The AI Energy Bubble: Biggest Opportunity or Biggest Trap?

AI is running into a problem even the most powerful chips can't solve: not enough electricity. The power shortage is real. Today on Dumb Money: Are AI energy stocks the biggest investment opportunity of the decade… or has Wall Street already priced in the future? Yesterday, Wall Street got two very different answers to the same question. How much is AI's electricity problem actually worth? Vertiv reported revenue growth of 61%, orders more than doubled, and its stock jumped 16%. Meanwhile, GE Vernova reported $29 billion in orders, a backlog of $197 billion. They raised their guidance, and that stock went down. Bloom Energy, after getting crushed 20% in two days last week, was back up 18% in three days. AI has a problem that even the world's most powerful chips can't solve. Not enough electricity. Morgan Stanley now estimating AI global technology power shortfall of 38 gigawatts through 2028. That's roughly the output of 38 large nuclear reactors that just don't exist. The big question: who's gonna actually make money solving the electricity problem? Today on Dumb Money, the AI energy bubble. Is it one of the greatest investment opportunities of the last decade or is it a trap? We break down why Chris is loading up on even more Bloom Energy along with Amazon (reporting after hours tonight), the circular financing argument that Jon Stewart and Ed Zitron are pushing (and why it's spectacularly wrong), why AI succeeding spectacularly is actually threatening Google's search revenue (Chris called this a year ago and got roasted), the AI mode problem that's killing traditional search monetization, why Reddit is sitting on the most valuable data licensing deals that are coming up for renewal mid 2027 (Google's $60M deal and OpenAI's $70M deal are jokes that should be billions), the Amazon earnings trade tonight and why Chris is deleveraging live during this episode (CapEx guidance could crush the stock even if everything else is perfect), why the market has never been dumber and you cannot trust it to react rationally to anything right now, the Bloom Energy thesis and what would actually make Chris sell (if they can't scale deployments with hyperscalers, he's out), why having an energy basket right now makes sense (Vistra, Constellation, GE Vernova alongside Bloom), the multi-decade power purchasing agreements that prove someone is spectacularly wrong about AI being a bubble, the Take Two and GTA 6 narrative shift (democratized game development through AI is now the bigger threat than whether the game is a hit), why Chris trimmed 75% of his Take Two position when the thesis changed, Tesla Model Y full self driving and why Chris just bought two Teslas after one demo drive in Dallas rush hour traffic, and the Theranos documentary that Chris has seen twice and is the movie event of the year. If you're trying to figure out who actually wins when AI needs 38 nuclear reactors worth of power that don't exist, this is the edge. Chapters: Chapters 00:00:00 Intro: AI energy bubble - biggest opportunity or biggest trap? 00:01:48 Everyone is an imbecile: The Jon Stewart AI denial problem 00:06:48 The circular financing myth and why AI adoption takes time 00:11:17 Google's AI mode problem: Success that threatens their business 00:14:52 Amazon's agentic shopping dilemma and why Chris is all in 00:21:24 The CapEx fear cycle and why investors don't trust CEOs anymore 00:32:59 The AI energy trade: Bloom, Constellation, Vistra and the power play 00:35:40 Bloom Energy: The misunderstood technology company everyone doubts 00:45:53 Reddit's data gold mine and the AI licensing renegotiation coming in 2026 00:59:22 Take-Two reality check: Why Chris sold 75 percent of his position 01:07:26 Amazon earnings anxiety: Deleveraging live and the market irrationality problem 01:22:28 Tesla FSD conversion: Dave's full self-drive revelation and the two Model Y purchase