Prediction: Energy Transfer (ET) Will Crush the S&P 500 in the Second-Half of 2026.

Energy Transfer (ET) is up 17% YTD versus the S&P 500’s 9%, driven by its toll-based midstream model that insulates cash flows from volatile oil and gas prices. The company secured long-term natural gas supply agreements to major AI/data-center demand and raised 2026 adjusted EBITDA growth guidance to 14%–16% (from 9%–12%). Valuation remains inexpensive at ~7x this year’s EBITDA with a 6.9% forward yield, and 2025 adjusted DCF of $8.2B covered $4.6B of distributions, supporting potential future distribution growth.

Analysis

ET is being treated less like a utility-style cash dispenser and more like a scarce natural-gas capacity owner with embedded AI load optionality. The market is paying up for contracted growth, but the real mechanism is capital rotation: investors are searching for yield with upside, and midstream names with visible volume expansion can reprice sharply if the market believes the cash flows are longer-duration than the usual toll-road model.

Second-order winners are the gas-gathering and transport peers that can credibly attach themselves to power demand: WMB, KMI, TRGP, and MPLX should see sympathy multiple support if ET proves that data-center and utility contracts convert into EBITDA faster than expected. By contrast, high-yield utility proxies like NEE, DUK, and SO are vulnerable to yield substitution if rate cuts or lower inflation make ET’s distribution look like a cleaner growth-plus-income package.

The key risk is that “AI infrastructure” is a narrative ahead of monetization. Many data-center projects are power- and permit-constrained, so contract announcements can sit in backlog for quarters before driving cash flow; if 10-year yields stay elevated, the rerating stalls because the equity is still fundamentally competing with bonds. A miss in next quarter’s guidance, or any sign that 2026 EBITDA growth is front-loaded and not repeatable, would likely cap the move.

Contrarian view: the stock may already be halfway to fair value on story alone, while the underlying EBITDA uplift is probably less explosive than the market is assuming. The best asymmetry is not chasing the name outright, but expressing the thesis as relative value versus rate-sensitive income sectors and lower-growth midstream peers.

Key Decisions for Investors

• Initiate a starter long ET position on any pullback toward the recent breakout zone; use a 1-3 month horizon and treat 7x EV/EBITDA as a momentum support level, with thesis failure if 2026 EBITDA growth is revised back below low-teens.
• Pair trade: long ET / short XLU for 1-3 months to express yield-substitution and AI-load optionality; risk is a sharp drop in long rates, which would help both legs but likely favor XLU less.
• For relative value, prefer ET over KMI and EPD on a 3-6 month horizon if the market keeps rewarding AI-linked infrastructure narratives; the pair should work only if ET continues to outperform on guidance and contract backlog expansion.
• Do not chase with calls unless the next earnings release confirms conversion of data-center contracts into near-term EBITDA; missing data is backlog duration and take-or-pay structure, which would make the options case too dependent on narrative.
• Set an alert on 10-year Treasury yields and ET’s forward yield spread: if ET yield compresses toward utility levels without a corresponding guidance raise, take profits on the rerating trade.