3 Dividend Stocks Sitting Outside the AI Power Trade -- And Still Winning

Enterprise Products Partners, Enbridge and Energy Transfer offer yields of 5.7%, 5.8% and 6.3%, respectively, as rising AI-related electricity demand could increase natural-gas pipeline volumes and create new infrastructure investment opportunities. U.S. energy demand is projected to rise 60% from 2025 to 2045, versus 10% growth in the prior 20 years, supporting the midstream growth thesis. Enterprise has raised distributions for 28 years and Enbridge has a 31-year dividend-growth streak, while Energy Transfer offers the highest yield but carries added risk following its pandemic-era distribution cut.

Analysis

The investable mechanism is not broad AI demand but incremental, locational gas burn at power-constrained data-center clusters. Existing pipelines benefit only where their systems connect to Gulf Coast supply, LNG export corridors, and constrained utility/load-growth regions; broad U.S. electricity-demand forecasts do not automatically translate into tariff or volume upside. EPD is best positioned for fee-based Gulf Coast/NGL and export-linked throughput, while ET offers more operating leverage to higher gas volumes but also greater execution and valuation sensitivity.

ENB's differentiated exposure is regulated gas distribution and contracted pipelines, which can convert load growth into rate-base expansion rather than solely volume growth. That should produce lower-beta earnings durability over 6-18 months, although CAD/USD creates a meaningful offset for USD-based returns. The second-order winner is likely gas-fired generation and gas infrastructure development in ERCOT, PJM and Southeast markets; the loser is merchant power or data-center developers unable to secure interconnection and firm fuel supply, not necessarily renewable developers broadly.

Near-term, this is unlikely to re-rate midstream materially absent evidence in quarterly guidance: new long-term contracts, committed expansion capital, or rising utilization on named systems. Consensus may be over-attributing all AI power demand to natural gas; gas demand can be displaced by nuclear uprates, storage, renewables plus transmission, and efficiency improvements. The key falsifier is a sustained weakening in LNG feedgas/pipeline utilization or management capital plans shifting toward low-return expansions without corresponding take-or-pay commitments.

ET's yield premium is compensation for a more fragile trust discount, not a free AI option. A credible path to narrowing that discount requires leverage discipline and distribution coverage to remain intact through a weaker commodity and volume environment; otherwise, its higher beta can underperform EPD despite comparable thematic exposure.

Key Decisions for Investors

• Prefer a 6-12 month long EPD / short ET pair for defensive midstream exposure: EPD has the cleaner balance-sheet and contract-quality profile, while ET carries greater downside if growth capex or leverage rises. Reassess if ET demonstrates sustained coverage improvement and committed, high-return expansion backlog.
• Accumulate ENB on broad energy-sector weakness for a 12-18 month rate-base and gas-utility load-growth thesis; size for CAD/USD volatility. Exit or reduce if regulatory outcomes impair allowed returns or utility capital spending materially exceeds internally funded capacity.
• Do not add directional AI exposure through NVDA based on this theme; pipeline cash flows are too indirect and too slow-moving to create a meaningful incremental semiconductor catalyst over the next 1-3 months.
• Set an earnings-season watch item for EPD, ET and ENB: require disclosure of incremental contracted capacity, utilization, expected project returns, and funding sources before underwriting an AI-related multiple expansion. Absent those datapoints, treat current yields as income trades rather than growth re-rating candidates.