Listen to the audio version of this article (generated by AI).
Tom Yeung here with your Sunday Digest.
Imagine that you own a Formula One race car.
It has a 1,000-horsepower engine, a carbon-fiber body, and tires with so much grip that they can almost stick to the ceiling. You hire the world’s best driver and spend millions tuning every part for speed.
Then race day arrives… and there is no fuel.
You may own one of the finest machines ever built, but it’s a very expensive paperweight without fuel.
That is the blind spot hiding inside the Magnificent Seven’s AI boom.
Amazon.com Inc. (AMZN), Microsoft Corp. (MSFT), Alphabet Inc. (GOOG), Meta Platforms Inc. (META), Nvidia Corp. (NVDA), Apple Inc. (AAPL), and Tesla Inc. (TSLA). These are all remarkable companies that have built some of the top F1 vehicles of the AI Revolution.
- Data centers…
- Custom chips…
- Advanced AI models…
And they all suffer from the same blind spot: They don’t produce the “fuel” that allows their multibillion-dollar AI investments to run. Instead, this role is filled by chipmakers… electrical utilities… data center construction firms… and other behind-the-scenes producers making the essential ingredients for the AI Revolution.
InvestorPlace Senior Analyst Eric Fry calls these components “Golden Rivets.” And in his recent free broadcast, he reveals why they are creating a $10 trillion opportunity that is even better than the one offered by the Magnificent Seven.
Today, I’m going to reveal one of these Golden Rivets and three top picks that are churning it out. To find out the rest, you’ll have to watch Eric’s Market Shock presentation here.
One Golden Rivet of the AI Revolution
The Golden Rivet I’m going to discuss is everywhere in our modern lives. But you can’t touch or taste it… and you’re not supposed to see or hear it (unless something has gone very, very wrong).
I’m talking about electricity… one of the greatest bottlenecks of the AI Revolution. And over the next several years, we are going to hear a lot about this invisible force.
That’s because electricity is expected to be the No. 1 reason for data center project delays. Analysts currently forecast that around 40% of all planned data centers for 2026 will get pushed into 2027… and the cause will be either the lack of power equipment (transformers, battery systems) or the inability to connect data centers to the main electrical grid. That will push construction planned for 2027 into 2028… and so on.
In other words, the Magnificent Seven companies are building massive power-hungry data centers, but they have nowhere to plug them in.
That’s going to create a bonanza for power utilities and electrical component makers that supply AI data centers. In fact, some power companies have sold out their production through 2030. High-voltage transformers and heavy-duty gas turbines are now even harder to obtain than the highest-end Nvidia chips, simply because there are none available.
Now, here are three electricity Golden Rivet companies that should benefit, from the riskiest to the least risky…
The Moonshot Bet
You might recognize my first pick from a Sunday Digest last year when the company still traded in the $8 range:
Fluence Energy (FLNC).
Fluence is a utility-scale energy storage provider. Think of it as storing power in a bottle: Fluence charges massive arrays of batteries when too much electricity is generated, and then dumps it back into the grid when it is needed.
Demand for Fluence’s services has been incredible. Virtually every AI data center needs battery backup systems, because gas turbines cannot spin up fast enough to keep up with sudden demand spikes. Fluence’s batteries give that extra jolt. Ask any child who has ever licked a 9V cell.
The popularity of renewables like solar and wind power has further charged demand for Fluence’s products. After all, AI data centers still need power when the sun doesn’t shine and the wind doesn’t blow. Many energy grids even let data centers jump in line for grid connections if they have on-site batteries. Fluence’s revenues are expected to rise 48% this year, and then another 24% in fiscal 2027 – some of the fastest growth rates in the business.
Keep in mind that the share prices of this promising startup sometimes trade wildly. The stock rose as high as $33.50 in January – a 4X increase from my July 2025 recommendation – before plummeting 50% back into the low-teens range. Fluence will remain unprofitable until 2027, so its stock price will depend on investor mood. Earlier this week, the stock plummeted 26% in after-hours trading before opening back up to almost where it started.
But the same volatility is now giving investors a second chance to buy shares. The stock is now trading under $14, and its flip from negative profits to positive next year should act as the catalyst that buy-and-hold investors need to all pile in all at once.
The Cheapest Entry Point
A less volatile way to play the electricity Golden Rivet is Legrand SA (LGRDY), a French company that builds the electrical plumbing inside data centers.
