Caterpillar's Power Generation Business Is Nearly as Big as Its Construction Segment. Here's What That Shift Means for the Stock's Multiple

Caterpillar’s power & energy unit grew revenue 17% YoY to $8.2B (nearly matching construction revenue of ~$8.3B), while operating profit rose to just over $2.0B, exceeding construction’s just-under $2.0B. The order backlog reached $72B, up 92% YoY for the three months ending June, driven by AI data-center power demand for conventional generators and gas turbines. Despite shares rising nearly 90% over the past 12 months, the article cites a consensus price target of $991.21—more than 20% above the current price—suggesting upside remains.

Analysis

CAT is being re-rated less as a construction-cycle proxy and more as a scarce-capacity supplier to AI infrastructure. The key mechanism is that data-center power demand is forcing customers into distributed generation and gas-turbine solutions now, which supports both mix and pricing; that can keep margins elevated even if unit growth moderates.

The second-order winners are the smaller, more leveraged power names such as PSIX and, to a lesser extent, NGS, where any spillover from hyperscale self-generation or gas-handling demand can move estimates faster than at CAT. By contrast, construction-only or broad industrial peers without power exposure may look cheap on an absolute basis but can still underperform if investors keep paying up for AI-adjacent capacity providers.

The contrarian miss is that backlog is a timing asset, not a permanent growth rate. If utility interconnects, permitting, or grid buildout catch up, the urgency premium behind onsite generation can fade quickly, and CAT’s >30x forward multiple leaves limited room for disappointment. Falsifiers are a deceleration in power order intake, margin compression in the power segment, or evidence that customers are shifting back toward utility power faster than expected.

Key Decisions for Investors

• Short CAT on strength over the next 1-3 months; the setup is attractive only if power orders/margins keep accelerating. Use a stop if next-quarter power backlog growth or segment operating margin re-accelerates meaningfully.
• Long PSIX as a smaller-cap leveraged beneficiary of distributed power demand; size modestly because liquidity and execution risk are high, but the upside is better if AI-driven genset demand broadens beyond CAT.
• Pair trade: long MSFT or GOOGL vs short CAT, beta-adjusted, to express the view that AI monetization duration sits with platform economics rather than equipment suppliers. Reassess after the next hyperscaler capex update and CAT earnings.
• If CAT sells off 5%-8% without any deterioration in power orders, take the other side tactically for the next earnings window; otherwise stay underweight and avoid chasing the multiple.