photonics layer which is exactly where LITE and COHR live. https://twitter.com/i/web/status/2021607826569789764 https://twitter.com/i/web/status/2021607826569789764"  
[X Link](https://x.com/valuechainkoala/status/2021607826569789764)  2026-02-11T15:30Z [--] followers, [---] engagements


"$AMAT: The HBM Tax Every AI Chip Has to Pay Chokepoint / valuechain role: Upstream equipment supplier for DRAM/HBM and advanced packaging sitting inside the memory/HBM CAPEX bottleneck. Trend: AI boom and a multiyear HBM/DRAM supercycle driving massive sustained fab and packaging investment. Strong catalysts shaping the story: [----] Signals WFE inflection with tools growing faster than overall semis reinforces $AMAT as longterm valuecapture on AI/memory capex. [----] Calls itself leading DRAM equipment supplier with outsized share anchors $AMAT at the core of memory scaling. [----] Says its tools"  
[X Link](https://x.com/valuechainkoala/status/2019463531968844286)  2026-02-05T17:30Z [--] followers, [--] engagements


"$MOD (Modine) turning AI heat into a cooling bottleneck play Role (bottleneck angle): Midstream data center cooling supplier (systems + components). Airedale pushes $MOD closer to the core of liquid and immersion cooling as racks get hotter. Trend: AI data centers are forcing a capex wave into advanced cooling (air liquid immersion). Catalyst chain: 2024: Buys TMG Core immersion cooling IP/assets adds real immersion tech (air + liquid + immersion stack). 2024: Data center sales +69% YoY to $294M + ramps capacity with [--] dedicated facilities proves demand is already scaling. Feb 2025: $180M new"  
[X Link](https://x.com/valuechainkoala/status/2022362635958665580)  2026-02-13T17:30Z [--] followers, [--] engagements


"No etch no HBM. No HBM no AI. $LRCX is the hidden choke point. Bottleneck/value chain role: Critical midstream supplier of etch and deposition tools for advanced chips embedded across leading fabs but without ASMLlike monopoly power. Trend: Pureplay leverage on the AI/HBM capex wave as rising process complexity drives heavier dependence on its etch/deposition tools. Strong catalysts driving the story: 2024: Management flags AI and HPC as key drivers of advancednode etch/ALD demand framing LRCX as core AI infrastructure plumbing. 2025: Labeled a highconviction buy in semi equipment on"  
[X Link](https://x.com/valuechainkoala/status/2020188308689531250)  2026-02-07T17:30Z [--] followers, [---] engagements


"Were already seeing hyperscalers quietly testing more than just GPUs custom AI ASICs like Googles TPUs Amazons Trainium and Metas MTIA are all about wiring hundreds or even thousands of chips into one tightly coupled cluster. The whole advantage of something like a TPU superpod is how densely it connects chips together with massive highspeed I/O and optical fabrics linking thousands of accelerators in a single logical system. Thats exactly the kind of architecture that pulls the bottleneck away from pure compute and toward optical connectivity where 800G/1.6T links 224G+ SerDes and ultradense"  
[X Link](https://x.com/valuechainkoala/status/2022346097847304396)  2026-02-13T16:24Z [--] followers, [---] engagements


"Feels like a lot of these Korea flows are really a leveraged bet on the AI memory supercycle. With Samsung and SK Hynix dominating HBM and advanced DRAM foreign inflows into Korea are basically a macro way to express conviction that AI capex will keep tightening the highbandwidth memory market over the next [--] years. https://twitter.com/i/web/status/2022359065930674191 https://twitter.com/i/web/status/2022359065930674191"  
[X Link](https://x.com/valuechainkoala/status/2022359065930674191)  2026-02-13T17:15Z [--] followers, [---] engagements


"All of these names are in an AI infrastructure supercycle where the real signal is capex not a choppy [----] price chart. The hyperscalers are guiding to roughly [------] billion dollars of [----] capital spend with the majority tied directly to AI data centers chips power and networks. They are effectively building nationalscale infrastructure while the market marks them down on nearterm margin pressure and AI fatigue. Looking at yeartodate performance (all negative) misses the point of a capex cycle that will take years and trillions of dollars to fully show up in earnings and cash flows."  
[X Link](https://x.com/valuechainkoala/status/2022712253103898660)  2026-02-14T16:39Z [--] followers, [---] engagements


"$AMAT is doing exactly what youd want in this tape: guiding to 20%+ semiconductor equipment growth in [----] off AI foundry/logic HBM and advanced packaging while already up 38% yeartodate with the rest of the infra winners. HBM4 is the engine behind that 16high AIclass stacks are dramatically more complex than legacy DRAM which means more process steps more tools per bit and a multiyear WFE tailwind for the companies that sit in the middle of those flows. https://twitter.com/i/web/status/2022713521788612947 https://twitter.com/i/web/status/2022713521788612947"  
[X Link](https://x.com/valuechainkoala/status/2022713521788612947)  2026-02-14T16:44Z [--] followers, [---] engagements


"@RealJGBanks PRY is an under the radar way to play this too Prysmian builds the high voltage cables and grid infrastructure that all this AI power demand will need"  
[X Link](https://x.com/valuechainkoala/status/2016531573345010060)  2026-01-28T15:19Z [--] followers, [----] engagements


"$ALAB is basically sitting on a key choke point in this list Astera Labs builds the highspeed connectivity and CXL/PCIe plumbing that lets GPUs memory and storage actually talk to each other at rack scale so as AI clusters get bigger its linklevel bottleneck becomes more valuable than any single app on top. https://twitter.com/i/web/status/2017251236362236061 https://twitter.com/i/web/status/2017251236362236061"  
[X Link](https://x.com/valuechainkoala/status/2017251236362236061)  2026-01-30T14:59Z [--] followers, [---] engagements


"In a world of agentic AI a lot of these apps are basically exposed: if a single AI layer can read/write email files CRM billing and support across tools it doesnt care whether its using $MNDY $ASAN $HUBS or $Brze it just optimizes for speed price and API access which turns many vertical SaaS products into interchangeable plumbing rather than destinations. https://twitter.com/i/web/status/2017256767210578215 https://twitter.com/i/web/status/2017256767210578215"  
[X Link](https://x.com/valuechainkoala/status/2017256767210578215)  2026-01-30T15:21Z [--] followers, [---] engagements


"Applied Materials $AMAT is the definition of a picksandshovels name here its equipment is needed to manufacture the advanced memory $MU logic and networking chips $NVDA $AVGO $ALAB that power AI infrastructure. When those downstream names enter an AI upcycle and foundries boost capex to meet demand $AMAT often feels the positive impact first long before the economics fully show up in the end applications. https://twitter.com/i/web/status/2017315463487750225 https://twitter.com/i/web/status/2017315463487750225"  
[X Link](https://x.com/valuechainkoala/status/2017315463487750225)  2026-01-30T19:14Z [--] followers, [----] engagements


