History Is Being Made. You Only Get One Chance!

← Retour au Tableau de Bord

URL YouTube

https://www.youtube.com/watch?v=s2Kqrf4emrs

Statut

Analyzed

Demandé Le

June 06, 2026 at 06:00 AM

Performance Globale

-5,30%

Recommandations

CIEN BUY
""And that pullback dropped right into my buy zone.""
Contexte: [01:05:48:20 - 01:05:58:02] After discussing Ciena’s Q2 results and a market-driven pullback, the speaker indicates the pullback reached their intended entry level.
Prix à la date de publication: $488,21
Prix de clôture du dernier jour: $462,34 (Jul 10, 2026)
Bénéfice/Perte: $-25,87 (-5,30%)

Transcription Complète

[01:00:00:03 - 01:01:29:12] In October, 2011, a 50-year   flooding event in Thailand took roughly  80% of the global hard drive capacity   offline. Seagate's gross margin moved from  around 20% to roughly 32% in two quarters,   and net profit nearly quadrupled in just a single  quarter. But here's the thing. Hormuz is not a   regional disruption like Thailand was. Because in  addition to oil, the straight handles roughly a   quarter of the global liquefied natural gas, the  primary shipping route for helium out of Qatar,   and a significant share of the global sulfur  supply that's responsible for sulfuric acid   that's used to etch circuit boards and chip  wafers. And helium also cools the lasers that   run every fiber optic cable inside a modern data  center. And it seals every high-capacity hard   drive. So when those inputs get harder to source,  a meaningful part of the global technology supply   chain completely stops working. The risk to  producers without pricing power is severe,   and the players that stand to win this setup are  the ones that are positioned exactly like Seagate   was in 2011. So in this video, I'm going to  be walking through 10 technology companies   that pass a very clean filter. Two quarter  margin expansion already on the record? Nine   contracts with named hyperscaler customers?  And production already booked through 2027?   The first six are going to be storage, optical,  and semiconductor names. And the last four are   just going to be AI power names, of which they  all happen to be the ones that I'm watching.   [01:01:29:12 - 01:01:36:08] want to address something right away because a lot   of these names have run very hot and several of  them are sitting at their near all time highs.   [01:01:36:08 - 01:01:53:04] so as I walk through these, watch for pullbacks.   You probably don't want to chase these straight  up, but keep this in the back of your head to   the supply constraints tied directly to these  companies. It's very real and they can easily push   these stocks well past their all time highs [01:01:53:04 - 01:01:53:12]   from [01:01:53:12 - 01:01:59:22]   Both of these things are true at the same time.  So high conviction in the business, but you need   to have discipline in the entry. [01:01:59:22 - 01:05:30:16]   But before I dive in, if you get any value from my  videos, please consider pressing the like button   so my channel can continue to grow. And of course,  as a reminder, I'm not a financial advisor,   and this is only for educational purposes.  However, if you'd like to see my watch list or   my trades that are happening in real time and my  portfolio, I'd love for you to join the community   on Patreon and Discord. Alright, the first name  on the list is Lamentum, and what they do sits   at the absolute physical bottom of every modern  AI data center. Lamentum makes the lasers and   the optical components that move data between  GPUs at AI training speeds. Specifically, they   manufacture externally modulated lasers, or EMLs,  and ultra high-power pump lasers that integrate   into the optical transceivers running between  racks of Nvidia and AMD chips. When a hyperscaler   builds a training cluster, the connectivity layer  tying thousands of GPUs into one logical machine   is photonics. And Lamentum is the supplier the  rest of the industry depends on for the highest   bandwidth tier. The reason Lamentum specifically  benefits from this Hormuz-style supply tightness   is their structural position. The 200 gigabit per  lane EML is the component required for the next   generation of 1.6 terabit optics. And Lamentum  is the dominant volume supplier of that part.   Per CPO market research, Lamentum is currently the  only company shipping 200 gigabit per lane EMLs   at volume, with Coherent operating in adjacent  photonic segments. And industry analysts project   a 36% EML supply shortfall just in 2026. So when  the macro environment tightens further, customers   stop negotiating and they start just locking  in capacity at the premium prices. Which is   exactly what Nvidia did with a $2 billion equity  investment and a multi-year purchase commitment in   March of 2026. Lamentum's gap gross margin tripled  from 16.6% in mid-2024 to 44.2% in the most recent   quarter. And the simple reason is segment mix.  