The Mag 7's $10 Trillion Blind Spot

The article argues the key “fuel” bottleneck for the Magnificent Seven’s AI boom is electricity: analysts expect ~40% of planned 2026 data centers to be delayed into 2027 due to insufficient power equipment and grid connectivity. It highlights upside to power “golden rivet” suppliers, including Fluence Energy (FLNC) with revenue expected +48% this year and +24% in fiscal 2027 despite high volatility, Legrand (LGRDY) trading at ~22.5x forward earnings with data-center sales rising above 32%, and Constellation Energy (CEG) showing a +6% weekly rebound after an earnings beat and a proposed 30%+ return potential toward ~$320 vs ~$265.

Analysis

The durable trade is not “AI” in the abstract, but scarcity rents in the grid-interconnect stack. The first-order winners are electrical equipment suppliers, nuclear generation, and any vendor with backlog already locked in; the second-order winners are the firms that can monetize expedited hookups and higher utilization without waiting on new generation buildout. That argues for ETN and CEG as higher-quality expressions, while FLNC is more of a financing-and-execution trade than a pure demand story.

The market is likely still underpricing how long lead times can persist, but it may be overpricing the immediacy of the upside. A true supply response for transformers, turbines, and transmission gear is measured in quarters to years, yet equity re-rating can happen sooner if order books keep compounding; the converse is also true if utilities slow approvals or if data-center project timing slips and backlog stops converting. That makes the next 1-3 quarters the key catalyst window: watch order growth, margin expansion, and forward guidance, not the promotional narrative.

The contrarian miss is that power scarcity can become an AI capex tax on the hyperscalers themselves. If MSFT, AMZN, GOOGL, and META cannot translate chips into deployed capacity, return-on-capital expectations for the entire AI stack can compress, eventually pressuring the semiconductor multiples that have already discounted perfect demand. The thesis is falsified if grid equipment lead times normalize faster than expected, if nuclear and gas additions accelerate, or if utility regulators block the direct-contract economics that underpin the premium story for CEG and peers.

Key Decisions for Investors

• Long ETN on pullbacks; prefer a 6-12 month horizon. Best risk/reward is backlog-driven compounding with lower execution risk than pure-play battery/storage names.
• Long CEG as the highest-quality scarcity beneficiary; expect a slower grind higher over 6-18 months if direct AI PPAs continue to convert. Falsify on weaker-than-expected earnings revisions or regulatory pushback on contract economics.
• Use FLNC only as a small tactical position or call-spread structure, not a core long. The upside is large if battery attachment rates surprise, but the business is still too dependent on sentiment and capital markets for large sizing.
• Fade enthusiasm in NVDA/MSFT on rallies if power constraint commentary worsens. Best expression is a relative-value short basket of MSFT/NVDA against long ETN/CEG, with the thesis that deployable watts matter more than shipped chips over the next 2-4 quarters.
• Avoid chasing LGRDY unless liquidity is sufficient; if accessible, it is the cleaner valuation arbitrage versus U.S. electrical peers, but the catalyst depends on North American mix and continued data-center order momentum.