XLU’s AI Power Story Crumbles as Texas Freezes Data-Center Demand
Texas regulators just froze data-center hookups, and one application fee cut AEP Ohio's pipeline by more than half overnight. Whether that unravels the entire investment case for the most popular utility ETF on the market depends on a number almost…
This post may contain links from our sponsors and affiliates, and Flywheel Publishing may receive compensation for actions taken through them.
The Utilities Select Sector SPDR Fund (NYSEARCA:XLU) has been marketed all year as the way to own the AI power buildout without buying AI stocks. That pitch depends on one number holding up: the sea of gigawatt-scale interconnection requests utilities cite to justify record capital plans.
Texas just poked a hole in it. ERCOT’s new Batch Zero review pauses fresh data-center hookups while regulators audit ownership, financing, water use, and on-site generation, and Reuters reported AEP Ohio’s data-center pipeline shrank by more than half after new application fees took effect.
That matters for XLU because its five largest utility positions, NextEra (NYSE:NEE | NEE Price Prediction) at about 13%, Southern (NYSE:SO) near 8%, Duke (NYSE:DUK) around 7%, Constellation (NASDAQ:CEG) at roughly 6%, and American Electric Power (NASDAQ:AEP) just above 5%, are the same companies whose earnings calls now revolve around hyperscaler contracts. If a chunk of that pipeline is a paperwork phantom, XLU’s growth story compresses to something much closer to a bond proxy trading at a 10-year Treasury yield of 4.77%.
How Requested Load Becomes a Multiple of Real Load
A single data-center developer can file interconnection requests with several utilities across several states for the same project, and none of those filings commits the developer to breaking ground. Requested capacity ends up counted several times, which is how U.S. utility queues balloon into figures no plausible construction schedule can serve.
The mechanism that separates real from phantom is money. Exelon showed it plainly this quarter, cutting its data-center pipeline to 36 gigawatts from 43, with only 11 gigawatts classified as high-probability and just 4 gigawatts backed by signed transmission security agreements and $1 billion of collateral. AEP is running the same experiment in Texas, where it has collected nearly $2 billion in cash or collateral against 45 gigawatts of Batch Zero load. Pipelines shrink when someone asks for a check.
Why a Regulated Utility Cares Who’s Really Coming
A regulated utility earns an allowed return on the capital in its rate base, which is why building substations and transmission lines is profitable, and building the wrong ones is dangerous. If a company energizes lines for a hyperscaler that never arrives, the cost lands on existing ratepayers, then on state commissions that can cut the allowed return when bills spike.
That is the risk hidden inside AEP’s $78 billion five-year capital plan and 7% to 9% EPS growth target through 2030, and it is why Exelon (NASDAQ:EXC) insisted its $41 billion plan through 2029 did not budge when the pipeline was trimmed. The prudent utilities are pricing to the audited subset.
What XLU Actually Owns, and Why It Isn’t Uniform
XLU holds roughly 30 names with $23.1 billion in net assets, per the State Street fund page, and the top five drive the fund. Those five are not the same business. NextEra and Southern are regulated wires-and-generation utilities with long capex tails; Constellation sells merchant nuclear power under bilateral contracts.
Constellation this quarter signed 920 megawatts of long-term nuclear PPAs with an average term of 18.5 years with investment-grade counterparties and raised guidance to $11.50 to $12.50 for 2026. That is a genuinely differentiated, AI-powered exposure, and it also explains why CEG is down about 15% year-to-date through September 4, 2026, while the rest of the fund plods along.
Does the Fund Deliver on the Pitch?
Not really; once you compare it to the market it is supposed to complement. XLU returned about 2% year-to-date, roughly 6% over one year, and 44% over five years through September 4, 2026. SPY returned about 13%, 19%, and 70% over the same periods.
Against Vanguard’s alternative, XLU looks like an expensive twin. VPU returned about 2% year-to-date, roughly 6% over one year, and 43% over five years, at a 0.09% expense ratio. If you want passive utilities, VPU is doing the same job for less. If you want the AI-power thesis specifically, owning CEG, NEE, and AEP directly concentrates you into the pipelines actually being audited rather than diluting them across water utilities and gas LDCs (we pulled together seven of the non-chipmaker suppliers riding the same buildout, from power to cooling to networking, in a free AI infrastructure report).
Is XLU ETF a Buy?
The setup looks balanced but skews cautious. XLU still owns the toll roads for whatever load does show up, and even a heavily discounted queue exceeds available generation after two decades of flat demand. But the fund’s valuation reflects the optimistic pipeline, not the audited one, and a 4.77% 10-year yield as of September 3, 2026 gives income buyers a cleaner alternative.
Investors seeking the AI-power thesis get more concentrated exposure through Constellation or NextEra directly, while those focused on defensive income get the same utility basket more cheaply through VPU.
Contact [email protected] for any questions or corrections.
Omor Ibne Ehsan is a writer at 24/7 Wall St. He is a self-taught investor with a focus on growth and cyclical stocks that have strong fundamentals, value, and long-term potential. He also has an interest in high-risk, high-reward investments such as cryptocurrencies and penny stocks.