The AI optics business now represents 86% of their   revenue. And the legacy commercial laser segment  that makes up the remaining 14% of the business   is actually shrinking or going down 15% year over  year. Next on the list is Sienna, which sells the   optical transport boxes that move data between  data centers when a single AI training cluster   spans more than just one building. And the product  driving the current cycle is WaveLogic 6 Extreme,   a coherent optical platform that lets hyperscalers  run a single logical GPU fabric across multiple   physical campuses. 90 plus global customers have  committed to it in the past two quarters. And the   four largest hyperscalers have lined up over $600  billion in 2026 capex that flows directly into the   boxes that Sienna makes. Sienna's specific benefit  comes from the structure of the Western optical   market. After Nokia had absorbed Infinira in early  2025. AI and optical networking just became a   two-vendor field, with Sienna holding roughly  half of the US market. In a two-vendor market,   when supplier costs rise across the Bill of  Materials, those costs do not get negotiated   away. Their CFO laid out the plan in the 2026 Q1  call. The company raised prices in the quarter   alongside engineering cost reductions. And  they guided two additional price increases,   kicking in during the second half of the year. [01:05:30:16 - 01:05:35:03]   The Q2 results just came out and Sienna  grew their revenue 40% year over   [01:05:35:03 - 01:05:41:03] Accelerating from 33% the quarter before and   adjusted earnings per share nearly quadrupled. [01:05:41:03 - 01:05:48:20]   raised their full year guidance to $6.3  billion, which works out to be about   32% growth for the year. [01:05:48:20 - 01:05:55:13]   And even after all of that, the market sold the  stock down hard because expectations going in   were just ridiculously high. [01:05:55:13 - 01:05:58:02]   And that pullback dropped  right into my buy zone.   [01:05:58:02 - 01:07:02:18] you've probably noticed the quant   ratings and the earnings transcripts that have  been pulling up for the first couple of names.   That all comes from one place, and that brings us  to today's sponsor, Seeking Alpha Premium. Now,   I have built my own research process over the  years, but Seeking Alpha earned a permanent   slot in my workflow, and I've been using it  now for the past two years. And there are two   features that I lean on the most. The earnings  call transcripts that let me search exactly what   management said about guidance without having to  sit through a full 90-minute call. Take Palantir,   for example. On their latest call, I pulled the  line where the CFO raised their full-year revenue   guidance to the $7.65 billion range, which  happened to be their largest raise ever in   all of about 10 seconds. And the quant ratings, I  happen to use as a data check against my own read,   especially where the model disagrees with me,  because it's sometimes more important than   just when it agrees. And the methodology holds up,  because Seeking Alpha's back-tested quant strategy   beat the S&P 500 every single year for the last  12 years. Use the link in the description to get   $30 off your first year and a seven-day free  trial. If you're looking for serious help with   your research, this is a great place to start. [01:07:02:18 - 01:22:36:07]   Third on the list is Seagate Technology. This is  the name where the Thailand parallel is obviously   going to be the most direct. Seagate makes the  high-capacity hard drives that hyperscalers use   for near-line storage, the data they want fast but  not GPU fast. Their Mosaic platform is the only   heat-assisted magnetic recording drive shipping  at scale, with two hyperscaler customers already   qualified. So roughly 87% of their shipments go  to cloud and data center customers. And here's   the thing. Every drive above 10 terabytes is  helium-sealed. That helium dependency is what   ties Seagate directly to the Hormuz situation.  Qatar's complex produces roughly 30-38% of the   world's helium supply. And Qatar Energy declared  force majeure on the facility in March 2026,   after the natural gas halt cascaded down into  the helium extraction. This outcome resulted   in Seagate having a very direct response. The  company raised production allocation pricing   by 20-30% and finalized bill-to-order contracts  with hyperscalers through fiscal 2027. So what   the financials show is the largest two-quarter  margin move across all of these ten names. Gap   gross margin expanded from 41% in December 2025  to 46.5% in March of this year. And their future   guidance for 2026 puts gross margin above  50% for the very first time. This is the   2011 Thailand pattern playing right out in front  of us. Only it's a little bit cleaner this time,   because in 2011 Western Digital lost share when  its Thailand fabs completely flooded out. But in   2026, both of these companies are going to benefit  from that same helium-led recovery. Which brings   me to the fourth company on the list, which is  Western Digital, which happens to have the cleaner   balance sheet version of the same Seagate trade.  