Transcription
So, a Patreon member had sent me a list of stocks last week, and two of them are up over 2,000% in just the past 5 years. They're all companies that support data center and AI buildouts, and they prove that this AI buildout has already paid off at least once at scale. And really, the question now is who carries it into the next chapter.
So, what I did is I took his list and then I found several more. So, I'm going to be covering 10 names in this video where most of them are going to be completely new to you. And keep in mind that it's important to get familiar with these names, but wait for the right moment to buy after you've done your research.
So, if you're new here, then you probably don't know me or the fact that I retired at 46, all thanks to my investments. And this is after a career with companies like Target and Amazon where I happen to have made them billions upon billions of dollars. So, before we dive in, I do want to point out that I'm not a financial adviser, and this is only for educational purposes.
Now, when the biggest names in tech are throwing money at AI, there are four layers of infrastructure before any of it ever reaches a single server rack. So, we're thinking about the contractors who physically build out the data centers or the equipment that goes inside. And then, of course, there's the cooling equipment or the hidden layer of fiber materials and the on-site power. So, I'm going to walk through the best names in each of those layers. And at the end, of course, I will tell you exactly how I would invest in this group if I were starting fresh. But if you want to see exactly what I'm buying and selling, I do give access to that information in my Patreon group. And I also provide additional research and analysis every week, like my watch list of stocks like the ones that I'm covering in this video.
We'll start with tier one, the contractors who physically build the data centers. The hyperscalers, Microsoft, Amazon, Google, Meta, and Oracle are spending over $700 billion on AI infrastructure this year. And this isn't just a one-year boom. McKenzie projects $7 trillion in global data center spending through 2030. And most of the companies on this list already have orders booked through 2028. I'm not going to lie, when I first read that $7 trillion number, I I had a hard time believing it because that happens to be 23% of the entire US GDP for 2025. And somebody has to physically construct all of that.
First on this layer is Comfort Systems USA where they install the mechanical, electrical, and plumbing systems inside hyperscaler data centers. When a hyperscaler wants one company to handle everything inside the building, Fix is the one that they call. What they have accomplished is transformative. Data center customers went from a third of revenue 2 years ago to over half of their revenue this past quarter. And Fix ended at a record 122.4 billion backlog that nearly doubled year-over-year. For context, if you had invested $1,000 in comfort systems in the spring of 2021, you'd have roughly $22,000 today. That is a 22 beggar in 5 years for an HVAC contractor. Seriously, who would have guessed that? And here's the kicker. Operating margin more than doubled in that window. Return on invested capital sits at 62% and they're carrying effectively zero debt. These are software company economics on a mechanical contractor.
Next is IEES Holdings where they handle the electrical work that brings data centers to life. Cabling, fiber, low voltage systems, the wires and boxes that connect the racks. Their communications segment serves data centers and the numbers there have been extraordinary where their total backlog jumped 62%. Now, what makes IEC different is size. They're running at around 12.6 6 billion in market cap, which means they're still small enough to compound aggressively if data center constructions stays at this pace through 2028. The revenue compounded at 23% every year for five straight years, while operating margin nearly tripled and earnings per share has multiplied nearly eight times. And that's what gets this stock to a 10bggger. Now, for me, the metric to watch on IEC is communication segment growth. As long as that segment just keeps printing 35% plus, then hyperscaler demand is intact and the runway just extends.
Now, we've been talking about the physical AI buildout and there's another buildout happening on the science side. And that brings us today's sponsor, Telescope Innovations, where I first covered them back in January, and the story has significantly improved since then. Developing a drug requires up to 30,000 experiments, and one chemist runs about five a week. Telescope's self-driving lab does a year of that work in a single week where AI designs experiments, robots execute, and machine learning optimizes the next round. Based on management's projections, their technology cuts development costs from $900 million to 300 million, and it shaves 5 years off a 10-year cycle. That's 50% more high margin patent monetization. They shipped their first commercial self-driving lab in December. And in January, Fizer installed a second one, moving the partnership from R&D pilot to full deployment. That's really big news because it validates the need for this technology. Their second existing product, Direct Inject LC, is already used by 15 of the 20 largest pharma companies. The team includes the only living two-time Nobel laurette in chemistry, as well as founder and CTO, Dr. Jason Hine. Collectively, this team is attributed with 34 patents, 450 academic publications, and management with solid prior exits within the space. And telescope has expanded into lithium refining with two proprietary processes and a strategic project with standard lithium, tying directly into the need for critical minerals, which then again ties back to the AI buildout, real revenue, top tier pharma transitioning from pilots to full deployment, and a stock pushing all-time highs. That's the setup that investors are looking for. If you'd like to learn more about Telescope Innovations with a ticker teif, please check out the link down in the description.
