Utility sector ETF XLU faces headwinds as Texas regulators temporarily freeze new data center demand approvals, stalling projected load growth.

Texas regulators have paused new data-center interconnections, and a recent application fee adjustment reduced American Electric Power’s Ohio pipeline by more than half overnight. Whether this development unravels the investment case for the market’s most popular utility ETF depends on a single metric—one that remains exceptionally difficult to verify.

This post may contain links from our sponsors and affiliates, and Flywheel Publishing may receive compensation for actions taken through them.

Throughout the year, the Utilities Select Sector SPDR Fund (NYSEARCA:XLU) has been marketed as the primary vehicle for capturing the artificial intelligence power buildout without direct exposure to AI equities. That narrative rests entirely on the persistence of a massive backlog of gigawatt-scale interconnection requests, which utilities cite to justify record capital expenditure plans.

Texas has introduced a significant complication. Under ERCOT’s new Batch Zero review process, fresh data-center hookups are paused while regulators audit ownership structures, financing arrangements, water consumption, and on-site generation capacity. Concurrently, Reuters reported that AEP Ohio’s data-center pipeline contracted by more than 50% following the implementation of new application fees.

This shift carries substantial weight for XLU. The fund’s five largest positions—NextEra Energy (NYSE:NEE) at approximately 13%, Southern Company (NYSE:SO) near 8%, Duke Energy (NYSE:DUK) around 7%, Constellation Energy (NASDAQ:CEG) at roughly 6%, and American Electric Power (NASDAQ:AEP) just above 5%—are precisely the companies whose recent earnings calls have centered on hyperscale computing contracts. If a meaningful portion of that projected pipeline proves to be administrative fiction, XLU’s growth narrative compresses into something resembling a traditional bond proxy, currently trading at a valuation aligned with a 10-year Treasury yield of 4.77%.

A single data-center developer can submit interconnection requests to multiple utilities across several states for the same facility, with none of those filings legally committing the developer to commence construction. Consequently, requested capacity is frequently counted multiple times, allowing U.S. utility interconnection queues to balloon into figures that no realistic construction schedule could ever support. The mechanism that distinguishes committed projects from speculative filings is financial collateral. Exelon demonstrated this clearly this quarter, reducing its data-center pipeline from 43 gigawatts to 36 gigawatts. Of that revised total, only 11 gigawatts were classified as high-probability, and merely 4 gigawatts were secured by signed transmission security agreements alongside $1 billion in collateral. AEP is conducting a parallel assessment in Texas, where it has collected nearly $2 billion in cash or collateral against 45 gigawatts of Batch Zero load. Interconnection pipelines consistently contract when developers are required to provide verifiable funding.

Regulated utilities earn an allowed return on the capital deployed within their rate base. While constructing substations and transmission lines generates profit, building infrastructure for unverified projects introduces severe financial risk. If a utility energizes lines for a hyperscaler that ultimately fails to proceed, those costs are passed to existing ratepayers and state regulatory commissions, which may subsequently reduce allowed returns when consumer bills spike. This dynamic underpins the risk embedded within AEP’s $78 billion five-year capital plan and its 7% to 9% EPS growth target through 2030. It also explains why Exelon (NASDAQ:EXC) maintained its $41 billion capital plan through 2029 despite trimming its pipeline. The most prudent utilities are now pricing their projections strictly around the audited, financially secured subset of demand.

XLU holds approximately 30 utility names with $23.1 billion in net assets, according to State Street’s fund documentation, with the top five holdings driving the majority of performance. These companies operate under fundamentally different business models. NextEra and Southern function as regulated wires-and-generation utilities characterized by extended capital expenditure cycles, whereas Constellation operates as a merchant nuclear power provider executing bilateral contracts. This quarter, Constellation secured 920 megawatts of long-term nuclear power purchase agreements with an average duration of 18.5 years against investment-grade counterparties, while raising its 2026 earnings guidance to $11.50–$12.50 per share. This represents genuinely differentiated, AI-driven exposure. It also clarifies why CEG has declined approximately 15% year-to-date through September 4, 2026, while the remainder of the fund continues to trade sideways.

The fund does not fully deliver on its stated pitch when measured against the broader market it aims to complement. Through September 4, 2026, XLU returned approximately 2% year-to-date, roughly 6% over one year, and 44% over five years. By comparison, the S&P 500 ETF (SPY) returned approximately 13%, 19%, and 70% over the identical periods. When evaluated against Vanguard’s sector alternative, XLU appears structurally inefficient. The Vanguard Utilities ETF (VPU) delivered nearly identical returns—approximately 2% year-to-date, 6% over one year, and 43% over five years—at a significantly lower 0.09% expense ratio. For passive utility exposure, VPU performs the same function at a fraction of the cost. For investors specifically targeting the AI-power thesis, holding CEG, NEE, and AEP directly concentrates exposure into the pipelines undergoing active regulatory scrutiny, rather than diluting it across municipal water providers and gas distribution companies. (For further detail, we have compiled seven non-chipmaker suppliers benefiting from this infrastructure expansion, spanning power delivery, cooling systems, and networking, in a complimentary AI infrastructure report.)

The current setup presents a balanced but cautiously weighted profile. XLU retains ownership of the critical transmission infrastructure required for any verified load, and even a heavily discounted interconnection queue surpasses available generation capacity following two decades of stagnant demand growth. However, the fund’s current valuation prices in the optimistic, unverified pipeline rather than the audited reality. With the 10-year Treasury yield standing at 4.77% as of September 3, 2026, fixed-income buyers have a more efficient alternative. Investors seeking concentrated AI-power exposure are better positioned to acquire Constellation or NextEra directly, while those prioritizing defensive income can access the same utility basket more efficiently through VPU.

Please direct questions or corrections to [email protected].