"@anandchokshi19 $ALAB Astera Labs is basically the signal highway inside the $NVDA stack: it solves the PCIe/CXL connectivity bottleneck between GPUs CPUs and memory so that all the AI compute Nvidia sells can actually be fed with data at full speed"  
[X Link](https://x.com/valuechainkoala/status/2017616032253300870)  2026-01-31T15:08Z [--] followers, [----] engagements


"$PRY is an underrated way to play the same theme from a different angle if copper is the bottleneck in the ground Prysmian sits on the bottleneck in the grid. As more capex chases new generation and transmission you need miles of highvoltage cable and interconnects to actually move that copperpowered electrons to data centers and cities and thats exactly where $PRY lives. https://twitter.com/i/web/status/2017987226982842793 https://twitter.com/i/web/status/2017987226982842793"  
[X Link](https://x.com/valuechainkoala/status/2017987226982842793)  2026-02-01T15:43Z [--] followers, [---] engagements


"One undertheradar name on this list is $LRCX Lam is basically part of the HBM manufacturing license. Without its etch and deposition tools qualified for advanced DRAM stacks Micron and the other memory players simply cant build highyield HBM at scale so as long as AI keeps pulling more HBM capacity Lam sits in a structurally critical spot in the supply chain. https://twitter.com/i/web/status/2018341495745413187 https://twitter.com/i/web/status/2018341495745413187"  
[X Link](https://x.com/valuechainkoala/status/2018341495745413187)  2026-02-02T15:11Z [--] followers, [---] engagements


"Nice overview but its worth remembering that the real leverage in this memory cycle isnt just $MU / $SNDK / Samsung / Hynix its also the WFE guys selling them the picks and shovels. $LRCX $KLAC and $AMAT provide the etch deposition and processcontrol tools you literally need to pattern stack and qualify HBM/DRAM dies; without that kit running 24/7 at higher and higher process intensity none of those record EPS numbers in memory ever show up. https://twitter.com/i/web/status/2018342612088791191 https://twitter.com/i/web/status/2018342612088791191"  
[X Link](https://x.com/valuechainkoala/status/2018342612088791191)  2026-02-02T15:15Z [--] followers, [----] engagements


"For me its the other way around: I start with understanding the businesses then I layer the technicals on top. If I dont know where a company sits in the value chain how it makes money and whether it has a real moat the [--] or 200week moving average is just noise. So my process is: first deeply understand the stocks Im investing in industry bottlenecks durability of the profit pool and only then use the technical zones to help with timing and risk management https://twitter.com/i/web/status/2018704233705464226 https://twitter.com/i/web/status/2018704233705464226"  
[X Link](https://x.com/valuechainkoala/status/2018704233705464226)  2026-02-03T15:12Z [--] followers, [----] engagements


"$CRCL is basically the onramp for institutional money into crypto. When funds fintechs or platforms want dollarlike exposure onchain without touching volatile tokens directly they use USDC as the bridge. On top of that USDC is now getting a fresh tailwind from Polymarket: all bets there settle in USDC and rising volumes are expected to add billions to USDCs market cap over the next couple of years which directly feeds into Circles fee and float economics. https://twitter.com/i/web/status/2019081265396240744 https://twitter.com/i/web/status/2019081265396240744"  
[X Link](https://x.com/valuechainkoala/status/2019081265396240744)  2026-02-04T16:11Z [--] followers, [---] engagements


"nteresting allocation but Im struggling to see the diversification here its essentially a concentrated space + RF bet with a big macro call on one theme. Two questions: Why size so heavily into space instead of spreading across other AI/infra beneficiaries (semis memory power defense etc.) Whats the game plan if/when flows rotate out of space do you have defined levels time horizon or alternative themes youre ready to recycle that capital into https://twitter.com/i/web/status/2020181790036553957 https://twitter.com/i/web/status/2020181790036553957"  
[X Link](https://x.com/valuechainkoala/status/2020181790036553957)  2026-02-07T17:04Z [--] followers, [---] engagements


"$BE is basically a timetopower arbitrage on the grid bottleneck. Big data centers can wait [---] years for new grid connections and transmission upgrades which completely breaks the AI buildout timeline. Blooms onsite fuel cells by contrast can be sited and turned on in months not years giving operators a fast modular way to add real megawatts where the grid is already tapped out. https://twitter.com/i/web/status/2020182968849850381 https://twitter.com/i/web/status/2020182968849850381"  
[X Link](https://x.com/valuechainkoala/status/2020182968849850381)  2026-02-07T17:08Z [--] followers, [----] engagements


"From this list the name that looks best positioned as a true brain in the stack is $AXON and the one sitting at the opposite end of that spectrum is $ADBE. Axon is building an endtoend system: sensors (TASERs body cams drones) cloud and AI on top of proprietary highfriction data. That combo of missioncritical workflows recurring SaaS and brutal switching costs for government customers gives it a real moat that AI actually strengthens rather than erodes. Adobe by contrast feels like the inverse setup. It still prints strong numbers but its moat is directly in the blast radius of AInative"  
[X Link](https://x.com/valuechainkoala/status/2020514647045025833)  2026-02-08T15:06Z [--] followers, [---] engagements


"$ALAB Scorpio is interesting here because it attacks the design bottleneck in memory not the fab CAPEX problem. By using AIdriven EDA and advanced chiplet/3D architectures Scorpio can help memory players squeeze more bandwidth density and performance per bit out of each insanely expensive fab line. In a world where only three giants can afford to build DRAM fabs anything that accelerates design cycles and improves memorycompute integration (HBM nearmemory inpackage) effectively creates extra supply from the same manufacturing choke point. https://twitter.com/i/web/status/2020520022049788165"  
[X Link](https://x.com/valuechainkoala/status/2020520022049788165)  2026-02-08T15:28Z [--] followers, [--] engagements


"The equipment side actually looks like one of the cleanest ways to play this HBM arms race. Every new HBM fab and every node jump (HBM3 HBM3E HBM4) drives more spend on etch deposition inspection and advanced packaging exactly where $AMAT $LRCX and $KLAC sit. They sell picks and shovels to Samsung SK hynix $MU and now CXMT/Huawei so they monetize capacity additions on both sides of the KoreaChina race while being less exposed to DRAM price swings and more tied to multiyear CAPEX budgets. https://twitter.com/i/web/status/2020523751385727289 https://twitter.com/i/web/status/2020523751385727289"  
[X Link](https://x.com/valuechainkoala/status/2020523751385727289)  2026-02-08T15:42Z [--] followers, [----] engagements


