© DenisTangneyJr / E+ via Getty Images
The AI data-center power crunch has turned utilities from bond proxies into growth stocks, and two funds sit at the center of that trade: the Utilities Select Sector SPDR Fund (NYSEARCA:XLU) and the Vanguard Utilities Index Fund ETF (NYSEARCA:VPU). They look nearly identical on a screener, but they make different bets on who actually captures the hyperscaler power boom. One concentrates on a handful of mega-cap operators with signed data-center deals. The other spreads across the entire utility complex, including mid-caps and water utilities that have almost no AI exposure at all.
What each fund is actually betting on
XLU tracks a subset of S&P 500 utilities. It holds roughly 30 names, cap-weighted, with NextEra Energy (NYSE:NEE | NEE Price Prediction) at 13.59%, Constellation Energy (NASDAQ:CEG) at 6.11%, and Vistra (NYSE:VST) at 3.36%. The top five holdings alone account for 39.30% of assets. That is a concentrated bet that the largest generators, especially the nuclear-heavy independents signing hyperscaler contracts, will keep repricing higher as AI load forecasts climb.
VPU tracks the MSCI US Investable Market Utilities 25/50 Index, which pulls in mid- and small-cap utilities alongside the large caps. That pushes the fund past 65 holdings and dilutes the AI-power names. Constellation and Vistra still appear near the top, but their weights are meaningfully smaller because the index reaches down into regulated water utilities, smaller gas distributors, and regional electric operators that have no data-center pipeline. VPU is a bet on the utility sector as a whole. XLU is a bet on the biggest names inside it.
Where the difference showed up
Concentration paid, but only barely. From January 3, 2023 through July 6, 2026, XLU returned 42.75% against VPU’s 42.05%. Over five years the gap widens slightly, with XLU up 62.63% versus VPU’s 60.64%. Year to date in 2026, XLU is up 7.53% compared with 7.11% for VPU.
The takeaway: XLU’s overweight to Constellation and Vistra during the 2023 through 2025 AI-utility rally added return, but VPU’s broader roster of regulated utilities kept it close because those names also re-rated on the same theme. The concentration edge is real. It is not enormous.
The practical comparison
| Metric | XLU | VPU |
|---|---|---|
| Expense ratio | 0.08% | 0.09% |
| Holdings | ~30 | 65+ |
| Top holding weight | 13.59% (NEE) | Lower (broader index) |
| Combined CEG + VST weight | 9.47% | Smaller due to dilution |
| YTD 2026 return | 7.53% | 7.11% |
| Five-year return | 62.63% | 60.64% |
The one-basis-point expense difference is trivial. What matters is that XLU’s construction rules force it to concentrate in the exact names driving the AI-power narrative, while VPU’s methodology deliberately spreads exposure into utilities the theme does not touch.
The verdict
For an investor buying utilities specifically to ride the AI data-center power buildout, XLU is the cleaner expression. Its 9.47% combined weighting in Constellation and Vistra, plus a heavy NextEra position, gives direct exposure to the generators signing hyperscaler contracts and pursuing nuclear restarts. VPU is the better choice for an investor who wants defensive utility exposure without making a concentrated bet on independent power producers. Its water and mid-cap holdings dampen both the upside and the downside from the AI theme.
What would flip the call: a regulatory or grid-connection setback that punishes the independent power producers specifically. In that scenario, VPU’s diversification into regulated utilities would cushion the drawdown that XLU’s concentration would amplify.
Contact [email protected] for any questions or corrections.