Third on this list is Quantis Services, the largest pureplay infrastructure contractor in the entire category. Power transmission, communications, and energy infrastructure work for hyperscalers, utilities, and large industrial customers across multiple sites simultaneously. They're the public company hyperscalers and utilities call when they need a billion dollars of work coordinated across multiple cities. Now, here's the stat that you should be seeing. Quantis backlog now sits at over 48 billion. That's bigger than the entire market cap of most of the companies on this list combined. Management also raised fullear revenue guidance into the $ 35 billion range. And a major portion of that backlog is locked in through long-term master service agreements rather than just a one-time project. That's the kind of revenue visibility that locks in years of forward earnings. The data center segment now represents around 10% of total backlog. And Quant has publicly disclosed that strategic partnerships with American Electric Power structured to support AP's 72 billion capital plan, including the transmission infrastructure that serves the data center market. Quanta is fundamentally a scale story. Revenue compounded at 20% annually for 5 years and the stock has returned more than seven times since 2021. For me, Quanta is the credibility anchor of the entire AI buildout. If hyperscalers are spending real money on real construction, Quanta catches a meaningful share of all of those dollars.
Now, we move on to tier two, and that's the equipment that goes inside the buildings. switch gear, transformers, power distribution, the hardware that every data center needs. Whether the racks are full of Nvidia chips or AMD chips, it just doesn't matter. The first name on this layer is Powell Industries, where they design and build the high voltage switch gear that controls the flow of electricity inside the data centers, utility substations, large industrial sites, and LG export facilities. Their CEO started strategically aligning the company with hyperscalers and collocation operators around 2022 and the financials reflect a complete operational reinvention. Powell was a small cap nobody in early 2021 that has run so far and so fast that the companies had to do a three for one stock split. And here's what really tells the story. In fiscal 2021, Powell was essentially just break even and their operating margin was nearly zero. 4 years later, that operating margin has expanded to nearly 20% while revenue more than doubled. That is exactly what an industrial pivot to data centers looks like when the execution just hits everything just right. In their most recent quarter confirmed the trend, Powell ended with a record $ 1.6 billion backlog and new orders are up 63% year-over-year. That happens to be the highest quarterly bookings in over 2 years.
Next is Eaton Corporation where they make the power management equipment that goes inside every major data center on Earth. This happens to be the switch gear, the busways, UPS systems, and the electrical distribution gear. The products that get specified in every Greenfield hyperscaler build. Eaten is by far the largest name on this layer at over 167 billion in market cap. And total data center orders were up over 200% year-over-year in their most recent reported quarter. and Electrical America's backlog reached a record 13.2 billion. Management has noted that data center construction backlog represents over a decade of normalized build rates. Eden also recently acquired Boyd, a leading liquid cooling company expected to generate $1.7 billion in annual sales. That's with over 80% coming just from data centers. Their operating margin has gone from 19% to over 24% in just 5 years and earnings per share has roughly tripled. The stock has returned more than three times since 2021, and that's a rare compounding for a $167 billion industrial company, especially with the data center momentum still building.
Now, we move on to tier three, which is cooling. The new generation of NVIDIA AI racks pull 132 kW of power, and that's nearly 10 times what a traditional server rack pulls. In the most blunt of terms, air cooling just can't handle that. Liquid cooling is the only answer. And first on this cooling list is Modine Manufacturing, where they design and build the liquid cooling distribution units, immersion systems, and the thermal management gear that the new generation of AI racks just simply can't run without. They happen to be the smallest pure play in this video. And the financial pivot from being a boring auto supplier to AI cooling specialist is the most dramatic transformation in the entire basket. Just look at where Modine was just 5 years ago. Fiscal 2021 was a complete disaster. Operating margin came in at a negative 5.4%. I mean, they were literally on the chopping block for being taken over. And then, of course, four years later, operating margin has flipped to a positive 11%. And the earnings per share has moved from a loss of more than $4 to a healthy gain. And the stock has returned roughly 17 times. Clearly, the data center pivot is what's carrying the company. Climate Solutions revenue grew 51% year-over-year in their most recent quarter, while data center sales alone were up 78%.