"$PRY is exactly the kind of name people roll their eyes at today (its just cables) that could be a quiet monster in [--] years. AI data centers EVs and industrial robotics all bottleneck on one thing: upgraded highercapacity power and connectivity infrastructure. Prysmian sits right in that choke point designing and installing the highvoltage grid links submarine interconnectors and data center cabling that physically move electrons and bits for these trends. In other words if AI EVs and robots are the apps $PRY owns a big piece of the plumbing they all need to keep scaling."  
[X Link](https://x.com/valuechainkoala/status/2020554178733425029)  2026-02-08T17:43Z [--] followers, [----] engagements


"$LRCX is one of the quiet winners in this whole HBM4 story. If Nvidia is standardizing Vera Rubin around HBM4 from Samsung and SK hynix that means even more wafers running through ultracomplex DRAM + 3Dstacked packaging flows exactly where Lams etch and deposition tools are missioncritical. Every incremental jump in HBM bandwidth and layer count makes the process harder not easier and forces Samsung/ Hynix to lean even more on bestinclass process equipment. So while the market is focused on who won the HBM4 socket $LRCX gets to sell the shovels into both Korean winners as they ramp capacity"  
[X Link](https://x.com/valuechainkoala/status/2020848120049852813)  2026-02-09T13:11Z [--] followers, [---] engagements


"$LRCX is one of the quiet winners in this whole HBM4 story. If Nvidia is standardizing Vera Rubin around HBM4 from Samsung and SK hynix that means even more wafers running through ultracomplex DRAM + 3Dstacked packaging flows exactly where Lams etch and deposition tools are missioncritical. Every incremental jump in HBM bandwidth and layer count makes the process harder not easier and forces Samsung/ Hynix to lean even more on bestinclass process equipment. So while the market is focused on who won the HBM4 socket $LRCX gets to sell the shovels into both Korean winners as they ramp capacity"  
[X Link](https://x.com/valuechainkoala/status/2020849688467906762)  2026-02-09T13:18Z [--] followers, [----] engagements


"Im much more cautious on $CRM it looks structurally exposed to AI rather than just oversold. Salesforce was already in a highly competitive market and AI is now eroding its moats around workflow complexity and lockin. If agents can automate logging routing and orchestration across tools the premium for a heavyweight $CRM suite shrinks and cheaper focused rivals can compete for the same budgets. https://twitter.com/i/web/status/2020870927181828236 https://twitter.com/i/web/status/2020870927181828236"  
[X Link](https://x.com/valuechainkoala/status/2020870927181828236)  2026-02-09T14:42Z [--] followers, [---] engagements


"$BE is literally doing this for power. Instead of waiting 35+ years for a new highvoltage grid connection Blooms fuelcell systems let data centers drop modular generation right onsite and turn gas into electricity within months not years. In an AI world where compute is scaling faster than the grid that ability to bypass interconnection queues and still get reliable lowcarbon megawatts is exactly the kind of bottleneck solution your thread is talking about. https://twitter.com/i/web/status/2020871729984192656 https://twitter.com/i/web/status/2020871729984192656"  
[X Link](https://x.com/valuechainkoala/status/2020871729984192656)  2026-02-09T14:45Z [--] followers, [---] engagements


"Memory equipment makers $LRCX $AMAT $KLAC are one layer upstream from the DRAM vendors so they ride the same capex wave with less exposure to pricing cycles. When Samsung SK hynix or $MU pour billions into HBM and DRAM capacity to serve hyperscalers a big chunk of that spend goes straight into deposition etch lithography metrology and advanced packaging tools. Those tools are sold by the memory equipment vendors. As node complexity rises and HBM stacks get taller each incremental bit of capacity requires more (and more sophisticated) equipment so tool makers capture a leveraged multiyear"  
[X Link](https://x.com/valuechainkoala/status/2021230112935415870)  2026-02-10T14:29Z [--] followers, [---] engagements


"One thing that rarely shows up in these upside lists is how differently these names are financed and $ORCL stands out there in a big way. Oracle is running with a debttoequity ratio thats several turns above [--] (think roughly 3.55.5x in recent years versus well under 1x for many large software peers) meaning its balance sheet is heavily funded by debt rather than equity. That leverage can juice EPS and buybacks on the way up but it also makes Oracle structurally riskier if rates stay higher for longer or if AI/cloud capex ever needs to be cut back because interest costs and refinancing risk"  
[X Link](https://x.com/valuechainkoala/status/2021231063469523074)  2026-02-10T14:33Z [--] followers, [---] engagements


"$LITE is basically part of the circulatory system of the AI data center stack. Lumentums InP lasers and highspeed optical transceivers are what turn electrical signals inside $NVDA / $AMD servers into light that can move 400G 800G 1.6T across racks and between clusters. As AI fabrics scale out copper hits reach and power limits so hyperscalers lean on players like $LITE for the EML lasers pluggables and emerging CPO/OCS systems that make dense lowlatency GPU clusters actually talk to each other at scale. https://twitter.com/i/web/status/2021233752395923726"  
[X Link](https://x.com/valuechainkoala/status/2021233752395923726)  2026-02-10T14:44Z [--] followers, [---] engagements


"Most of that CapEx is flowing into a few infrastructure layers: AI compute and servers (GPUs/ASICs accelerators racks OEMs like $NVDA $AMD $AVGO $SMCI $DELL). Memory and storage (HBM/DRAM NAND HDD/flash systems $MU $WDC $STX plus array vendors). Networking and optics (switches NICs photonics fiber $ANET $CSCO $LITE $COHR $GLW and peers). Power and cooling (grid gear UPS liquid cooling onsite generation $VRT $ETN $BE traditional utilities/energy). Data center real estate and construction (land shells fitout generators physical security via REITs and specialist builders)"  
[X Link](https://x.com/valuechainkoala/status/2021234757380501840)  2026-02-10T14:48Z [--] followers, [---] engagements


"Most of that CapEx is flowing into a few infrastructure layers: AI compute and servers (GPUs/ASICs accelerators racks OEMs like $NVDA $AMD $AVGO $SMCI $DELL). Memory and storage (HBM/DRAM NAND HDD/flash systems $MU $WDC $STX plus array vendors). Networking and optics (switches NICs photonics fiber $ANET $CSCO $LITE $COHR $GLW and peers). Power and cooling (grid gear UPS liquid cooling onsite generation $VRT $ETN $BE traditional utilities/energy). Data center real estate and construction (land shells fitout generators physical security via REITs and specialist builders)."  
[X Link](https://x.com/valuechainkoala/status/2021240986769207322)  2026-02-10T15:12Z [--] followers, [---] engagements


"@geoffreykarren @jukan05 $LITE $COHR"  
[X Link](https://x.com/valuechainkoala/status/2021494927587524843)  2026-02-11T08:02Z [--] followers, [---] engagements


"$LEU is doing a lot more than just another uranium stock its a strategic energysecurity asset. Its the only Western player that can currently enrich uranium to HALEU levels which is a hard requirement for most nextgen SMRs. Without a reliable HALEU supply chain the whole SMRs as a gridscale solution to the AI/datacenter power bottleneck story stalls before it starts. https://twitter.com/i/web/status/2021603705045299461 https://twitter.com/i/web/status/2021603705045299461"  
[X Link](https://x.com/valuechainkoala/status/2021603705045299461)  2026-02-11T15:14Z [--] followers, [---] engagements


