Caterpillar Inc.

Caterpillar is the world's largest construction/mining-equipment maker, but the thesis-relevant business is Power & Energy: large reciprocating engine generator sets (2–10 MW) for backup/prime power, and Solar Turbines industrial gas turbines (up to ~38 MW) — the two hardware families that electrify and backstop hyperscale AI data centers.

• Q2 FY2026 (end 6/30/2026): sales & revenues $20.543B, +24% YoY — first $20B+ quarter in company history (10-Q / 8-K EX-99.1, filed 2026-08-05 / 08-04).
• Adjusted EPS $8.17 (+73% YoY) vs $6.20 consensus; adjusted operating margin 21.9% (+430 bps); GAAP operating income $4.295B (XBRL OperatingIncomeLoss, 10-Q).
• Segments (Q2): Power & Energy $8.2B (+17%), profit $2.0B (+30%), margin 24.6%; Construction Industries $8.3B (+35%); Resource Industries $4.6B (+20%); Financial Products $1.1B (+10%).
• Power generation sales +72% YoY (data-center demand); Power & Energy sales-to-users +33%.
• MP&E free cash flow $5.1B record (+$2.8B YoY); $2.2B returned to shareholders; dividend +8%.
• Balance sheet: cash $6.71B (6/30/2026), long-term debt $30.70B (12/31/2025), R&D $616M in Q2.
• Valuation (stockanalysis.com, 2026-09-11): market cap $376.3B, TTM revenue $74.73B (+18.4%), P/E 35.3, fwd P/E 28.5, 52wk range $423–$1,073, analyst PT $975.61 (Buy).

• monetization model: One-time equipment sales; aftermarket parts/consumables; service/repair; Cat Financial leasing/loans; rental. (10-K Item 1, Item 7: sec.gov ; Accession 0000018230-25-000010)
• revenue mix: FY2024 external sales/revenues $64.809B: Energy & Transportation ~38% ($24.81B), Construction Industries ~37% ($24.09B), Resource Industries ~18% ($11.94B), Financial Products ~6% ($3.97B). (10-K Segment note; same accession)
• who pays: End-user operators—contractors, miners, energy/transport, governments, OEMs, rental fleets—and independent dealers; Cat Financial finances buyers/dealers. (10-K Item 1; same accession)
• margin_value_capture: FY2024 operating profit margin ~20.2%; value accrues in aftermarket parts/services and Cat Financial, offsetting cyclical new-equipment margins. (10-K Item 7; same accession)

• customer classes: Industrial Infrastructure Owners, Utility - Grid Operators, Government - Program Buyers, Data Center Operators, Industrial Automation - OEMs, Water - Environmental Infrastructure Buyers, independent dealers, rental fleets, mining/oil & gas operators, railroads. (10-K Item 1; sec.gov ; Accession 0000018230-25-000010)

Caterpillar operates through three reportable segments:

Segment-level values above are from the FY2024 Form 10-K; exact segment cells should be re-verified against the filing and are marked needs primary source validation where they could not be live-pulled. [S1]

Caterpillar’s revenue model is not only machine sales. The company monetizes through:

1. New equipment sales — machines, engines, generator sets, locomotives, and turbines sold through independent dealers.
2. Aftermarket parts and services — replacement parts, remanufacturing, field service, digital/telematics subscriptions, and dealer service revenue.
3. Financing — Caterpillar Financial Services provides dealer and end-user financing, generating interest income and used-equipment residual exposure.

This is important for the conviction framework: the aftermarket and service stream is stickier and higher-margin than new machine sales. The 10-K does not break out aftermarket revenue on the income statement, but the recurring nature is visible in the segment operating margins and in management’s strategic emphasis on “services growth.” Exact aftermarket revenue dollar figure: needs primary source validation. [S1]

Using the segment figures above, the mix is approximately:

• Construction Industries: 36.4%
• Resource Industries: 20.7%
• Energy & Transportation: 43.6%
• Other/eliminations: -0.7%

Derived from segment sales divided by consolidated sales; segment figures need primary source validation. [S1]

The key structural insight is that Energy & Transportation has become the largest segment. That means oil and gas, power generation, marine, rail, and industrial engine demand are at least as important to CAT’s revenue mix as construction.

Caterpillar’s most decision-useful reported margin is operating profit margin. The company does not present a consolidated gross margin as cleanly as a software or consumer business because its cost structure includes high raw material, freight, labor, and dealer support costs. Consolidated GAAP operating profit exact figure needs primary source validation. Full-year 2024 operating margin was approximately 19.5% based on the earnings release; needs primary source validation against the 10-K. [S1][S4]

Segment operating margins are materially higher in Construction and Resource than in Energy & Transportation because of aftermarket/service mix:

Those segment margins are derived from segment sales and segment profit; exact values must be re-pulled from the 10-K. [S1]

The immediate buyers are Caterpillar’s independent dealers. End users are contractors, miners, oil and gas producers, power plant owners, railroads, marine operators, and industrial fleets. Caterpillar Financial Services finances both dealer inventory and end-user purchases.