Next on the cooling list is Carrier Global, where they make commercial and data center cooling systems at scale as one of the largest commercial HVAC companies in the world. Carrier is a different kind of pick than the rest of the names on this layer. The 5-year return is roughly 55% and it's the lowest in this entire video. An operating margin has actually contracted from the high teens to under 10% as they integrated the viceman acquisition and they divested the fire and security business altogether. So what even earns Carrier a place on this list? You just need to simply look at their most recent quarter. Global data center orders were up over 500% year-over-year. And of course, another thing I like to look at is the data center backlog, which now fully covers their $ 1.5 billion fullear sales target with room to exceed it. Their integrated quantum leap data center offering, which combines precision chillers, coolant distribution units, and building management software, has booked roughly 300 to$400 million in orders within its first year. Carrier is now publicly working with Nvidia on cooling designs for the next generation AI infrastructure.
Now, we'll move on to the fourth tier, which happens to be the hidden layer. This is the fiber, the cooling components, and the behind the meter power that none of this works without. But that also sits just one or two steps removed from the GPU. And they don't show up in most analyst models. The first name on the hidden layer is Beldin, where they make the data center fiber, the high-speed interconnects, and the cabling that physically connects the GPUs inside the hyperscaler campuses. Their broader business spans industrial automation, broadband, and smart building infrastructure. Beldon's been a steady recovery story since 2020. revenue compounding at roughly 9% annually. Operating margin expanding from the high single digits to over 11% and the company moving from a net loss to solid profitability. Solid execution, but not the fireworks of the small cap names of the earlier layers.
Now, here's where it gets a little bit interesting. On the same day that Beldin reported their most recent earnings, they announced that they're acquiring Ruckus Networks for $1.8 billion in cash. That's roughly 40% of the Beldin's market cap going into a single acquisition. Ruckus adds Wi-Fi 7 access points, enterprise switches, and cloud management networking to a company that was already running the fiber backbone. In just one simple move, Beldon goes from cable company to a full stack networking infrastructure play. The stock got hit pretty hard on the deal because the market doesn't yet trust this integration. While every other name in this video is making new highs, Beldon is sitting at one of the cheapest forward multiples of the basket at 15.7 times its forward earnings. Despite delivering revenue up over 11%. This is one you just don't jump into, but you definitely want to watch it and do a little extra research.
Next on the hidden layer is Mueller Industries, where they're the only vertically integrated manufacturer of copper tube, brass rod, and forgings in North America. They supply critical cooling components into the plumbing, HVAC, refrigeration, and industrial systems that the data centers absolutely rely on. The most recent quarter was extraordinary. Net income jumped nearly 52% and diluted earnings per share climbed to $216, while gross margin expanded nearly 300 basis points. Mueller also raised their quarterly dividend by 40%. The sixth consecutive year of double-digit dividend increases. That is the kind of capital return move that you make when management is highly confident in forward cash flow. Here's what really tells their story. Revenue compounded at 12% annually for 5 years while their operating margin roughly doubled to over 21% and earnings per share multiplied more than five times. The stock returned more than 6 and a half times since 2021.
Last on the hidden layer is Bloom Energy, where they design and manufacture solid oxide fuel cells that generate electricity on site behind the meter and without waiting for the utility grid to even be connected. They're the only public pure play on behind the meter power generation for data centers in the entire market. Their most recent quarter came in well ahead of consensus. Revenue was up 130% year-over-year, gross margin reaching 31.5%. But here's where they really shine. Oracle agreed to procure up to 2.8 gawatts of Bloom's fuel cell systems with 1.2 gawatt already contracted and more than half of Bloom's data center backlog now comes from hyperscalers, Neoclouds, and collocation providers beyond Oracle. Total contracted backlog sits at roughly $20 billion, and they're scaling manufacturing capacity at hundreds of megawatts per quarter.
Those Patreon stocks that I open with proved that this AI buildout paid off once it began to scale. These 10 names are how I would carry it into the next chapter and how I'm prioritizing them. Now, I'm not saying that I would invest into all of these, but I am going to say this is how I'm going to prioritize them in a ranking order. So, how would I actually invest $100 across this group if I were starting fresh?
I'd start with $30 into tier one, which is the contractors. They have the most direct hyperscaler exposure and the cleanest balance sheets on this layer. Then, I'd put $25 into tier two, the equipment makers. This layer combines world-class capital efficiency with the largest single backlog on the list. Then I'd put $20 into tier three, which is cooling. It's the highest growth subsegment of the entire AI buildout. And last, I'd put $25 into tier 4, the hidden layer. This is where the rare value names and the contrarian setups begin to live. My point of view is that I'm going to put higher allocations going to the names with the cleanest balance sheets and the most direct hyperscaler exposure. And of course, as I trickle down, the smaller allocations go to the higher volatility names, just like most of my investments.
I have to give a big shout out to the Patreon group because they surfaced so many of these companies that allowed me to dig even further. And hopefully you got some value from that. And as always, thank you so much for watching.