"If I had to express this in positions Id be buying across the chokepoints of the AI stack not just the headline GPUs. $BE as the offmeter power plant for AI: Blooms solidoxide fuel cells let data centers bypass slow congested grid interconnects and stand up dedicated highdensity power in months instead of years which is exactly where the AI buildout is hitting a wall. In a world where timetocompute is capped by timetopower onsite generation becomes the real bottleneck breaker. $LRCX as the HBM4 enabler: the shift to 16high HBM4 stacks and AIdriven memory capex means more etch deposition and"  
[X Link](https://x.com/valuechainkoala/status/2022711441858478124)  2026-02-14T16:36Z [--] followers, [---] engagements


"Valuations matter but AI isnt just hype sitting on top of a Shiller chart its already showing up as real revenue and earnings growth for the biggest index weights. A high Shiller P/E on backwardlooking earnings can look scary but if AI structurally raises the earnings power of these businesses over the next decade todays multiple can compress without a 40% crash. The key question isnt only P/E = [--] therefore bubble but how much of this multiple is backed by durable AIdriven cashflow growth https://twitter.com/i/web/status/2022716061905502562 https://twitter.com/i/web/status/2022716061905502562"  
[X Link](https://x.com/valuechainkoala/status/2022716061905502562)  2026-02-14T16:54Z [--] followers, [---] engagements


"$SOFI sits in one of the most profitable corners of finance: a depositfunded highyield consumer bank wrapped in a digital platform. As it scales its own balance sheet and crosssells multiple products into the same young growing member base each incremental customer becomes more profitable. Thats why the pivot to sustained profitability with softwarelike margins matters: once the fixed regulatory and tech costs are in place additional volume drops through at attractive returns on capital. In that sense SoFi isnt just fintech its a vertically integrated capitalefficient franchise in one of the"  
[X Link](https://x.com/valuechainkoala/status/2022717967381742026)  2026-02-14T17:01Z [--] followers, [----] engagements


"$Axon. Their moat isnt nice UX SaaS its owning the whole regulated workflow: body cams + tasers + sensors feeding straight into Axon Cloud evidence + chainofcustody locked in for years and multiyear contracts baked into police procedures. AI just makes that stack more powerful (search transcription autoreports) it doesnt replace it. https://twitter.com/i/web/status/2022721778993246705 https://twitter.com/i/web/status/2022721778993246705"  
[X Link](https://x.com/valuechainkoala/status/2022721778993246705)  2026-02-14T17:17Z [--] followers, [---] engagements


"Nuclear value chain for the AI power era. Mining + conversion (UF6) $CCJ $UUUU Enrichment (Bottleneck #1) $LEU $URA Fuel fabrication (Bottleneck #2) $BWXT $SMR $OKLO Reactor design + licensing $SMR $GEV Operators (O&M) $DUK $SO $NEE $EXC"  
[X Link](https://x.com/valuechainkoala/status/2014390103972733211)  2026-01-22T17:30Z [--] followers, [---] engagements


"$LFST turning the mentalhealth tech boom into actual billable visits Behavioral health demand is exploding. The chart says it: +43.7% since [----]. Apps can create intent but clinicians convert it into reimbursed care. LFST sits on that choke point with about 7.4k clinicians and about 7.9M yearly visits. Catalyst chain: [----] Revenue $1.25B vs. $1.06B in [----] (19% YoY): the theme is already a billiondollar P&L line not just a slide in a deck. [----] Clinician base 7.4k+: LFST is scaling the real constraint in mental health licensed provider capacity not just more apps. [----] Q2  1.7M visits in a"  
[X Link](https://x.com/valuechainkoala/status/2017651592103641549)  2026-01-31T17:30Z [--] followers, [--] engagements


"Feels like $AXON belongs on the AI winners list too. Their moat isnt just hardware its the ecosystem: Lockedin devices (TASERs body cams incar video) deployed across entire agencies. Axon Evidence a sticky AIpowered cloud platform that stores and manages petabytes of digital evidence with huge switching costs and compliance/regulatory hooks. AI for Axon isnt a side quest it makes the core product better (autotagging transcription report drafting search) on top of data only they have which is exactly the kind of structural advantage you want in an AI software winner."  
[X Link](https://x.com/valuechainkoala/status/2020847296213045455)  2026-02-09T13:08Z [--] followers, [---] engagements


"In my early years in the market I used to look for opportunities inside the chart. Another high another breakout another pattern. I felt that if I could just spot the right shape Id spot the right stock. Slowly I realized something uncomfortable: the chart is only a summary of what already happened. Everyone who has already done a lesson a year or a decade on the stock is baked into that line. It doesnt tell me why there is an opportunity there only how it looked in hindsight. The first time I broke down an industry not a stock something changed. Instead of asking Does the chart look good I"  
[X Link](https://x.com/valuechainkoala/status/2020913084072644747)  2026-02-09T17:30Z [--] followers, [--] engagements


"I turn messy industries into a clean structure. Every week: the mega catalyst shaping an industry the value chain: who gets paid who gets squeezed the bottleneck: where pricing power concentrates the tickers and catalysts: the practical takeaway Research you can actually use. Full research: link in bio"  
[X Link](https://x.com/valuechainkoala/status/2021171022276288517)  2026-02-10T10:34Z [--] followers, [--] engagements


"Id actually flip the $ADBE take on its head here to me its one of the names most at risk from AI not the safest way to play it. Adobes moat was always about distribution + lockin + workflow complexity. Generative AI is attacking all three at once: cheaper simpler AInative tools are compressing pricing power and core use cases (image video design) are being automated away at the edges. So while the current numbers still look fine I see ADBE as in a structurally dangerous spot where AI is eroding the moat in big steps not strengthening it which makes that 45% undervalued AI winner narrative"  
[X Link](https://x.com/valuechainkoala/status/2020845797345018101)  2026-02-09T13:02Z [--] followers, [----] engagements


"AI Data Center Liquid Cooling Value Chain + Profit Pool Planning and Design (1520%) Choose the cooling method and design how it fits into the racks. Tickers: $VRT $NVT $JCI $MOD $FIX Parts and Materials (1520%) Build the physical kit: cold plates connectors pipes tanks and coolant. Tickers: $VRT $NVT $MOD CDUs and Pumps (Bottleneck 1) (2025%) The heart that moves coolant and controls heat for 200kW+ racks. Tickers: $VRT $NVT $MOD System Integration and Data Center Build (Bottleneck 2) (2025%) Install and connect everything in new builds and retrofits. Tickers: $VRT $JCI $TT $FIX Operations"  
[X Link](https://x.com/valuechainkoala/status/2021637859367653522)  2026-02-11T17:30Z [--] followers, [--] engagements


