Forget XLU: These 3 Utility Dividend Stocks Offer More Exposure to AI

If you own the Utilities Select Sector SPDR Fund (NYSEARCA:XLU), you bought it for a reason: cheap, diversified access to America’s regulated utilities, a low beta, and a dividend check that arrives whether the market rallies or rolls over. XLU has done that job for two decades, and its $23.1 billion in net assets says plenty of investors agree. But XLU’s roughly 2.7% yield and 3.78% one-year return understate what is happening inside the sector right now, and three of its own top holdings are the reason.

What XLU Actually Owns, and Why It Dilutes the Story

XLU is top-heavy, with NextEra alone making 12.9% of the fund. The next four positions, Southern, Duke, Constellation, and AEP, make up another roughly 25%. The remaining 60% is a long tail of water utilities, gas distributors, and slower-growing regional names like Atmos, CenterPoint, Ameren, and PPL. That tail is what keeps XLU’s yield near the sector average and its earnings growth close to GDP. The AI and data-center demand supercycle is not evenly distributed across those 30-plus holdings. It is concentrated in a handful of them, and you can own those directly.

Southern Company: The Southeast Data-Center Magnet

Southern Company (NYSE:SO | SO Price Prediction) yields 3.35% on a forward dividend of $3.04, meaningfully above XLU. More importantly, Q2 adjusted EPS came in at $1.13, and management said data center usage was 55% higher than the prior-year quarter. Georgia Power just signed a 3.2 gigawatt, 25-year contract with OpenAI, and total contracted large-load agreements now exceed 17 gigawatts by the mid-2030s, backed by roughly $21 billion of collateral. Southern has raised its dividend for more than two decades, most recently to $0.76 per quarter. Trading at a 19x forward P/E, it captures the Southeast growth story XLU only partially expresses.

Duke Energy: The Compounding Dividend Machine

Duke Energy (NYSE:DUK) yields 3.53% and just raised its quarterly payout to $1.085, marking over 20 years of consecutive annual dividend increases. Duke beat consensus for a fifth straight quarter with Q2 adjusted EPS of $1.43, reaffirmed 5% to 7% long-term EPS growth through 2030, and guided to the top half of that range starting in 2028. CEO Harry Sideris said Duke is “deploying more than $1 billion per month” in regulated capital, with 7.8 gigawatts of signed data-center agreements and $5 to $10 billion of upside to the current five-year capital plan. At a 18x forward P/E, you are paying utility multiples for a growth ramp that XLU averages away.

American Electric Power: The Transmission Toll Road

American Electric Power (NASDAQ:AEP) is the pure transmission play. Commercial load in its vertically integrated segment jumped 14.9% in Q2, and management raised 2026 EPS guidance to $6.25 to $6.55. The company’s $78 billion five-year capital plan is expected to produce nearly 11% rate-base CAGR, with contracted load additions now at 69 gigawatts through 2030, of which 45 gigawatts sit in ERCOT. AEP yields 3.08% and targets 7% to 9% annual earnings growth, with an expected CAGR above 9% through 2030. That is roughly double what the average XLU holding will deliver.

Tradeoffs You Are Accepting

A three-stock sleeve of SO, DUK, and AEP concentrates you in regulated electric utilities in the Southeast, Carolinas/Midwest/Florida, and 11-state AEP footprint. You lose XLU’s exposure to water, gas distribution, and independent power producers like Vistra and Constellation. You take on single-state regulatory risk, and you have to rebalance yourself. With the 10-year Treasury at 4.67%, none of these yields dominate risk-free income, so the case rests on dividend growth plus rate-base compounding, not on income alone.

How to Think About the Switch

In a tax-advantaged account, rotating out of XLU into an equal-weight basket of SO, DUK, and AEP is a clean trade: no capital-gains friction, higher blended yield, and direct exposure to the load-growth names already inside XLU. In a taxable account, weigh embedded gains from XLU’s 139% ten-year run before selling; a partial rotation, funded with new capital rather than a full liquidation, often makes more sense. If you want XLU’s diversification and defensive character above all else, stay put. If you want the AI power-demand tailwind expressed at full strength, the three names are already sitting in your ETF, just diluted, and the same buildout is pulling in the cooling, networking, and equipment suppliers we profiled in a free report on seven AI infrastructure stocks that aren’t chipmakers.

Contact [email protected] for any questions or corrections.