Caterpillar
Cyclical capital-equipment OEM: sells machines, engines, gen-sets and turbines plus a high-margin recurring aftermarket (parts + service); a captive finance arm (Cat Financial) funds dealer/customer purchases. Distribution runs through ~150 independent dealers, not direct.
The thesis on this name
State of Data-Center Power
The reciprocating-engine and gen-set leader for both backup and prime on-site datacenter power. Large reciprocating-engine backlog has grown >3.5x since Jan-2024, CAT is tripling large-engine capacity from 2024 levels, and some orders extend well into 2028 (fact); a Jan-2026 alliance supplies 2GW of G3516 gas gensets to a Microsoft/NVIDIA-linked West Virginia campus (fact). The asymmetry: CAT is a diversified industrial whose Power & Energy datacenter torque is under-credited relative to pure-plays, giving exposure to on-site power scarcity with construction/mining diversification cushioning the AI-capex beta. A high-potential expression at an industrial multiple.
Earnings, margins, COGS & capex
FY2025 was a record-revenue year ($67.6B, +4%) but only mid-single-digit growth as core construction/mining stayed soft; Q1 FY26 then re-accelerated to +22% YoY ($17.4B) as data-center power demand inflected. Power & Energy (the renamed Energy & Transportation segment) is now the growth engine, with reciprocating-engine and turbine backlog tied to AI data centers driving total backlog to a record $63B (+79% YoY). Tariffs are the main margin drag (~$710M unfavorable manufacturing cost in Q1 FY26; ~$2.4B FY26 headwind cited), partly offset by price and volume.
Income statement — where each revenue dollar goes
% of revenueOf every $1 of revenue, ~72¢ is cost of goods and ~12¢ operating expense, leaving ~17¢ of operating profit (~13¢ net).
Revenue trend
Margins
flat/down (estimate; tariff pressure)
down (tariffs)
roughly flat
down vs 2024
down (~11% FCF decline YoY)
COGS structure
COGS is dominated by raw materials and components — steel/metals, engines/powertrain, hydraulics, electronics/controls — plus large in-house manufacturing labor and overhead. The current swing factor is TARIFFS (~$2.4B FY26 enterprise headwind cited), on top of freight and warranty; volume leverage and price are the offsets, and aftermarket parts carry structurally higher margin than whole-machine sales.
Capex
Capex funds the ~3x large-reciprocating-engine capacity expansion (adds ~15GW annual output; bulk spend 2027-2029), turbine/Solar Turbines capacity, plus routine plant maintenance, autonomy and digital. Intensity is rising from the historical ~4-6% range as the engine ramp accelerates (fact on plan; intensity estimate).
Latest earnings
Beat — adjusted EPS $5.54 vs ~$4.62 consensus (~$0.92 / ~19% beat, largest of trailing five quarters).
Low-double-digit FY2026 sales growth with improved margins vs January guidance; raised 2024-2030 enterprise CAGR to 6-9% and power-gen to >3x 2024 sales by 2030. No single hard FY26 revenue number in the release.
- Backlog
- $63B record (+79% YoY)
- Power & Energy (Q1 FY26)
- $7.0B (+22% YoY)
- Construction Industries (Q1 FY26)
- $7.16B (+38% YoY)
- Adjusted operating margin
- 18.0%
Growth drivers
- AI data-center power — reciprocating-engine backlog >3.5x since Jan-2024; Power & Energy +22% YoY Q1 FY26 (fact)
- Capacity tripling to ~3x 2024 levels (+15GW/yr) converting demand to revenue 2026-2029 (fact)
- High-margin aftermarket parts + service growing with the installed base
- Record $63B backlog (+79% YoY), some orders into 2028, gives multi-year visibility (fact)
- Raised targets: 6-9% enterprise CAGR to 2030; power-gen >3x 2024 sales by 2030 (fact)
- Eventual construction/mining + energy-transition-minerals cyclical recovery (estimate)
Reported financials — SEC EDGAR
Audited GAAP figures pulled from SEC filings · latest filing 2026-02-13. The audited primary-source spine — not financial advice.
Revenue — annual (GAAP)
Margins & balance sheet — FY’25
Bull & bear
CAT has transformed from a cyclical machinery name into a leveraged, supply-constrained beneficiary of the AI data-center power build-out, with a record $63B backlog and a tripling of large-engine capacity converting demand into multi-year earnings.
- Power & Energy is inflecting: +22% YoY in Q1 FY26, reciprocating-engine backlog >3.5x since Jan-2024, orders extending into 2028 — durable, visible growth (fact).
- Capacity tripling to ~3x 2024 levels (+15GW/yr) means today's constraint is supply, not demand — revenue is largely pre-sold as plants come online 2027-2029 (fact).
- Management raised long-term targets (6-9% enterprise CAGR; power-gen >3x 2024 sales by 2030), a rare bullish guide-up for a 100-year-old industrial (fact).
- Both-ends position: CAT sells backup gen-sets AND prime-power turbines/engines, plus high-margin aftermarket on a growing installed base — recurring revenue compounds.
