Want to Make a Fortune From the AI Power Boom? Buy These 2 Industrial Stocks
Source: Nasdaq

Goldman Sachs estimates U.S. data-center power demand could approach 108GW by 2030, underpinning a bullish outlook for utility American Electric Power and data-center infrastructure supplier Vertiv. AEP has 69GW of contracted load additions through 2030, a $78B five-year capex plan, and raised 2026 operating EPS guidance to $6.25-$6.55 while targeting more than 9% annual operating-earnings growth through 2030. Vertiv guided for 2026 sales of $14B, roughly 31% organic growth, supported by a $15B backlog, but its approximately 43x forward P/E leaves the shares vulnerable if data-center spending or order timing weakens.
Analysis
The AI-power theme splits into two distinct earnings profiles: AEP monetizes load growth through a regulated-rate-base model, while VRT monetizes it through equipment content and execution. AEP's upside depends less on aggregate data-center demand than on timely regulatory recovery, allowed ROE, and whether incremental generation/transmission is financed without pressure on credit metrics; a higher-rate environment can absorb much of the nominal rate-base growth through equity issuance and interest expense. The key 1-3 month catalyst is clarity on large-load interconnection treatment and cost allocation, while the 6-18 month issue is whether hyperscalers accept tariffs that protect residential customers from stranded infrastructure costs.
VRT has stronger near-term operating leverage but materially greater expectation risk. The important variable is not data-center construction volume alone: liquid-cooling and high-density power architectures raise VRT's revenue per megawatt, but also invite share capture by Eaton (ETN), Schneider Electric, Johnson Controls (JCI), and private cooling suppliers. At a premium earnings multiple, even a modest order-to-revenue conversion delay or backlog normalization can drive multiple compression before fundamentals deteriorate; management pipeline commentary is not equivalent to cancellable-order protection.
Consensus is likely underweighting the bottleneck shift from chips to grid connection, permits, transformers, and gas-turbine availability. That favors regulated transmission owners and electrical-equipment suppliers with installed capacity, but it also means a meaningful portion of announced compute projects may slip rather than disappear. AEP is the lower-beta way to express delayed-but-persistent load growth; VRT is a higher-beta expression of accelerated rack-density adoption and should be owned only against explicit evidence of sustained book-to-bill and margin expansion.
AllMind Terminal
AI-powered research, real-time alerts, and portfolio analytics for institutional investors.
Request TrialMarket Sentiment
Overall Sentiment
moderately positive
Sentiment Score
0.48
Ticker Sentiment
Key Decisions for Investors
- Initiate a 6-12 month long AEP / short XLU pair at equal beta weights. The trade isolates AEP's incremental load and transmission exposure from broad utility-rate sensitivity; reassess if regulatory treatment reduces allowed returns, net debt-to-EBITDA deteriorates, or large-load commitments fail to convert into approved capital projects.
- Maintain VRT as a tactical long only on post-earnings confirmation that orders, backlog conversion, and gross-margin guidance remain intact; use a 10-15% position stop or defined-risk put spread because valuation leaves limited tolerance for a single-quarter delivery slip. A 15-25% upside case requires continued high-density cooling mix gains, while downside is sharper if order growth decelerates materially.
- For a more balanced AI-infrastructure basket, pair long VRT with long ETN rather than adding broad semiconductor exposure. ETN provides electrical-distribution and power-quality exposure with less dependence on cooling-specific execution; reduce VRT relative weight if ETN's order growth accelerates while VRT's book-to-bill falls below 1x.
- Set an event alert around ERCOT and other state commission interconnection/cost-allocation rulings over the next 1-3 months. A decision requiring utilities or legacy ratepayers to bear a disproportionate share of dedicated data-center infrastructure would weaken the AEP thesis and favor suppliers such as ETN and VRT, which can still sell into projects regardless of utility return treatment.
More News
- AI funding fears rattle Wall Street. Here's how we're adjusting our portfolio
- Nvidia's role as AI financier faces first big test as neoclouds take market hit
- Goldman says these top tech stocks have plenty of upside ahead of earnings
- Will AI expose the rickety infrastructure of the internet?
- Apax, Nordic Capital and Dalmore circling UK supplier JLA Group
- Block’s new voice model sounds eerily human