2 AI Infrastructure Stocks So Cheap, I Almost Can't Believe It

The AI infrastructure boom is increasing electricity demand, but returns on the spending remain uncertain. The article highlights two infrastructure investments with different risk profiles; one is described as offering a dividend yield above 6%, long-term growth prospects, and potential mid-teens annualized total returns.

Analysis

The investable distinction is not simply “AI power demand,” but who bears construction, utilization, and financing risk. Existing assets with take-or-pay or otherwise durable contracts could capture incremental demand with less exposure to uncertain data-center buildout than merchant projects or speculative capacity additions. Conversely, contracted revenue is not automatically low risk: customer concentration, contract duration, escalation clauses, and counterparty credit determine whether the cash flow can support a high yield.

Near term, the AI narrative may lift power and utility multiples before new load materially contributes to earnings. Over 1–3 months, the key checks are signed, enforceable power agreements and company guidance on funded capacity—not announcements or projected demand. Over 6–18 months, grid connection delays, permitting, equipment availability, and higher financing costs could defer returns; regulated utilities may also face scrutiny over who pays for grid upgrades. A sustained rise in long-term rates is a particular headwind to high-yield infrastructure valuations.

The contrarian risk is treating a quoted yield and projected mid-teens return as evidence of a bargain: without the company names, valuation, debt maturity schedule, and contract terms, neither claim is independently assessable. No security-level trade is justified from this item alone.

Key Decisions for Investors

• Do not initiate a position solely on the article’s return and yield claims. Obtain the two company names, current valuation, debt maturities, capex commitments, and contract details before underwriting either security.
• Set a 1–3 month watchlist trigger: favor contracted power and grid-infrastructure exposure only if filings or counterparties verify signed capacity, credible payment terms, and funded interconnection plans. Treat aspirational data-center demand as a catalyst watch, not booked earnings.
• For any eventual long position, size against rate and execution risk; reassess if long-term yields rise materially, project schedules slip, or guidance fails to convert contracted capacity into operating cash flow. These would challenge the yield-support and growth thesis.
• Avoid assuming all utilities or power producers benefit equally. Compare contract duration, customer concentration, balance-sheet capacity, and regulatory cost recovery before choosing a relative-value exposure.