3 AI Power Stocks to Buy Now for Growth, Value and Breakout Potential

Zacks highlights Quanta Services, Powell Industries, and MYR Group as AI-driven power-infrastructure beneficiaries, citing accelerating data-center electricity demand and projected U.S. power-demand growth of 50% or more by 2050. Quanta reported a record $53.4B backlog and is forecast to grow revenue 38% in 2026, while Powell's backlog rose 69% year over year to $2.4B, including a data-center order worth more than $400M. MYR Group's backlog increased 20% year over year, with adjusted EPS projected to rise 61% in 2026; all three stocks are presented as materially below prior highs despite favorable earnings outlooks.

Analysis

The investable bottleneck is shifting from generation headlines to interconnection, substation capacity, switchgear and field execution. PWR is the highest-quality expression because its scale, utility relationships and project-management capability should let it monetize multi-year transmission work with less customer concentration than POWL; its premium multiple is therefore defensible, but leaves it most exposed to a broad AI-capex de-rating. MYRG offers the cleaner mean-reversion setup if its recent operational recovery proves durable, though its greater fixed-price contract exposure makes labor availability, weather and project-cost discipline more important than the data-center narrative.

POWL has the largest earnings torque but also the least forgiving setup: a small number of large engineered-equipment awards can create uneven revenue recognition, working-capital swings and abrupt backlog comparisons. The key second-order constraint is not announced data-center spend but utility interconnection queues, transformer/switchgear lead times and available generation; delays at any of these points defer electrical-equipment deliveries rather than eliminate demand. This argues for favoring contractors with recurring utility maintenance and transmission exposure over companies priced on a handful of hyperscale awards.

Over the next 1-3 months, earnings revisions and order commentary can support a rebound in the group, but a recovery should not be confused with a secular re-rating. For 6-18 months, the upside case requires sustained utility capex authorization and data-center projects moving from announced to contracted power supply. Falsifiers are declining backlog conversion, margin erosion from labor/material inflation, or a meaningful slowdown in hyperscaler capex guidance; a rise in long-end yields would also compress the sector's elevated duration-sensitive multiples. Consensus appears too focused on AI demand and insufficiently on execution capacity, making selective exposure preferable to a broad ‘AI power’ basket.

Key Decisions for Investors

• Initiate a 6-12 month long PWR / short POWL pair, sized dollar-neutral: PWR offers more diversified transmission and utility exposure, while POWL is more vulnerable if large-order timing or customer concentration disappoints. Target 15-20% relative outperformance; stop if POWL reports two consecutive quarters of backlog acceleration with stable or expanding gross margin while PWR's electric backlog conversion weakens.
• Accumulate MYRG in tranches only after the next quarterly backlog and gross-margin update confirms recovery; use a 9-12 month horizon. The reward is a valuation normalization toward higher-quality grid-contractor peers, but exit on a material reduction in operating-margin guidance or backlog contraction, which would indicate that the rebound is project-specific rather than structural.
• Do not chase POWL solely on technical oversold signals. Treat a new disclosed data-center or utility order, together with evidence of positive free-cash-flow conversion, as the entry trigger; otherwise maintain it as a watchlist name because a single delayed mega-project can dominate near-term estimate risk.
• Hedge sector beta through a modest long XLU or regulated-utility exposure against contractor longs if rates are stable: utilities are the entities ultimately earning regulated returns on a substantial share of grid investment. Remove the hedge if long Treasury yields rise sharply, since both utility valuations and customer willingness to accelerate capital programs can weaken.