"What worries me is less the stock story and more what this pivot means for Bitcoin over time. When big listed miners lock up multiyear power contracts to run AI workloads that energy is effectively taken off the table for securing the Bitcoin network slowing hashrate growth and making the system more dependent on a shrinking set of large players. In the best case AI cash flows keep miners solvent through crypto cycles; in the worst case capital and power quietly migrate away from BTC hashpower concentrates and the longterm security budget of the chain becomes a lot more fragile than most"  
[X Link](https://x.com/valuechainkoala/status/2022349504876265981)  2026-02-13T16:37Z [--] followers, [--] engagements


"When AWS says it expects to be sold out of [----] capacity while AI demand keeps accelerating what they are really saying is that power and infrastructure not customers are the constraint. Theyre pouring roughly [---] billion dollars into capex doubling datacenter power by [----] and still cant build AI capacity fast enough which is exactly why offmeter power (like $BE) and gridside infrastructure become the real bottleneck breakers for the next leg of the AI cycle. https://twitter.com/i/web/status/2022716389996245374 https://twitter.com/i/web/status/2022716389996245374"  
[X Link](https://x.com/valuechainkoala/status/2022716389996245374)  2026-02-14T16:55Z [--] followers, [---] engagements


"When AWS says it expects to be sold out of [----] capacity while AI demand keeps accelerating what they are really saying is that power and infrastructure not customers are the constraint. Theyre pouring roughly [---] billion dollars into capex doubling datacenter power by [----] and still cant build AI capacity fast enough which is exactly why offmeter power (like $BE) and gridside infrastructure become the real bottleneck breakers for the next leg of the AI cycle. https://twitter.com/i/web/status/2022715189800407514 https://twitter.com/i/web/status/2022715189800407514"  
[X Link](https://x.com/valuechainkoala/status/2022715189800407514)  2026-02-14T16:50Z [--] followers, [---] engagements


"$FIX (Comfort Systems) the hidden bottleneck behind AI data center buildouts Role (bottleneck): Downstream builder and integrator for data centers. $FIX gets paid when hyperscalers need power and cooling installed fast at scale. Trend: The AI data center boom is making construction speed and execution the real constraint. Catalyst chain: 2024: Profitability jumps double-digit revenue growth record backlog. Data centers shift from side work to the core engine. Q1 2025: Backlog $8.12B (up 16.5% YoY). About 37% of revenue from technology (including data centers). Q2 2025: Modular capacity 2.7M"  
[X Link](https://x.com/valuechainkoala/status/2022725022222291161)  2026-02-14T17:30Z [--] followers, [--] engagements


"AI computing is not really limited by chips anymore. Without advanced connectivity optics and SerDes AI clusters hit a hard bandwidth wall. That is why companies like $ALAB and $MRVL sit in the critical third step of the AI value chain"  
[X Link](https://x.com/EliavShfaim/status/2003068656331411652)  2025-12-22T11:42Z [--] followers, [---] engagements


"For years banks avoided stablecoins due to regulatory uncertainty. The GENIUS Act (2025) marks a shift from aggressive enforcement to a clear federal framework"  
[X Link](https://x.com/valuechainkoala/status/2005703776414622126)  2025-12-29T18:13Z [--] followers, [--] engagements


"Green policy shut refineries and froze new capacity so refining became the real bottleneck. Now Venezuelas heavy sour crude returns at a discount and Gulf Coast complex refiners are the gatekeepers $VLO They buy cheap feedstock turn it into diesel and jet fuel and export it to emerging markets. Value shifted away from crude producers and toward refiners that can process the dirty barrel. https://twitter.com/i/web/status/2012215494757982436 https://twitter.com/i/web/status/2012215494757982436"  
[X Link](https://x.com/valuechainkoala/status/2012215494757982436)  2026-01-16T17:28Z [--] followers, [--] engagements


"In energy markets the bottleneck is shifting from oil supply to the global ability to refine heavy dirty barrels. Catalyst flow: 2010s: ESG + tighter emissions rules start killing refinery economics 20142020: Underinvestment and closures quietly shrink downstream capacity 20172020: Venezuela sanctions choke off heavy crude exports 20202022: COVID accelerates refinery shutdowns; demand returns to a tighter system 20222024: Partial easing lets Venezuelan barrels return but only in limited volumes 20242025: More refinery closures across the U.S./Europe deepen the structural crunch Late 20252026:"  
[X Link](https://x.com/valuechainkoala/status/2012940551117529133)  2026-01-18T17:30Z [--] followers, [--] engagements


"Value Chain: Venezuelan heavy crude comeback 1) Upstream: extract heavy crude $CVX $XOM $COP $TTE 2) Oilfield Services: drill and service $SLB $HAL $BKR 3) Transport: ship crude globally $FRO $DHT $INSW 4) Complex Refiners (Bottleneck): refine dirty barrels $MPC $PSX $VLO $PBF 5) Distribution: sell refined fuels $XOM $SHEL $BP $TTE Takeaway: cheap heavy crude + limited refining = pricing power. https://twitter.com/i/web/status/2013302937372762524 https://twitter.com/i/web/status/2013302937372762524"  
[X Link](https://x.com/valuechainkoala/status/2013302937372762524)  2026-01-19T17:30Z [--] followers, [--] engagements


"Thesis: The winners in the heavy crude comeback are the Gulf Coast refiners that can actually process it. Why $VLO: Complex refining setup (coking + hydrocracking) built for heavy sour barrels Structural shortage in complex refining keeps margins elevated Catalyst flow: 2024-25: closures + underinvestment keep cracks tight Q3'25: 97% utilization $13.14/bbl margin shows strong operating leverage 2025-26: widening light vs heavy spreads benefits complex refiners Optional upside: renewable diesel + exports add a second earnings lever Takeaway: If heavy barrels flood back complexity wins."  
[X Link](https://x.com/valuechainkoala/status/2013665451981840830)  2026-01-20T17:30Z [--] followers, [--] engagements


"$BE: The Power Wall Is Here and Bloom Is the Shortcut AI data centers cant wait years for grid hookups. Bloom adds on-site megawatts using modular solid oxide fuel cell power plants. Catalyst chain: 2024: 80MW SK E and S deal proved Bloom can scale onsite power 2025: multi-billion data center power deals (Oracle Brookfield) 2025: installed base above 1.5GW scaling toward 10GW per year [----] to 2026: about 900MW with a major US utility 2026+: IRA credits improve economics Will hyperscalers choose Bloom over waiting for the grid https://twitter.com/i/web/status/2016926816624054455"  
[X Link](https://x.com/valuechainkoala/status/2016926816624054455)  2026-01-29T17:30Z [--] followers, [---] engagements


