The sector in the cross hairs of the bond sell-off looks poised for a bounce, says Mike Khouw

Utilities Select Sector ETF XLU has surrendered its Q1 outperformance as the fastest Treasury-yield rise since 1994 pressured rate-sensitive utility valuations, with independent power producers such as Constellation, NRG and Vistra seeing P/E multiples compress by 2 to 8 turns. The article argues that AI-driven data-center electricity demand and long-term power purchase agreements remain intact, while XLU now trades near its 10-year average P/E of 17.8x. It proposes a bullish November XLU 40/43 call spread for a net $0.85 debit, targeting a rebound toward recent highs over seven to eight weeks.

Analysis

The cleaner expression is not broad regulated utilities but a selective long in merchant generation versus rate-base-heavy utilities. CEG and VST can monetize tightening capacity and data-center load through wholesale pricing, nuclear/dispatchable capacity value, and bespoke contracts; traditional XLU constituents face the opposite near-term mix of financing costs, regulatory lag, and capex dilution. NRG is more nuanced: its retail book can cushion volatility, but its upside depends on retaining generation margins rather than passing power-price gains back to customers.

The key 1-3 month catalyst is not simply lower Treasury yields; it is evidence that hyperscaler power procurement is converting into contracted economics rather than nonbinding demand commentary. Announced PPAs, interconnection commitments, capacity-auction outcomes, and upward 2027-28 EBITDA guidance would force estimates higher and distinguish CEG/VST from the broader utility basket. A lower-rate move without contract or forward-power-price confirmation would likely create only a beta bounce, with XLU outperforming merchants rather than a durable rerating in IPPs.

Consensus may be underestimating the financing bottleneck as a competitive advantage. Higher rates and constrained transmission buildout can delay new gas, renewable, and grid capacity, increasing the scarcity value of existing dispatchable fleets; that is structurally favorable for CEG and VST over 6-18 months. The counter-risk is that hyperscalers self-supply via behind-the-meter generation, regional demand forecasts fail to translate into actual load, or political pressure caps retail/wholesale power returns; each would impair the scarcity thesis before reported demand does.

The published XLU call spread is a weak vehicle for the stated thesis because XLU's regulated exposure dilutes merchant-power upside. The trade is also promotional in character given the disclosed ownership and absence of contract-level EBITDA sensitivity. Use broad utilities only as a tactical rates hedge, while reserving directional risk for generators with identifiable data-center and capacity-market exposure.

Key Decisions for Investors

• Initiate a 3-6 month long CEG / short XLU pair, sized beta-neutral: CEG offers the highest-quality dispatchable-power scarcity exposure, while the XLU short funds rate-sensitive regulated capex risk. Target 10-15% relative outperformance; exit if 10-year Treasury yields fall materially but CEG forward power curves and consensus EBITDA do not improve within one quarter.
• Prefer long VST over NRG for the same horizon where ERCOT capacity tightness is the primary thesis; use NRG only if retail-margin disclosures confirm that rising load is accretive rather than competitively passed through. Cap the position ahead of summer reliability data, where weather-driven price spikes can reverse quickly.
• Do not execute the cited XLU November 40/43 vertical without verifying spot, expiry, and current implied volatility; the quoted strikes/premium may be stale. If IV remains below its 12-month percentile and rates stabilize, use a 2-3 month defined-risk XLU call spread only as a tactical rates-reversal overlay, not an AI-power allocation.
• Set event alerts for new hyperscaler PPAs, PJM/ERCOT capacity-market results, and 2027-28 EBITDA guidance revisions. Add to CEG/VST only after a disclosed contract includes duration, capacity, and pricing mechanics; otherwise treat data-center announcements as narrative rather than earnings catalysts.
• Falsification: reduce merchant-generator exposure if forward power prices/capacity values decline more than 10%, regulatory proceedings introduce return caps or windfall measures, or managements fail to translate load demand into upward free-cash-flow guidance by the next two earnings cycles.