Data centers could consume up to 17% of total U.S. electricity by 2030, forcing a shift to behind-the-meter power and driving a backlog surge for AI power hardware. GE Vernova reported/accumulated a 100GW gas-turbine backlog through 2030, with turbine pricing projected to rise to $600/kW by 2027 (200% above 2019), while Eaton saw data center orders jump 200% in late 2025 and its electrical backlog reach a record $19.6B by early 2026. The trade-off is near-term margin pressure—Eaton’s Electrical Americas margin down 180 bps from heavy capex and backlog liquidation—even as GE Vernova is up ~62% YTD (P/E ~30.8) and Eaton is up ~26% since January (P/E ~39), both priced for strong continued hyperscaler spending. Overall, the structural demand tailwind is strong, but execution and macro capex risk could quickly pressure these elevated valuations.
The market is reallocating the AI trade from compute scarcity to power scarcity, which changes where the economic rent sits. That favors electrical infrastructure vendors and gas-generation enablers, but the key nuance is that backlog is not the same as cash conversion: these businesses are now trading on industrial-like visibility with software-like multiples, which makes execution the binding constraint. In the next 1-3 months, the stock reaction should hinge less on order growth than on whether management can show margin stabilization as factories scale and inventory/working capital rise.
The second-order winners are the suppliers of grid-adjacent hardware and localized generation, while the hidden losers are anything reliant on cheap, unconstrained data-center expansion: some software names, data-center REITs with power bottlenecks, and utilities that cannot monetize demand fast enough because of interconnection lag. A faster-than-expected utility buildout or regulator push to streamline interconnects would reduce the urgency premium, while higher copper, electrical steel, or labor costs would compress margins and delay revenue recognition even if bookings stay strong.
Contrarianly, the consensus may be underestimating how cyclical these "sure things" remain. GEV and ETN can keep outperforming structurally, but at these valuations the stocks need flawless capex conversion and no hyperscaler spending pause; any growth air pocket could trigger multiple compression before the backlog turns into earnings. The better expression is relative-value, not blind outright beta: own the enablers, but be disciplined about entry because the market is already paying for a lot of the 2027 story today.
AI-powered research, real-time alerts, and portfolio analytics for institutional investors.
Request TrialOverall Sentiment
mildly positive
Sentiment Score
0.15
Ticker Sentiment
