AI Energy Stocks: Jon Najarian’s Power Thesis, Explained

AI energy stocks have become their own sector in the last two years, and the reason is a simple collision: AI data centers need power faster than the grid can deliver it. Jon Najarian’s Infinite Power presentation for Banyan Hill is one of the better-evidenced pitches in this wave, because almost every structural claim in it checks out against published sources. His core framing is the memorable one: “You can pay for the grid that’s failing or you can profit from the grid that’s replacing it.”

Our full Infinite Power teardown covers the presentation claim by claim. This piece explains the AI energy thesis and how to evaluate it.

The Demand Story, Verified

The numbers Najarian builds on are real:

  • The ASCE’s 2025 Infrastructure Report Card gave U.S. energy infrastructure a D+, down from C- in 2021, with 70% of power transformers 25 years or older and 60% of circuit breakers 30 years or older
  • The University of Texas Energy Institute’s roughly $5 trillion grid replacement estimate is genuine
  • The DOE’s warning that blackouts “could increase by 100 times in 2030” under certain conditions is a real report finding
  • BlackRock’s 148-gigawatt data-center demand figure reflects published estimates
  • Grid interconnection timelines run four to ten years, while an AI data center needs power in months

The emblematic case is xAI’s Memphis Colossus: 150 megawatts needed, about 8 available from the local utility, and a solution built from truck-delivered gas turbines standing up roughly 400 megawatts in about 122 days. That is the behind-the-meter pattern, and more than 60% of new data center developers are now planning their own generation according to Bloom Energy’s data center power report, a figure the presentation cites accurately.

The Supplier Pattern

Najarian’s proof that suppliers capture the value is a three-stock highlight reel, and each number is real: Vertiv returned 1,200% over five years as the data-center power-and-cooling supplier, Generac returned 850% in under three years through the generator surge, and Quanta Services compounded 8,600% over its life as the grid contractor. His summary of the pattern: “The biggest winners in any boom aren’t the companies on the front pages. They’re the suppliers.”

That is a favorable selection, and he presents it as one. But the pattern has a structural reason behind it: the AI companies themselves are expensive and widely held, while the equipment suppliers start smaller and re-rate when order books turn.

The Pick and Its One Big Risk

The actual recommendation is a Houston-based company that leases gas-turbine “power pods” under 10-to-15-year contracts, what the pitch calls “power as a service.” The clue profile, a $6 billion market cap, NYSE listing, roughly 2 gigawatts under contract across three hyperscalers, and revenue that doubled from just over $300 million in 2023 past $600 million in 2024, sits behind the $79 Market Rebel Report trial.

The risk the presentation itself concedes: “in 2024, this company derived 96% of its Power Pod revenue from a single customer. And that customer was Elon Musk’s XAI.” Najarian’s counter is that the company “still owns over one gigawatt of new contracted hyperscaler capacity coming online through 2028” regardless. Both things are true, and customer concentration of that scale deserves top billing in any risk assessment of the theme.

Reading the Institutional Names the Right Way

The presentation lists five institutional holders as evidence the smart money has arrived: Vanguard with over 900,000 shares, Millennium Management with more than 660,000 shares, plus BlackRock, Morgan Stanley, and VanEck. That list deserves one honest footnote. Vanguard and BlackRock hold nearly every S&P 500 constituent through index funds, so their presence in a register is not the same as an active bet. Millennium is the different case: it is an active hedge fund, and a 660,000-share position is a decision somebody there made on purpose. That distinction matters more than the raw names.

The presenter’s own history with institutional flow is worth knowing too. Jon Najarian spent 16 years as a CNBC regular, sold his firm Mercury Trading to Citadel in 2004, and built the HeatSeeker system for tracking unusual options activity with his brother Pete. The book that came out of that work, Follow the Smart Money, is the backbone of his SMART money criteria: Size, Much larger than average, Aggressive, Reason, Timing. Whether or not the Houston pick becomes the asymmetric winner he calls the highest conviction call of his career, the framework of watching what large, deliberate money does before the crowd notices is a genuinely useful discipline, and it is the honest core of the whole AI energy pitch.

How to Evaluate Any AI Energy Stock

The framework the presentation suggests, whether you subscribe or not, is a fair one:

  1. Is the company exposed to behind-the-meter demand, or to grid build-out? Both grow, on different timelines.
  2. Does revenue recur (leases, service contracts) or arrive one-time (equipment sales)? “Power as a service” converts a boom into an annuity.
  3. Is the customer base concentrated? The Memphis story is dramatic precisely because one company dominated it.
  4. Is the entry valuation already reflecting AI hype? The pattern stocks re-rated before the mainstream arrived; a theme this covered offers thinner edges than 2020 did.

For the same demand story from other publishers, our data center power demand piece and Homestacks coverage are useful comparisons, and the full Infinite Power teardown covers the Vertiv precedent in detail.

The honest summary of the AI energy thesis: the shortage is quantified and real, the supplier pattern has real historical examples, and the specific picks each carry their own concentration and valuation risks. Najarian’s framing of his own conviction is calibrated about right: “I’m not saying it’s certain. I’m saying it’s the highest conviction call that I’ve made in years.”

NewsletterVetter is an independent publication. We receive compensation from some of the services we review through affiliate links. Nothing on this site is investment advice. Always do your own research.