The filing states that Caterpillar’s products are sold principally through an independent dealer network. Exact dealer count and branch count: needs primary source validation against the 10-K. [S1]

Caterpillar’s 10-K does not name a single end customer that supplies 10% or more of total revenue. The usual disclosure is that no single customer accounted for more than 10% of total sales and revenues. Therefore, for this memo, no top customer is invented. The customer base is diversified across global dealers, contractors, miners, oil and gas operators, rail, marine, and power generation. [S1]

Caterpillar’s 2024 result is a strong pricing-power case: total sales fell 3% but diluted EPS rose from $20.12 to $21.62. [S1] That combination implies CAT offset volume declines with price realization, cost discipline, or favorable mix. Exact price realization, volume, and currency bridge for FY2024: needs primary source validation against the 10-K or 8-K.

The worked calculation below isolates the earnings impact of a modest price increase in the largest segment.

From the segment disclosures, the buyer classes are:

• Construction Industries: heavy civil and infrastructure contractors, residential and nonresidential builders, road builders, quarry operators, rental companies, and municipalities.
• Resource Industries: mining companies, quarry operators, and heavy construction/mining contractors.
• Energy & Transportation: oil and gas producers, power generation developers/utilities, marine operators, rail operators, and industrial fleets.

The 10-K does not name specific customers. It does not identify any 10% customer. Therefore, no buyer names are invented in this memo. [S1]

• Mining companies buy based on productivity, fleet commonality, autonomous haulage, and life-cycle cost. These buyers are less price-sensitive and more uptime-sensitive.
• Contractors buy based on dealer support, financing, rental availability, and residual value. They are more price-sensitive but value dealer proximity.
• Oil and gas/power customers buy engines and turbines based on emissions compliance, fuel flexibility, lead times, and service network.
• Governments and infrastructure agencies buy through dealers or contractors; they value regulatory compliance and local service.

Caterpillar sells through an independent dealer network. The dealer network holds inventory and provides aftermarket service. This means CAT’s reported sales are partly a function of dealer inventory decisions, not final end-market demand. Deere’s filing shows the same dealer-channel dynamic. [S1][S2]

Caterpillar - industrial machinery and backup power generation for datacenter/AI buildout.

The AI buildout's binding constraint has shifted from GPUs to industrial reciprocating engines + gas turbines. A hyperscaler commissioning a 1 GW campus needs ~100× 10 MW gensets (or ~26× 38 MW turbines), and factory lead times are 40–107 weeks — so engines must be ordered before site work begins. CAT is the #1 large-genset OEM and one of only two US industrial gas-turbine makers at scale (with GE Vernova).

Q3 2026 earnings expected 2026-11-04 (Zacks/MarketWatch). Last print Q2 FY2026 (2026-08-04): record $20.5B sales (+24% YoY), Power & Energy $8.2B with power-generation +72% on data-center demand; backlog $72B. Parked pending panel.

1. Lead-time capacity ceiling. Large data-center genset factory lead times run 40–107 weeks across Western OEMs (Manufacturing Mag, 2026-06-04), with Linesight putting data-center-scale units in the Americas at 100–105 weeks. CAT is taking large recip-engine orders into 2028 and turbine orders into 2030 (Q2 2026 earnings call). Consequence: a campus commissioning in 2027 must order engines ~2 years ahead — power hardware, not GPUs, is now the schedule-critical path.

1. Backlog → revenue visibility. Backlog hit a record $72B (+$9B sequential, +92% YoY), with 59% (~$42.5B) expected to ship in the next 12 months (Q2 8-K EX-99.1). That is roughly 1.3× a full year of Power & Energy segment revenue, giving multi-quarter revenue visibility independent of new orders.

1. Incremental margin flow-through. Power & Energy Q2 segment margin was 24.6% (+250 bps YoY) on $8.2B sales (+17% YoY), with segment profit $2.0B (+30%). Management cited >40% incremental margin in Q2 (Q2 call) — each incremental data-center power dollar flows ~40¢ to segment operating profit, making the DC mix margin-accretive.

1. "Speed to power" wedge. CAT claims gas genset power generation can begin <1 year from order (vs ~5-year large-frame gas-turbine lead times at GE Vernova). CAT's restarted 10 MW medium-speed gas recip engine line adds ~1.5 GW over 18 months (shipments Q4 2026), incremental to the 65 GW large-engine capacity target for 2030. Capacity math: a 1 GW campus = 100× 10 MW recip engines or ~26× 38 MW turbines.

1. 202(c) demand tailwind. DOE issued six Section 202(c) emergency orders in two months (May–Sep 2026) authorizing grid operators (PJM, SPP) to direct data centers onto their own backup gensets during grid stress (PJM hit ~166 GW record demand July 2026). This reclassifies CAT gensets from rarely-run insurance to grid-reliability resources — higher utilization and faster aftermarket parts/replacement cycles.

1% × $28,231 million = $282.3 million incremental revenue.

$282.3 million × 30% = $84.7 million incremental segment operating profit.

$84.7 million × (1 - 0.22) = $66.1 million after-tax EPS contribution.

$66.1 million ÷ 493.5 million shares = $0.13 per share.