Nobody's Talking About This Energy Company, but It's Quietly Signing the Power Deals Feeding the AI Build-Out

Clearway Energy Group agreed to build nearly 1.2GW of renewable power for Alphabet/Google data centers, with first projects expected in 2027-2028, and Clearway says it will be paid substantially higher fixed pricing than prior arrangements. It also signed 600+MW of PPA extensions for wind repowering through 2041 with fixed pricing more than 2x prior contracted/merchant levels, supporting its forecast to grow CAFD/share from $2.12 last year to $2.90-$3.10+ by 2030. The article flags risks including a lowered 2026 CAFD outlook from El Niño-driven wind weakness and execution/asset acquisition at fair terms, but overall points to steadier cash flow and dividend growth tied to AI-driven power demand.

Analysis

CWEN is the cleaner way to express the AI power thesis because the economics are turning from narrative into contracted cash flow. The key mechanism is duration: every recontracting event that resets old wind/solar output to hyperscaler-backed pricing improves visibility on distributable cash, which should support both dividend growth and a lower equity risk premium. That makes CWEN more attractive than the higher-beta “future technology” power names, where valuation is still mostly contingent on execution and capital intensity.

The second-order winner is not just CWEN but the broader owner of scarce, already-permitted power interconnects. In a constrained grid, the asset that matters is not the turbine or the reactor concept but the ability to deliver firm megawatts on time; that should slowly re-rate contracted renewables and co-located power developers, while pressuring speculative AI power equities that need many more milestones before cash flow. The Google exposure is also a signaling event for other hyperscalers: if one large buyer pays up for certainty, comparable contracts across the sector should reset higher over the next 1-3 quarters.

Near-term risk is that weather noise and financing assumptions obscure the underlying improvement, so the stock may not move linearly with the story. The real falsifier is not AI demand; it is a weaker-than-expected 2026 CAFD path or acquisition terms that are too expensive relative to replacement cost. Longer term, the market may be underestimating the optionality of co-located digital infrastructure, but that is a 2029-2031 story, not a current earnings driver.

Key Decisions for Investors

• Buy CWEN on pullbacks over the next 2-6 weeks; frame as a 6-12 month total-return position where dividend growth plus multiple support can compound if recontracting terms hold.
• Pair trade: long CWEN / short BE or OKLO for a relative-value expression of AI power demand. CWEN has contracted cash flow and lower funding risk; BE/OKLO remain far more dependent on future execution and sentiment. Reassess if BE/OKLO raise guidance or secure large project-finance commitments.
• Set an alert on CWEN for any 2026 CAFD revision or acquisition pricing commentary. A second cut to CAFD or overly rich drop-down pricing would break the thesis and likely cap the rerating.
• For lower-risk exposure, own GOOG/GOOGL as the counterparty beneficiary rather than a pure power-name bet; the upside is modest, but it captures the value of securing long-duration energy at scale without taking project development risk.