2 AI Stocks You Never Saw Coming -- and They Come With Dividends

The article highlights a new AI-driven demand stream for Caterpillar and Ford, centered on data center power generation and battery energy storage. Caterpillar’s AI-related engine, turbine, and generator sales could be a higher-margin growth driver, while Ford Energy plans roughly $2 billion of initial investment with deliveries targeted for late 2027 and Morgan Stanley estimating $500 million to $600 million of run-rate EBIT at 20 GWh capacity. Both names are framed as dividend-supported ways to gain exposure to AI infrastructure and power needs.

Analysis

The market is starting to re-rate “AI exposure” away from pure semis and toward the physical bottlenecks: power, cooling, backup generation, and on-site storage. That broadens the winning set, but it also means the second-order beneficiaries are likely to be slower-burning than chip names: order books can inflect now, while revenue and margin realization may lag by 12-24 months as utilities, data-center developers, and hyperscalers work through interconnection and procurement cycles.

CAT is the cleaner way to express that thesis because it already has distribution, service, and installed-base leverage in heavy equipment and power systems. The incremental upside is not just unit sales; it is mix shift into higher-margin recurring service, controls, and power-generation systems, which can compound faster than headline equipment demand. The risk is that the market has already begun capitalizing this optionality, so near-term upside could stall if investors decide the AI-power story is already embedded before the 2027/2028 earnings bridge becomes visible.

F is more controversial and more contrarian: the opportunity is real, but it is effectively a venture bet inside a mature cyclical balance sheet. If the storage business scales, the equity could get credit for a new EBITDA stream with industrial-like service durability; if it does not, capital intensity and execution risk will drag on returns. This creates a cleaner setup for option structures or a spread trade than for outright long equity, because the payoff is convex but delayed.

The market may be underestimating how quickly AI power demand can spill into adjacent beneficiaries such as electrical gear, grid equipment, and industrial service providers, while overestimating how fast new entrants can win contracts without a long reliability track record. The key catalyst is not the hype cycle itself but when data-center developers move from planning to purchase orders; until then, the trade can drift on narrative. A reversal would likely come from slower hyperscaler capex or easing power constraints through utility buildout, which would compress the scarcity premium across the theme.

Key Decisions for Investors

• Add to CAT on 6-12 month pullbacks; best risk/reward is a long-duration hold into 2027/2028 as AI power spend converts into higher-margin service and power-systems revenue.
• Use F as a call option on storage execution: prefer a small starter position or long-dated call spreads over outright size, with 12-24 month horizon and high execution risk.
• Pair long CAT / short a slower-moving industrial proxy without AI-power exposure to isolate the infrastructure re-rating; this captures thematic inflows while reducing market beta.
• Consider a basket long in electrical/grid beneficiaries alongside CAT rather than chasing only semis; the first-order capex winners are often the picks-and-shovels names with the longest contract visibility.
• If CAT or F rerate sharply on narrative alone, trim into strength: the core risk is a valuation overshoot before backlog converts into earnings.