Western Digital completed a full spin-off of its   Sandisk flash business in February 2025, leaving  the parent company as a pure play hard drive maker   for the very first time in its modern history.  Roughly 89% of revenue now comes from cloud and   hyperscaler customers. And the technology path  differs from Seagate. Where Seagate is leaning on   heat-assisted magnetic recording, Western Digital  is leaning on Ultra-SMR, which is a higher density   approach that's already shipping 40 terabyte  volume drives in early 2026. Western Digital   captures the helium constraint through this same  channel as Seagate, with the same hyperscaler   customer base on the receiving end. But what  Western Digital did with that setup is lock it   down a little bit longer. Five of the company's  top seven hyperscale customers are now on a   multi-year long-term agreement through 2026, and  one of those customers committed through calendar   2027, and select agreements extended to 2028 and  also 2029. They raised average selling prices   roughly 9% year-over-year on the most recent  quarter alongside the "Build to Order" book.   Western Digital also crossed 50% gross margin for  the very first time in its history in Q3 of this   year. And the segment data shows the milestone is  essentially a cloud business achievement. Because   the cloud segment is now 89% of revenue and grew  48% year-over-year. Combined with the "cleanest   balance sheet across these 10 names," that is  what makes Western Digital the cleaner side of   the same Seagate trade. Now fifth on the list is  Credo Technology Group. And it's the only name   in this video whose business gets stronger  when the helium and the laser supply chain   breaks even further. Credo makes the high-speed  copper cables that connect AI chips and switches   inside a data center rack. These are called active  electrical cables. For long-distance connections,   fiber optic cables win hands down on power  efficiency. Which is why fiber dominates   the internet backbone. But for short connections  inside a single server rack, copper actually wins.   Because every fiber link needs a laser and the  power-hungry electronics that convert electrical   signal to light and then back again. Credo's  smart copper cables skip that conversion entirely,   using roughly half of the power of a comparable  fiber transceiver at short reach. And Credo holds   about 73% of that market. Now here's why Hormuz  drives demand. Every short-reach fiber link in   a data center needs a laser. Rare materials and  helium to keep the laser cool. When those inputs   get harder to source because of Hormuz constraint,  hyperscalers can't really get enough fiber   transceivers at any price. They are essentially  forced to architect more of their builds around   Credo's copper for every link where copper works.  Now this isn't exactly a temporary substitution.   It is a redesign of how the racks get wired. And  Credo is the structural beneficiary on both the   cost side and the demand side of that shift. In  the most recent fiscal quarter, Credo tripled   its revenue compared to the same quarter a year  earlier. And gross margin expanded to over 68%. If   you were to look back a year ago, the company was  losing money. But in their most recent quarter,   they were earning almost 39 cents of net profit  on every dollar of revenue. That's the exact type   of trajectory of a company on the right side  of the trade. Our next company is Astera Labs,   and Astera is the chip company that benefits when  hyperscalers redesign their racks away from fiber.   Astera makes three kinds of connectivity chips  that sit between AI processors. The first is a   retimer, which keeps high-speed signals clean as  they travel through copper inside of a rack. The   second is a smart cable product that competes  head-to-head with Credo. And the third is a   fabric switch that connects many GPUs into a  single logical machine. Nvidia, AMD, Microsoft,   Google, and Meta, they all use Astera Silicon. And  of course, Astera holds a board seat on OpenUALink   standard with the hyperscalers built specifically  so they could avoid being trapped by Nvidia's   standards. Now Astera's angle on all of this is  rack redesign. When fiber optic connectivity gets   really expensive and constrained, hyperscalers  do not want to reduce the numbers of GPUs that   they're ordering. Instead, they redesign the  racks, packing more processors into a single   rack or a pod, and connecting them with copper  and short-reach links. Because that side of the   supply chain is not affected by Hormuz.  