"Conviction is great but 93% in two smallcap names isnt just conviction its lotteryticket concentration. Volatility I can live with: permanent damage if one thesis breaks is harder to recover from. Even if $RKLB $ONDS crush it position sizing and some diversification matter just as much as being right on the story. https://twitter.com/i/web/status/2016063107051786568 https://twitter.com/i/web/status/2016063107051786568"  
[X Link](https://x.com/valuechainkoala/status/2016063107051786568)  2026-01-27T08:17Z [--] followers, [---] engagements


"Fair questions. A few thoughts from the IREN side: Bare metal + training only is a strawman. The Microsoft deal is a 5year $9.7B AI cloud contract for training and inference capacity anchored on 3GW of secured lowcost power thats not a onecycle trade its infrastructure. Workloads will absolutely tilt toward inference but both training and inference need the same scarce inputs: land power cooling and highdensity racks. Those are exactly the layers where IREN is vertically integrated and hardest to replicate. [--] years out IRENs edge isnt we rent GPUs its being a scaled owner of AIready campuses"  
[X Link](https://x.com/valuechainkoala/status/2016182630744469548)  2026-01-27T16:12Z [--] followers, [--] engagements


"@ArkkDaily you can almost group it into three layers: Brain $NVDA $AMD $TSM $GOOGL $AMZN $PLTR (AI compute data). Body $TSLA $DE $AVAV $RKLB $KTOS $ACHR (robots autonomy space/defense). Nerves & test gear $TER $LHX $TRMB (test comms sensing)"  
[X Link](https://x.com/valuechainkoala/status/2016185887109820637)  2026-01-27T16:25Z [--] followers, [---] engagements


"@curious_founder If 48GW of data centers are going offgrid Id bet on the powertoparkinglot plays: BE (onsite gas/solidoxide) IREN (AI campuses sitting on multiGW power) FLNC (big batteries to firm it all)"  
[X Link](https://x.com/valuechainkoala/status/2016430549351489845)  2026-01-28T08:38Z [--] followers, [---] engagements


"@optionscjp OpenAI is a big customer not the whole thesis. AI demand not one lab drives $ORCL / $AMD / $AVGO / $CRWV. If OpenAI stumbles the GPUs and data centers wont sit idle theyll just serve the next model winner"  
[X Link](https://x.com/valuechainkoala/status/2016434832360829285)  2026-01-28T08:55Z [--] followers, [---] engagements


"@thexcapitalist If youre worried about that transformer bottleneck $POWL is basically a picksandshovels play here they dont sell AI they sell the heavy electrical gear every new AI data center must buy to get power from transformers into the racks"  
[X Link](https://x.com/valuechainkoala/status/2016526924365840406)  2026-01-28T15:00Z [--] followers, [---] engagements


"@DeItaone The most ironic part is that a private stablecoin issuer is now behaving more like a shadow central bank than most actual central banks quietly stacking gold in a bunker while telling everyone fiat is the problem"  
[X Link](https://x.com/valuechainkoala/status/2016534873427583300)  2026-01-28T15:32Z [--] followers, [---] engagements


"Funny thing is the SELL bucket here is where the nuance really lives youve got everything from a fully scaled ad/search monopoly $GOOG to a maturing platform with real defense/AI rails $PLTR to plain vanilla gold proxies. Lumping them together might work for a shortterm trade but from a fundamentals/positioning lens the spread in business quality moats and cashflow durability inside that list is way too wide for a single blanket call. https://twitter.com/i/web/status/2017250105384902870 https://twitter.com/i/web/status/2017250105384902870"  
[X Link](https://x.com/valuechainkoala/status/2017250105384902870)  2026-01-30T14:54Z [--] followers, [----] engagements


"@KresoOptions Selling options only looks like printing money until volatility hits youre literally taking on small steady gains in exchange for rare massive losses that can blow up your account or create theoretically unlimited downside on naked calls"  
[X Link](https://x.com/valuechainkoala/status/2017317636808331615)  2026-01-30T19:22Z [--] followers, [---] engagements


"$AMAT is effectively the toll collector on every advanced memory wafer: its tools are embedded across HBM and DRAM process steps so it captures value from each chip produced. As AI becomes more memorycentric and more DRAM capacity shifts to HBM AMATs content per wafer and revenue per memory chip scale with every incremental dollar flowing into the memory stack. https://twitter.com/i/web/status/2017620082247995726 https://twitter.com/i/web/status/2017620082247995726"  
[X Link](https://x.com/valuechainkoala/status/2017620082247995726)  2026-01-31T15:24Z [--] followers, [---] engagements


"People obsess over Fibonacci levels but almost nobody asks where a company actually sits in the value chain. $RKLB isnt just another space stock its evolving into a fullstack space infrastructure prime (launch + satellites + critical subsystems) sitting right on the bottleneck for small and mediumlift access to orbit and turnkey space systems. Once you understand that role and how scarce it is you get a much clearer picture of longrun economics than any retracement line can give you. https://twitter.com/i/web/status/2017985371007549545 https://twitter.com/i/web/status/2017985371007549545"  
[X Link](https://x.com/valuechainkoala/status/2017985371007549545)  2026-02-01T15:36Z [--] followers, [---] engagements


"On this list $TSM is about as close as you get to a single point of failure. If advanced-node production at TSMC really stopped it wouldnt just hit NVDA/AMD/AVGO it would cascade into the cloud names and even the big platforms because the entire stack is levered to a steady flow of cuttingedge wafers. In that scenario most of these [--] picks dont just wobble they gap down together as the market reprices what AI cloud and devices look like without the silicon engine underneath. https://twitter.com/i/web/status/2017989550534980079 https://twitter.com/i/web/status/2017989550534980079"  
[X Link](https://x.com/valuechainkoala/status/2017989550534980079)  2026-02-01T15:52Z [--] followers, [---] engagements


"$RKLB is a nice launch pureplay but the eyes should be on the operating systems that will orchestrate satellites payloads and autonomous missions think of what $KTOS is doing with OpenSpacestyle softwaredefined ground and network stacks. Launch ends up a relatively small slice of the value pool because once youre in orbit most of the economics move to who controls routing tasking bandwidth allocation and data flows not who provided the initial ride upstairs. https://twitter.com/i/web/status/2017991979834236959 https://twitter.com/i/web/status/2017991979834236959"  
[X Link](https://x.com/valuechainkoala/status/2017991979834236959)  2026-02-01T16:02Z [--] followers, [---] engagements


"The numbers dont really support that take $IREN is still overwhelmingly a Bitcoin miner today. In FY24 it did about 184M dollars of revenue from Bitcoin mining vs roughly 3M from AI cloud and even in Q3 FY25 around 141.2M of 148.1M in revenue came from mining while only 3.6M came from AI services. https://twitter.com/i/web/status/2017996184020042206 https://twitter.com/i/web/status/2017996184020042206"  
[X Link](https://x.com/valuechainkoala/status/2017996184020042206)  2026-02-01T16:19Z [--] followers, [---] engagements


