3 Stocks That Pay Dividends and Offer AI Exposure

The article highlights Broadcom, Vertiv and Caterpillar as dividend-paying beneficiaries of AI-related data-center investment. Broadcom yields 0.7% and has delivered a 13.1% five-year annualized dividend-growth rate, while Caterpillar yielded 0.8% and returned $7.9 billion through dividends and buybacks in FY25. Vertiv, yielding 0.1%, is positioned to benefit from demand for data-center power, cooling and IT infrastructure; Vertiv and Caterpillar carry Zacks Rank #2 (Buy) ratings.

Analysis

This is low-information promotional research rather than a new fundamental catalyst; the immediate signal is weak and should not justify chasing any of the named AI-infrastructure equities. The more useful implication is that investor positioning is broadening from accelerators into the physical bottlenecks of data-center deployment—power conversion, cooling, backup generation, and networking—which can sustain relative earnings revisions even if GPU unit growth moderates.

VRT has the cleanest operating leverage to incremental data-center capex, but also the greatest de-rating risk if hyperscaler build schedules slip: its valuation depends on backlog conversion and margin execution, not its negligible income yield. CAT is a less crowded, cyclical beneficiary of on-site/bridge power demand; however, power-generation sales are unlikely to offset a broad construction, mining, or dealer-inventory downturn. AVGO remains the higher-quality AI exposure because custom silicon and networking content can compound with scale, but its multiple is vulnerable if large customers slow AI capex or ASIC adoption merely shifts spend away from GPUs rather than expanding total infrastructure budgets.

The contrarian view is that the next constraint is utility interconnection and grid availability, not availability of servers or cooling equipment. That favors longer-duration electrical infrastructure beneficiaries such as ETN, PWR and GEV more than the most visible data-center equipment names over 6-18 months. A reversal would be signaled by weaker hyperscaler capex commentary, VRT book-to-bill falling below 1x, or evidence that temporary generation demand is being displaced by delayed projects rather than converted into permanent capacity additions.

Key Decisions for Investors

• No event-driven trade on the article itself; treat it as a positioning indicator and wait for earnings/backlog data rather than adding to VRT on retail-income framing.
• Over 1-3 months, favor a relative-value pair: long CAT / short VRT in equal beta-weighted sizing if VRT rallies materially ahead of its next results. CAT offers less AI purity but lower single-theme valuation risk; exit if VRT reports sustained order acceleration and backlog conversion above expectations.
• For 6-18 months, build a basket long ETN, PWR and GEV versus a modest underweight in high-multiple AI hardware proxies, including VRT, as grid and interconnection constraints become the limiting factor. Falsify on a material decline in utility capex plans or a broad relaxation in interconnection queues.
• Maintain AVGO as preferred large-cap AI infrastructure exposure only on pullbacks or after confirmation that networking/custom-silicon revenue is growing independently of one customer cycle. Reduce if management signals slower aggregate AI infrastructure spend or if customer concentration drives a guidance reset.