3 Energy Stocks Racing to Fix AI's Power Problem

The article argues the AI data-center power bottleneck is increasingly driving geothermal investment, citing Fervo Energy securing multibillion-dollar geothermal partnerships with Google and other hyperscalers tied to its Utah Cape Station. It projects those contracts could start flowing through revenue in the back half of 2026, potentially triggering a re-rating over the next 12–18 months, while Ormat is cited as showing ~50% top-line and ~30% bottom-line growth from mature hyperscaler demand. Baker Hughes is framed as a “picks-and-shovels” way to gain sector exposure as it supplies geothermal drilling equipment, with the key risk flagged as execution delays.

Analysis

This is less a geothermal trade than a capital-allocation signal: hyperscalers are effectively pre-buying dispatchable electrons, and the first money should accrue to the names that can deliver capacity now, not the purest long-duration story. That argues for ORA and BKR as the cleaner earnings expressions, while FRVO remains a financing-and-execution narrative until actual megawatts show up in reported numbers.

The second-order winner is the drilling and subsurface-services stack. If the AI power race persists, BKR should monetize across the entire buildout even when individual projects slip, whereas pure-play developers remain exposed to well productivity, interconnect timing, and project-level dilution. The market is probably overpaying for 'new energy' optionality in small caps and underpricing the scarcity value of companies that can close contracts, permit sites, and service wells at scale.

The main risk is timing, not demand: a delay in AI capex, a slower interconnection cycle, or disappointing well performance would push the thesis from months to years and hit FRVO first. Falsifiers are simple: if 1H26 contract wins do not convert into visible revenue/backlog by 2H26, or if ORA's operating leverage stalls, the sector should trade back as a story stock cohort rather than a cash-flow theme. Contrarian view: the best trade may be quality over purity, because the market tends to chase the most exotic beneficiaries before realizing the bottleneck sits in execution, not headlines.

Key Decisions for Investors

• Long ORA on pullbacks for a 6-12 month hold: cleaner visibility into contracted cash flows and less binary execution risk than early-stage peers; use earnings/backlog updates as the catalyst path.
• Long BKR as the picks-and-shovels expression over 6-18 months: monetizes the entire geothermal buildout regardless of which developer wins; thesis fails if order growth does not outpace broader oilfield-services cyclicality.
• Pair trade: long ORA / short FRVO for 6-12 months to express quality-vs-story; best risk/reward if the market keeps rewarding 'proof of cash flow' over pre-revenue optionality.
• Do not chase FRVO until the next reporting cycle confirms backlog-to-revenue conversion; if you must own it, treat it as a high-volatility event trade only after evidence of commercial scaling.
• Set an alert on hyperscaler capex commentary and interconnection/permitting headlines; a slowdown there would compress the entire geothermal premium quickly and is the most likely catalyst to fade the move.