This includes components like:
- Busways. Overhead power highways that feed AI servers
- Breakers. Heavy-duty protection against power surges
- Power distribution units. Complex power strips for individual servers
- Monitoring equipment. Measurement equipment to identify potential failures
In other words, Legrand moves electricity from a data center’s main power electrical room to each server and manages the things that can go wrong. Data centers now make up more than 32% of sales, up from 15% in 2023.
Legrand is one of the cheapest companies in the business because it primarily trades on the Euronext Paris, where it is valued like a European wiring company. Shares trade at just 22.5X forward earnings. That is far lower than Legrand’s American-traded peers, including Eaton Corp. Plc (ETN) at 30X and Rockwell Automation Inc. (ROK) at 31X. These U.S. firms actually have lower data center revenue shares.
In addition, Legrand has already raised its 2026 guidance twice and has acquired the capabilities it needs for a 2028 industry-wide switch to a new 800-volt standard. That should keep driving sales higher. And as for that whole “European wiring” caricature… roughly half of its sales now go to North America.
That means we should begin to see a convergence between this French company and its American peers. My base case is for a 30% return, and possibly higher if insatiable AI data center demand keeps pushing expected earnings higher from here.
The Quiet Compounder
Finally, the power industry’s bluest of blue-chip award goes to Constellation Energy Corp. (CEG), America’s largest producer of nuclear energy. The Baltimore-based firm operates roughly two dozen nuclear reactors, which generate enough power to supply 16 million typical American homes. It also has a large portfolio of wind, solar, natural gas, and hydroelectric plants, which can supply another 11 million homes.
Nuclear energy is a particularly excellent source of electricity for AI data centers. Reactors are very low-cost once they are built and provide the kind of cost stability that tech companies prize. Fuel makes up less than 20% of a nuclear power plant’s cost, compared to 65% to 80% for gas power plants.
Nuclear plants also produce electricity 24/7, giving them an advantage over solar and wind, which require the expensive batteries (often from Fluence) to properly run.
That’s made Constellation’s shares slightly more expensive than its peers, especially those that focus on gas power. CEG trades at 23X forward earnings, compared to a 19.7X sector average. Constellation also holds somewhat high debts because of a 2025 acquisition of another power producer, Calpine Corp.
However, the premium could be worth it for three key reasons:
- De-rating. Constellation’s shares have fallen 33% since its October 2025 peak, putting prices back at long-term averages on a P/E basis.
- Direct deals. Constellation has increasingly fueled its growth with direct contracts with AI data centers, which bypass pricing caps set by regulators. Many of these projects are due to get switched on in the next year.
- Guidance. Earnings estimates have mostly trended higher, and the company has an excellent history of beating these estimates. Shares rose 6% this week after Constellation announced another earnings beat.
Constellation’s stock should not rise as quickly as Fluence or Legrand. Its size and stability make fireworks less likely. Nevertheless, I still expect the stock to grind higher from around $265 to $320 in the next year or so, making it an appropriate bet for risk-averse investors.
The Other Golden Rivets of the AI Revolution
In late-2025, Microsoft CEO Satya Nadella revealed that his firm had AI chips sitting on shelves because the company didn’t have enough power to install them.
“You may actually have a bunch of chips sitting in inventory that I can’t plug in,” Nadella said in an online interview. “In fact, that is my problem today.”
This is Microsoft we’re talking about… a $3.7 trillion firm. And they couldn’t find enough electricity to run the chips they had bought.
It turns out Microsoft is not the only Mag 7 company suffering from AI bottlenecks. Nvidia has been forced to delay production of its highest-end chips because its suppliers couldn’t keep up. Apple has been forced to raise prices of iPhones from a “a hundred-year flood” of memory chip shortages. And Elon Musk’s xAI has turned to using dozens of gas turbines without permits to power its “Colossus” supercomputer in Memphis because Tesla’s “Megapack” systems can’t provide enough juice.
Everywhere you look, there are bottlenecks in the AI buildout.
That’s where Eric’s Golden Rivets come in. These are the firms producing these components in such short supply.
In his free Market Shock presentation, he’s revealing his top picks, including 15 free stocks, that he believes will profit from the growing AI shortages.
Until next week,
Thomas Yeung, CFA
Market Analyst, InvestorPlace