And every time a hyperscaler does that,   the chip content per rack goes up. Meaning more  retimers, more fabric switches, and more smart   cables. Oh, and guess what? Astera makes all  three. And their new 320-lane fabric switch,   which just started shipping in volume in May of  this year, is specifically designed to support   exactly that redesign. Most recently, Astera grew  revenue 93% year over year, while expanding gross   margin to 76%. That is a fabulous chip designer  growing at hyperscaler speed, with gross margin   above three-quarters of revenue. A combination  that does not exist really anywhere else in this   video. Now we're going to move on to AI power side  of the list, starting with Constellation Energy,   which operates the largest nuclear fleet in the  United States. 21 reactors with 22 gigawatts of   capacity. At the beginning of the year, the  company closed a $16 billion acquisition of   Calpine, adding 33 gigawatts of natural  gas and geothermal generation. Combined,   Constellation is now the largest private sector  power producer in the country. Its biggest   customers are the hyperscalers building AI data  centers everywhere. Microsoft signed a 20-year   contract for the restored Three Mile Island. Meta  signed a separate 20-year contract for the entire   1.1 gigawatt output of the Clinton, Illinois  plant. Both, of course, are at hyperscaler   premium pricing. And their connection to Hormuz  comes through natural gas. The Strait carries   roughly a quarter of the world's seaborne LNG.  And when global natural gas supply tightens, U.S.   natural gas gets pulled overseas to fill that gap.  And of course, U.S. gas prices rise as a result,   which pushes up electricity prices on the grid,  because gas plants set the going rate for most   of the day. Constellation's nuclear plants run  at a fixed marginal cost of around $32-$35 per   megawatt hour, which makes them the cheapest  source of power on the grid. And of course,   when the price rises because of gas, the  nuclear fleet captures the entire spread,   more revenue and more margin. And the big thing  that changes for Constellation is their customer   base. Five years ago, the company was a regulated  utility with slow-growing investments. Today, it's   the dominant counterparty for the hyperscalers  that are chasing 24-7 carbon-free power, with two   20-year contracts already locked in and a combined  fleet that has more than doubled. Their management   team is already guiding 2026 earnings up nearly  20% year-over-year simply because of the strength   of the Kalpine integration and the rising power  prices. And right on the heels of Constellation   comes Talon Energy, the concentrated single-asset  version of the same trade. Talon operates the   Susquehanna nuclear plant in Pennsylvania, with  2.5 gigawatts of generating capacity. In June of   last year, the company signed a 17-year agreement  for $18 billion with Amazon Web Services. That   single contract covers 76% of Susquehanna's  entire output. Talon also closed two natural   gas plant acquisitions in November 2025, and they  signed a $3.4 billion gas plant deal in January   of this year, building out a multi-jigawatt  dispatchable fleet alongside the nuclear asset.   Talon captures the same gas price spillover that  benefits Constellation, but through a little bit   of a different lever. The Amazon contract is  fixed price for the first 10 years and then   resets to a fixed margin above market prices for  the remaining seven. So with natural gas, prices   stay elevated because of Hormuz pressure through  the late 2020s and into the 2030s. Talon does not   have to renegotiate to capture that upside because  the reset is going to be automatic, and the margin   rides the energy and capacity prices wherever  they go. The real driver is the operational   activation that's happening in spring of this  year. Talon is in the middle of a refueling   outage in Susquehanna, and the transmission  reconfiguration that turns Amazon contract   from announced to in-service is being completed  at the same time. In their latest quarter, their   revenue had grown 79% year over year and adjusted  free cash flow quadrupled. The Amazon contract   has not yet started delivering its full economics,  and the full ramp continues through 2032. Now the   third name on the power side of the list is Bloom  Energy, and Bloom is the one that bypasses the   grid entirely. Bloom makes solid oxide fuel cells.  