AI Infrastructure Stocks: Barclays Highlights Top Gas Power Equipment Plays

Barclays says AI infrastructure spending could exceed $1 trillion annually, driving strong demand for gas power generation equipment and related components. The note highlights multiple beneficiaries, including Caterpillar, GE Vernova, Howmet, Baker Hughes, Cummins, Bloom Energy, FTAI Aviation, Woodward, ATI and Generac, many of which reported earnings beats, raised guidance, or saw analyst price-target increases. The setup is supportive for the sector and could move individual names, especially those tied to turbines, backup power and data-center energy needs.

Analysis

This is less a broad AI trade than a bottleneck trade: the real monetization sits in the picks-and-shovels layer that can convert hyperscaler capex into shipped megawatts, not in the platforms spending the money. The most attractive names are the ones with constrained capacity, high switching costs, and multi-quarter backlog visibility — especially equipment and component suppliers where pricing can still lag demand by 1-2 quarters. That creates a second-order setup where order books compound faster than consensus models, while margin leverage appears with a delay.

The market is likely underestimating how self-reinforcing the power buildout becomes if AI demand stays strong into 2027-2028. As utilities, data center developers, and cloud platforms all compete for the same generation hardware, lead times should extend, which tends to benefit OEMs and critical subcomponent suppliers more than diversified industrials. The cleaner expression is not “AI semis,” but the supply chain that enables incremental electrons: turbines, control systems, hot-section materials, backup generation, and dispatchable alternatives.

The main risk is duration mismatch. Several of these equities are already discounting a strong capex cycle, so near-term beats may not matter if investors start worrying about normalization after 2028 or a capex digestion phase sooner. A second risk is substitution: if grid interconnection or nuclear/renewables policy accelerates faster than expected, some of the gas-power urgency can fade, pressuring names that are most levered to a prolonged thermal generation shortage.

The contrarian angle is that the market may be overpaying for visible order growth while underpricing margin compression from capacity expansion. Companies announcing aggressive buildouts may be signaling future demand, but they also telegraph that supply is catching up, which can cap pricing power later in the cycle. The best risk/reward likely sits one layer deeper in enabling components where capacity is tighter and re-pricing power is stronger than at the headline OEM level.

Key Decisions for Investors

• Go long WWD / HWM / ATI as a basket for 3-6 months: these are the cleanest leveraged plays on turbine and hot-section component scarcity, with better pricing power than the headline OEMs; use pullbacks to add, target 15-20% upside, and cut if order growth decelerates for two consecutive prints.
• Pair long BE vs short ORCL for 6-12 months: if AI infrastructure spending keeps shifting toward behind-the-meter power solutions, BE has direct convexity to on-site generation demand while ORCL captures the capex burden; this isolates the power scarcity theme from broader AI software optimism.
• Long GEV on a 6-9 month horizon with a tight stop on valuation multiple compression: the setup is strongest if capacity expansion converts into margin expansion, but the stock is vulnerable if investors start to fear peak-order growth before 2028.
• Buy GNRC on weakness as a tactical hedge to grid-delay risk over the next 2-4 quarters: data center backup generation demand is a second-order beneficiary if interconnection queues worsen, with upside if backlog converts faster than expected.
• Avoid chasing CAT after strength; instead, use it as a relative short against WWD or HWM if industrial rotation fades: CAT is more exposed to the market already pricing in the capex cycle, while the suppliers have more room for estimate upgrades.