3 Pipeline Stocks Getting Bigger Without Taking on More Risk

AI data-center power demand is positioned as a growth catalyst for U.S. natural-gas pipeline operators Kinder Morgan, Williams and Energy Transfer. Kinder Morgan has a $9.6B project backlog, including $8.8B in natural-gas projects, while Williams is participating in six data-center power projects totaling $9.6B and targets at least 11% annual EBITDA growth through 2030. Energy Transfer plans up to $5.9B of growth capex, including projects supplying Permian gas to Texas, New Mexico and Arizona data-center markets, and offers a 6.8% distribution yield.

Analysis

The investable issue is not aggregate AI power demand but whether proposed data-center loads convert into creditworthy, long-duration transportation contracts before utilities secure generation and interconnection. WMB has the cleanest exposure to constrained Southeast/Atlantic demand corridors and can capture both regulated-like pipe returns and behind-the-meter power optionality; that optionality deserves a premium only once customer commitments, generation ownership economics, and permitted capacity are disclosed. KMI offers lower-beta backlog conversion, but its broad footprint makes AI-driven EBITDA upside less material to consolidated results.

ET is the highest-torque expression if Permian associated-gas supply remains structurally discounted and Texas/Southwest data-center development proceeds. The second-order risk is that data-center developers may choose self-generation closer to load, reducing long-haul transport needs, while ERCOT interconnection delays can push demand realization out by years even where gas is abundant. A weaker oil-price environment would also slow Permian drilling and associated-gas growth, potentially tightening ET's supply-side advantage while impairing volume assumptions across its broader network.

Near term, this is unlikely to be a durable rerating catalyst absent signed contracts or upward EBITDA guidance; pipeline equities already trade on project backlog credibility, distribution growth, and rates. Over 6-18 months, the key differentiator will be permitted projects entering service on time at sub-6x build costs, versus capex escalation and customer concentration that turn nominal backlog into lower-return growth. Consensus may be underestimating electric-power bottlenecks: gas pipeline capacity alone has limited value if generation equipment, transmission, and utility approvals remain constrained.

Key Decisions for Investors

• Pair trade over 6-12 months: long WMB / short KMI in equal dollar amounts. WMB has greater exposure to premium demand corridors and higher embedded growth optionality; exit if WMB fails to convert announced power projects into contracted backlog by the next two earnings cycles or if the valuation premium exceeds incremental EBITDA visibility.
• Accumulate ET only on broad midstream or energy-price weakness, targeting a 12-18 month total-return thesis driven by distribution growth plus project commissioning rather than an immediate AI rerating. Size below WMB because MLP structure, Permian production sensitivity, and higher execution complexity raise downside; reassess if Permian gas basis tightens materially or growth capex rises without distributable-cash-flow coverage improvement.
• Maintain KMI as a defensive income holding rather than an AI beta trade for the next 3-6 months. Add only following disclosed contract awards that raise forward EBITDA guidance; absent that evidence, its upside is likely capped by its mature asset base and interest-rate sensitivity.
• Monitor ERCOT and Southeast utility interconnection queues, turbine delivery schedules, and disclosed customer credit support. Delays or cancellations in these inputs are the earliest falsifier for the data-center gas-demand thesis and should trigger reduction of WMB/ET exposure before project-level impairments appear in guidance.