These are not the backup generators that sit   idle in a hospital basement. They happen to be  the small power plants that run 24 hours a day,   converting natural gas or hydrogen directly  into electricity through an electromechanical   reaction at the customer's location. The customer  simply plugs the data center into the fuel cell   as a primary source of power, and they don't even  touch the grid. Bloom is the only fuel cell maker   shipping at gigawatt scale, and they have signed  three transformative contracts in the past year.   Oracle even locked in a master agreement for up  to 2.8 gigawatts. An AEP signed a $2.6 billion   agreement covering Amazon and another customer.  And Brookfield committed $5 billion to finance   further Bloom deployments at AI sites. So where  Constellation and Talon capture Hormuz through gas   price spillover, Bloom captures it on the demand  side. Connecting a new AI data center to the grid   takes 5 to 7 years because every US utility is  already overwhelmed by AI capacity requests.   So when you layer on Hormuz volatility on top of  that bottleneck, hyperscalers really stop waiting   and they start acting. The obvious choice is just  to build their own power on site. Bloom can ship   a working power plant to a data center within a  few months, not having to wait years. The Oracle,   AEP, and Brookfield contracts are proof that  the largest AI buyers in the world have made   their decision and have really started to commit  at scale. After a decade of losses, Bloom flipped   to positive operating income in 2024, and they've  been accelerating ever since. In their latest   quarter, their revenue has grown 130% year over  year, but the product line grew 208% on its own.   And their product line runs at a 35% gross margin  versus 18% for services. So that mix shift to   product is driving both the top line growth and  the margin expansion. Now we'll move on to the   next name on the list, which is Powell Industries.  And that happens to be the small cap pick within   the list. Powell designs the custom engineered  electrical equipment that sits between a power   source and a builder's actual circuits. Their core  products are switch gear, the high voltage devices   that route and protect electrical flow, and bus  duct and heavy duty conductor that carries large   amounts of electricity inside of a facility. When  a hyperscaler builds an AI campus, somebody has to   engineer the substation equipment that connects  to the data hall, and Powell is one of the only   specialists that they can go to for that type of  work. On the data center side, the company secured   a $400 million mega order for AI data center  switch gear. That single order is the largest   project in Powell's history. And on the oil  and gas side, when the Strait of Hormuz creates   risk to international natural gas capacity, the  structural case for US-based LNG exporters begins   to strengthen. And those exporters need exactly  the kind of switch gear that Powell sells. Oil and   gas already represent 29% of Powell's backlog, and  management has explicitly tied the outlook to US   exporter advantages from the international supply  constraints. The segment data shows where Powell's   growth is really coming from. Electric utility  revenue, the bucket where utility routed data   center work gets billed, grew 49% year over year.  And in oil and gas, the legacy industrial segment,   it stayed relatively flat. And the results are  really showing up in their backlog, where it had   started at $1.8 billion, and it's already been  bumped up to $2.2 billion. To boil it all down,   the AI grid work is completely taking over their  business. That's a good thing. Now, that happens   to be the 10 names. The Hormuz constraint flows  into the technology supply chain through three   different channels. The helium and rare materials  that go into lasers and high capacity hard drives   travel straight through the straight. And that  flow is what drives the margin movers at Humentum,   Siena, Seagate, and Western Digital. That same  constraint creates demand for the copper and   chip level alternatives that bypass the laser  supply chain entirely, which is what Credo and   Astera Labs are doing right now. And the  natural gas that sets electricity prices   gets pulled overseas when global LNG begins  to tighten, which is the mechanism driving   margin expansion at Constellation, Talon, and  Bloom Energy on the power side. Powell happens   to sit at the intersection of both, selling the  switchgear that goes into AI data centers and US   LNG export facilities at the same time. So every  one of these companies is showing two-quarter   margin expansion with demand locked in for the  next year. As I said at the beginning, that's   my filter. The companies that pass it are the ones  positioned to convert this constraint into margin,   not just absorb it. As always, thanks so much  for watching, and I'll see you the next time.