- Q1 FY26 ~19% EPS beat and record backlog show operational execution; ~$8B/yr returned to holders supports the stock even before the engine ramp fully hits.
At ~49x trailing earnings CAT prices in years of uninterrupted AI-power demand, while its core construction/mining business is only growing ~4%, tariffs are eroding margins, and the data-center order book is exposed to AI-capex digestion and aggressive competitor capacity additions.
- Valuation is the core short: ~48-51x trailing / ~39x forward leaves no margin of safety if data-center order growth slows or backlog gets cancelled.
- The growth is concentrated and potentially cyclical — if hyperscaler capex digests or grid build-out catches up, the recip/turbine demand spike can reverse fast.
- Tariffs are a real, recurring drag (~$2.4B FY26 cited), and GAAP operating margin already slipped YoY despite +22% sales — pricing may not fully offset.
- Competitors (Cummins, Rolls-Royce mtu, GE Vernova, Mitsubishi, Siemens Energy, Kohler) are all adding data-center power capacity into the same demand — risk of oversupply and price erosion by 2027-2029.
- Core Construction + Resource Industries remain mature and macro-sensitive; a construction/mining downturn would swamp the data-center tailwind and expose the cyclical base.
What it is worth
Blend of forward P/E (~39x current) against the ~6-9% raised long-term enterprise CAGR + data-center mix shift, cross-checked with EV/EBITDA and the $63B backlog underpinning forward revenue. The debate is multiple, not demand: at ~49x trailing CAT is priced as a structural AI-power compounder, not a cyclical industrial.
~$650-750
AI-power demand digests or competitor oversupply hits pricing; CAT re-rates back toward a ~18-22x industrial-cyclical multiple.
~$950-1,050
roughly current; high-double-digit EPS growth offset by gradual multiple compression toward ~32-35x forward.
~$1,250+
sustained 6-9% growth, data-center mix lifts margins, multiple holds ~39-45x forward as backlog converts cleanly.
Re-rating is real and demand-backed, but the multiple bakes in years of uninterrupted data-center growth — risk/reward is asymmetric to the downside on any AI-capex digestion. Not financial advice.
SWOT
Strengths
- Dominant scale and brand in heavy equipment + engines — ~150-dealer global distribution and parts network competitors can't replicate quickly
- Power & Energy franchise (Cat reciprocating engines + Solar Turbines) is a top-2 supplier of data-center backup AND prime power — a rare both-ends position
- Record $63B backlog (+79% YoY) gives multi-year revenue visibility, some orders into 2028
- High-margin recurring aftermarket dampens the equipment cycle and grows with the installed base
- Strong cash generation ($11.7B FY2025 operating cash flow) funding both the engine capacity build and ~$8B/yr of buybacks+dividends
Weaknesses
- Core Construction + Resource Industries are mature, cyclical and have been roughly flat-to-soft — total FY2025 growth was only +4%
- Tariffs are a direct, sizeable margin drag (~$2.4B FY26 cited — ~$710M Q1 unfavorable manufacturing cost) compressing GAAP operating margin
- Long, lumpy lead times — capacity to meet data-center demand doesn't fully arrive until 2027-2029, so near-term upside is supply-constrained
- Consolidated balance sheet carries ~$43-44B reported debt (mostly captive finance), which can obscure industrial leverage for casual readers
- Heavy exposure to global construction/mining capex and commodity prices — a macro downturn hits the legacy core hard
Opportunities
- AI/data-center power demand is structural and early — gas/recip prime power is displacing grid-connection waits, expanding CAT's addressable market
- Raised long-term targets (6-9% enterprise CAGR, power-gen >3x 2024 sales by 2030) signal management conviction in a multi-year run
- Microgrids + hydrogen/fuel-cell pilots (e.g. Microsoft) and battery-hybrid systems open adjacent power-product revenue
- Mining electrification and energy-transition minerals (copper, lithium) can drive a future Resource Industries upcycle
- Services/digital + autonomy attach rates raise lifetime value per unit sold
Threats
- Data-center power demand could prove cyclical/over-built — an AI-capex digestion phase would hit the new growth narrative and the rich multiple
- Stretched valuation (~48-51x trailing) prices in sustained data-center growth — any backlog cancellation or order slowdown de-rates the stock
- Intense competition from Cummins, Rolls-Royce mtu, GE Vernova, Mitsubishi Power, Siemens Energy, and Kohler/Rehlko, several adding capacity into the same demand
- Tariff/trade-policy escalation and steel/component cost inflation can persistently pressure margins
- Permitting, emissions regulation and grid-services rules could shift the fuel mix away from CAT's strongest products
Moats, dependencies & bottlenecks
Moats
global parts/service) Decades to replicate; locks in lifetime parts/service revenue and uptime SLAs hyperscalers need.
CAT/Solar Turbines reliability is the buying criterion for mission-critical data-center power; switching risk is operational, not just price.
Tripling large-engine capacity is a real lead, but rivals are also expanding — advantage is timing and footprint, not exclusivity.