"Fair on looking at FY25 but the revenue split still tells the story. In FY25 IREN did $501m total revenue: $484.6m from Bitcoin mining vs just $16.4m from AI Cloud. So even in [----] roughly 30x more came from BTC than AI the decoupling is about the forward runrate not the actual P&L mix yet. https://twitter.com/i/web/status/2017998172677636491 https://twitter.com/i/web/status/2017998172677636491"  
[X Link](https://x.com/valuechainkoala/status/2017998172677636491)  2026-02-01T16:27Z [--] followers, [--] engagements


"The $9.7B Microsoft deal is massive but its a 5year contract that ramps as 200300MW comes online first with the rest of the 1.4GW only following later. In FY25 reported revenue was still about $484M from Bitcoin mining vs roughly $16.3M from AI Cloud the $1B AI runrate is a forward target not the current revenue mix yet. https://twitter.com/i/web/status/2018001102709444820 https://twitter.com/i/web/status/2018001102709444820"  
[X Link](https://x.com/valuechainkoala/status/2018001102709444820)  2026-02-01T16:38Z [--] followers, [--] engagements


"The real upside from this trend probably sits less with the raw miners and more with the processors the handful of Western companies $MP $UUUU that can actually refine separate and upgrade these critical minerals into batterygrade chemicals rare earth oxides and magnets are the ones likely to capture the fattest most durable margins as supply chains reshore. https://twitter.com/i/web/status/2018339390355878300 https://twitter.com/i/web/status/2018339390355878300"  
[X Link](https://x.com/valuechainkoala/status/2018339390355878300)  2026-02-02T15:03Z [--] followers, [---] engagements


"What this list is really telling you is where Wall Street thinks the picks and shovels of the AI/semis boom are. $NVDA $AVGO are the obvious winners but $ASML $LRCX $AMAT and $KLAC are the ones selling the lithography etch/deposition and processcontrol gear that every advanced logic and HBM fab must run to keep shipping bits. As long as the industry is starved for cuttingedge capacity those tools are the toll booths on the way to all of the earnings growth everyone is modeling. https://twitter.com/i/web/status/2018345861856518468 https://twitter.com/i/web/status/2018345861856518468"  
[X Link](https://x.com/valuechainkoala/status/2018345861856518468)  2026-02-02T15:28Z [--] followers, [---] engagements


"$PLTR and $IREN dont really belong in the same bucket. Palantir is a software + data platform with sticky highmargin contracts embedded deep in customer workflows. IREN is mainly an AI infrastructure / datacenter and power play pivoting out of mining with capitalintensive buildout and a few big AI cloud deals. Slapping the same forward P/S on both because they share an AI buzzword ignores where they sit in the value chain and how durable their profit pools are. If youre bullish on IREN the case should be about contract quality unit economics and capex risk not PLTR is 50x so IREN should 10x."  
[X Link](https://x.com/valuechainkoala/status/2018693391106683033)  2026-02-03T14:29Z [--] followers, [---] engagements


"The way I see it the Nikkei piece is mixing two very different stories. In the global AI memory value chain $MU / SK Hynix / Samsung have already rotated capacity to HBM3E/4 and structurally reduced their China exposure. That profit pool lives with hyperscalers and advanced nodes not with domestic Chinese DRAM. Where CXMT / YMTC do matter is as a longterm price & capacity risk in legacy DRAM/NAND and for secondtier players like Nanya/Winbond that live in that segment. So the right question isnt Will this kill MU/Hynix its: How does Chinese capacity change the shape of the memory profit pool"  
[X Link](https://x.com/valuechainkoala/status/2018710261620731973)  2026-02-03T15:36Z [--] followers, [----] engagements


"I actually think were already past the point where this is all capex no payoff. Look at how the megacaps are talking: $MSFT is openly saying AIenhanced cloud and productivity are adding billions in new revenue $GOOG is leaning on AI to drive search/ads and ship software faster internally and $AMZN is tying its AI spend directly to expected AWS and retail productivity gains. The capex wave is huge no doubt but its not a blind leap off a cliff. For the strongest players you can already see AI moving from narrative to P&L. https://twitter.com/i/web/status/2018713456132735224"  
[X Link](https://x.com/valuechainkoala/status/2018713456132735224)  2026-02-03T15:49Z [--] followers, [---] engagements


"One big thing missing here is leverage. If you walk into a bank with a stock portfolio youll usually get limited volatile margin at best. Walk in with real estate and theyll happily give you longdated fixedrate financing against it. That access to cheap stable leverage can turn a 3% nominal appreciation asset into much higher equity returns over time if you manage the risk well. https://twitter.com/i/web/status/2018714141460926888 https://twitter.com/i/web/status/2018714141460926888"  
[X Link](https://x.com/valuechainkoala/status/2018714141460926888)  2026-02-03T15:52Z [--] followers, [---] engagements


"$LRCX is quietly levered to that impossible $MU move. HBM4 isnt just another memory node the 16high stacks are expected to consume materially more wafers per bit (industry chatter is 30% more wafers vs prior gen) which means Micron Samsung and SK Hynix have to push a lot more silicon through etch/deposition steps to hit their AI capacity targets. Who sells the highaspectratio etch and related tools that make those HBM4 TSVs and capacitor structures possible Lam. So as HBM4 ramps and DRAM makers crank capex $LRCX sits in the middle of that spend as the pickandshovel beneficiary behind MUs"  
[X Link](https://x.com/valuechainkoala/status/2018719008564928937)  2026-02-03T16:11Z [--] followers, [---] engagements


"$CRM sits in one of the most crowded links in the software value chain: generic enterprise CRM and sales tooling. Even before AI it was a brutally competitive space; now AI copilots from Microsoft HubSpot and others are attacking its core workflows. In that context AI feels more like a threat to CRMs pricing power and differentiation than a clear tailwind. https://twitter.com/i/web/status/2019083631042670730 https://twitter.com/i/web/status/2019083631042670730"  
[X Link](https://x.com/valuechainkoala/status/2019083631042670730)  2026-02-04T16:20Z [--] followers, [---] engagements


"HBM4 is sneaky good news for the tool makers behind Microns big Idaho fab. As HBM stacks move to 16high each terabit of memory is expected to consume materially more wafers ballpark 30% more versus prior HBM generations. That means $MU and the rest of the DRAM gang will need more etch/deposition capacity to feed those lines and $LRCX which dominates highaspectratio etch for TSVs and 3D DRAM is positioned to capture that extra 30% wafer intensity directly in equipment orders and services. https://twitter.com/i/web/status/2019085528008712292 https://twitter.com/i/web/status/2019085528008712292"  
[X Link](https://x.com/valuechainkoala/status/2019085528008712292)  2026-02-04T16:27Z [--] followers, [---] engagements