Lowers customer cost of ownership and smooths sales; a structural enabler more than a standalone moat.
gas recip, turbines, microgrid, hydrogen pilots) Lets CAT win whichever fuel/permitting path a site favors; partly matched by Cummins and Rolls-Royce mtu.
Dependencies
Amazon, Google, Meta, neoclouds + colos) The marginal growth dollar now rides on AI infrastructure spend; a capex digestion phase directly hits backlog conversion.
COGS and tariff exposure; price spikes or supply disruption squeeze margins on long-lead orders.
~$2.4B FY26 tariff headwind cited; further escalation or relief swings margins materially.
Sales and service flow through dealers; relationship is durable but adds a layer between CAT and end-customer demand signals.
Two-thirds of revenue still rides global construction and resource capex and commodity prices.
Advantages
- Top-2 incumbent in data-center power with both backup gen-sets and prime-power turbines/engines under one roof.
- ~$63B record backlog (+79% YoY) gives unusual multi-year revenue visibility for an industrial.
- Unmatched ~150-dealer global service/parts footprint — the uptime guarantee hyperscalers require.
- High-margin recurring aftermarket that grows with every engine/turbine installed.
- Balance sheet and ~$11.7B operating cash flow large enough to self-fund the capacity tripling and keep returning ~$8B/yr to holders.
- Solar Turbines + reciprocating-engine engineering breadth lets CAT win across fuels and permitting paths.
Weaknesses
- Core Construction and Resource Industries are mature, cyclical and only growing low-single-digits.
- GAAP operating margin slipping YoY on tariffs (~$2.4B FY26 headwind) despite strong volume.
- Near-term data-center upside is supply-constrained until 2027-2029.
- Valuation (~49x trailing) leaves little room for any demand disappointment.
- Heavy macro/commodity sensitivity in the legacy two-thirds of the business.
- Consolidated reported debt (~$43-44B, mostly captive finance) can mislead on industrial leverage.
Bottlenecks
- Large-engine manufacturing capacity — demand exceeds supply until the 2027-2029 build-out lands; near-term revenue is gated by plant output, not orders.
- Long lead times and component supply (engines, turbines, switchgear) that stretch delivery into 2028 for some orders.
- Tariff and input-cost inflation compressing the GAAP margin even as volume grows.
- Skilled manufacturing labor and supplier ramp needed to triple engine output without quality slippage.
- Permitting/emissions approvals at customer sites that can delay prime-power project timing.
Top signals & trends
Top signals
The single best leading indicator of data-center revenue; watch for any deceleration or cancellations.
Direct read on data-center power demand converting to revenue.
On-time plant ramps unlock revenue; delays cap upside. Watch capex run-rate.
Persistent tariff drag (~$2.4B FY26) testing whether price/volume can hold margins.
Upstream demand signal; a capex digestion phase would pressure CAT's order book.
Industry-wide expansion into the same demand risks oversupply/price erosion by 2027-2029.
Trends
Structural demand displacing slow grid interconnects; the core re-rating driver for CAT.
Pushes data centers to behind-the-meter gas turbines/engines where CAT is strong.
Direct ~$2.4B FY26 margin headwind; ongoing uncertainty on input costs.
Expands product TAM but raises competition and emissions/permitting complexity.
Legacy two-thirds of revenue remains soft-to-flat; an upcycle is upside optionality, a downturn a risk.
Longer-term Resource Industries tailwind from electrification minerals.
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
Primary raw-material input; tariff-exposed COGS driver.
Electronics / power-electronics & controls suppliers Engine controls, switchgear, sensors for gen-sets and turbines.
Bearings / hydraulics / driveline component makers Specialized mechanical components for engines and machines.
Move heavy equipment and long-lead engines/turbines globally.
Hyperscaler; collaborated with CAT on data-center backup power incl. hydrogen fuel-cell pilot.
Hyperscaler data-center power buyer (backup + prime generation).
Hyperscaler driving on-site data-center power demand.
Hyperscaler scaling AI data centers requiring on-site power.
American Intelligence & Power (AIP) / Boyd CAT alliance Strategic alliance to deploy 2GW of dedicated hyperscale-AI power (2026); Monarch Compute Campus, WV.
Legacy core: contractors, miners, oil & gas for machines and power.
Closest data-center-genset rival; QSK95 high-speed engines, expanding Fridley MN capacity +30% (Feb 2026). Co-leader with CAT in data-center generators.
UK-listed; data centers now >80% of Power Systems revenue, taking 2027-2028 orders. Direct high-speed gas-engine competitor.
Gas-turbine leader (aeroderivatives from GE jet engines) for larger prime-power; order book into 2028. Competes at the turbine end vs Solar Turbines.
German-listed; one of the gas-turbine 'Big Three' accepting orders into 2028. Prime-power turbine competitor.
Japan-listed; gas-turbine 'Big Three' member; competes for large prime-power data-center projects.
US private; established backup-power gen-set supplier to data centers, competes at the smaller backup end.