"HALEU is quietly moving from experimental fuel to strategic infrastructure. If SMRs become the on-site power solution for AI data centers HALEU becomes the real bottleneck. Here is the catalyst chain: Dec 2021: DOE launches the HALEU Availability program and signals multi-ton demand for advanced reactors. Jul 2023: TerraPower and $LEU expand plans to build commercial HALEU enrichment capacity in Ohio for the Natrium timeline. [----] to mid 2026: DOE formalizes a federal allocation process and commits to supplying HALEU to multiple reactor developers. 2025: U.S. policy shifts toward nuclear fuel"  
[X Link](https://x.com/valuechainkoala/status/2014027714437718063)  2026-01-21T17:30Z [--] followers, [--] engagements


"Nuclear is a layered value chain. The bottleneck gets tighter as you zoom in. Layer [--] (macro chain): Mining conversion enrichment fuel fabrication build operations back end. Bottleneck: enrichment and advanced fuel capacity. Layer [--] (enrichment): LEU is mature. Bottleneck: HALEU enrichment for SMRs and Gen IV. Few licensed players almost no scale. Layer [--] (HALEU): The bottleneck is a triangle you must solve at once. Licensed HALEU cascades long-term DOE or offtake contracts and conversion plus deconversion integration. That is where timelines break and leverage concentrates. Tickers: $LEU"  
[X Link](https://x.com/valuechainkoala/status/2014752491288526849)  2026-01-23T17:30Z [--] followers, [--] engagements


"Greenland Rare Earth Value Chain: Bottlenecks and Profit Pool Mining (10 to 15%) $CRML $MP $USAR $LYC Separation and chemistry (25 to 35%) Bottleneck #1 $NEO $MP $LYC $ILU Alloys and powders (15 to 20%) Bottleneck #2 $NEO $USAR NdFeB magnets (20 to 30%) Bottleneck #3 $NEO OEM demand (15 to 25%) $TSLA $F $RTX $GM Recycling (5 to 10%) $NEO Takeaway: Greenland headlines start with mining but the money sits in separation and magnets"  
[X Link](https://x.com/valuechainkoala/status/2014760043577577659)  2026-01-23T18:00Z [--] followers, [---] engagements


"$LEU is not just a uranium name. It is a U.S. enrichment bottleneck. Role: a unique American centrifuge operator sitting on the midstream choke point for LEU and HALEU supply. Trend: Nuclear Renaissance on-site power for AI data centers. Catalyst flow: 2023: Piketon begins HALEU production and delivers the first [--] kg under a DOE contract. Jun 2025: delivers [---] kg of HALEU to DOE over [---] kg total a real track record. Jun 2025: DOE extends the HALEU contract through mid [----] with options for up to [--] more years. Sep 2025: announces multi-billion dollar expansion plans at Piketon to scale"  
[X Link](https://x.com/valuechainkoala/status/2015114876696326593)  2026-01-24T17:30Z [--] followers, [---] engagements


"Chipmaking Equipment: The Profit Pool Map Behind Leading Edge Nodes Approx share of the profit pool by layer: Lithography (25 to 30%) Main bottleneck $ASML $INTC Deposition (20 to 25%) $AMAT $LRCX $TEL Etch and patterning (15 to 20%) $LRCX $AMAT $TEL Cleaning CMP wafer prep (10 to 15%) $TEL $KLAC $ENTG Metrology and inspection (10 to 15%) Secondary bottleneck $KLAC $AMAT Advanced packaging and hybrid bonding (5 to 10%) $AMAT $TEL $INTC Which company is the biggest bottleneck https://twitter.com/i/web/status/2015847202175033704 https://twitter.com/i/web/status/2015847202175033704"  
[X Link](https://x.com/valuechainkoala/status/2015847202175033704)  2026-01-26T18:00Z [--] followers, [---] engagements


"$IREN: Land + Power Leverage to the AI Buildout Role: controls scarce land and cheap power for hyperscale AI sites. Catalyst timeline: H2 2025: 75MW Blackwell ready AI campus (Horizon 1) demonstrates GPU-first execution. Nov 2025: 9.7B multi-year AI cloud deal with Microsoft anchors demand. [----] to 2027: up to 2.75GW permitted power in Texas expands the runway. [----] to 2028: Sweetwater multi-GW buildout is the next scale step (execution risk). Takeaway: AI scales on power. IREN is positioned upstream of demand. Where do you see the stock in [--] months: higher or lower"  
[X Link](https://x.com/valuechainkoala/status/2016209594671149467)  2026-01-27T18:00Z [--] followers, [---] engagements


"Hypersonic Missiles Value Chain: Where the Profit Pool and Bottlenecks Sit R&D and design (10 to 15%) $LMT $RTX $NOC $KTOS Propulsion (25 to 30%) Main bottleneck $RTX $NOC $LMT $KTOS Thermal materials (20 to 25%) Bottleneck $HXL $RTX Guidance and avionics (10 to 15%) $LMT $RTX $GD Testing and production (15 to 20%) $LMT $RTX $NOC $KTOS Launch and C2 (5 to 10%) $RTX $LMT $NOC Whats your pick for the biggest Golden Dome winner Deep dive is in my bio"  
[X Link](https://x.com/valuechainkoala/status/2016571986219241547)  2026-01-28T18:00Z [--] followers, [--] engagements


"How the New Defense Budget Could Supercharge [--] Weapon Profit Pools Smart munitions + 155mm the consumables that never stop Modern wars burn through guided rounds fast. Replenishment + upgrade from dumb to guided keeps volume demand structurally high. Tickers: $LMT $RTX $NOC $GD $RHM $ESLT $POWW Air & missile defense the fastest re-rating corner of defense Drones cruise missiles and ballistic threats are forcing countries to buy layered defense radars interceptors and C2. This is where budgets can compound not just refill. Tickers: $RTX $LMT $NOC $LHX $ESLT Loitering munitions + long-range"  
[X Link](https://x.com/valuechainkoala/status/2017289204913623490)  2026-01-30T17:30Z [--] followers, [--] engagements


"Profit Pool Map: Memory Manufacturing Equipment (DRAM and HBM) If AI runs on HBM this is where the tool profits concentrate. 1) Lithography (Chokepoint #1) EUV and DUV patterning that sets density and scaling limits for DRAM and HBM. 2025% Tickers: $ASML NIKON CANON 2) Deposition and Etch Thin-film deposition and high-aspect-ratio etch for DRAM capacitors and interconnects. 2530% Tickers: $AMAT $LRCX $TEL 3) CMP Cleaning Metrology Planarization wafer cleaning and yield control as memory complexity rises. 1520% Tickers: $AMAT $KLAC $TEL 4) Advanced Packaging (Chokepoint #2) TSV microbumps"  
[X Link](https://x.com/valuechainkoala/status/2018738755386085488)  2026-02-03T17:30Z [--] followers, [---] engagements

Limited data mode. Full metrics available with subscription: lunarcrush.com/pricing