Power Gen Secular Story Real but Priced; Tariff Drag and Cyclical Resets Offset — Initiating Coverage at Hold

Rating Action

This is our initiation of coverage on Caterpillar (CAT). We initiate at Hold based on a balance of forces: a confirmed secular Power Generation pull from data-center electricity demand that is real and multi-year, offset by an active tariff overhang, mid-cycle dealer-inventory and pricing resets in Construction Industries, and a multiple that already largely reflects the Power Gen narrative.

We are not initiating at Outperform because three of the most important variables — the run-rate FY tariff drag (still being negotiated globally), the duration and magnitude of the CI price-merchandising lap, and the realized cadence of new Power Gen capacity coming online in late 2026/2027 — will be materially better understood at the November 4 Investor Day and through the Q3/Q4 prints. We expect to revisit the rating when those data points clear. We are not initiating at Underperform because the backlog is at a record, ME&T free cash flow remains strong, capital return is healthy, and the underlying business momentum — particularly the order book in Power Gen, mining trucks, and Solar Turbines — argues against a contrarian short call.

Results vs. Consensus

A clean adjusted-EPS miss against a top-line beat. The composition is what matters more than the headline: revenue beat by ~$220M on better volume + financial products growth, EPS missed by ~$0.16 on tariff drag plus deferred-compensation expense plus unfavorable price realization in CI.

Segment Performance

Construction Industries (CI) — Cyclical Reset Mid-Lap

• Sales: $6.2B, -7% YoY; in line with internal expectations.
• Segment operating profit: $1.2B, -29% YoY; segment margin 20.1% (-600bps YoY).
• Tariff impact: ~170bps drag on margin; CI bears ~55% of the Q3 tariff impact.
• Regional sales: NA -15%, LatAm -20%, EAME +13%, APAC +6%. NA sales-to-users +3% (better than expected) on resi + non-resi growth, partially offset by lower rental fleet loading. EAME sales-to-users up on Africa + Middle East, weakness in Europe. APAC sales-to-users slightly down; China about flat YoY (below internal expectations after a strong Q1). LatAm sales-to-users down but better than anticipated.
• Dealer dynamics: Machine dealer inventory -$400M Q/Q (in line with expectations). Dealer rental revenue continued to grow despite lower rental fleet loading. The merchandising programs (attractive Cat Financial rates) are driving sales-to-users and pulling forward inventory absorption.
• Forward read: Bonfield framed the price headwind as halving in Q3 vs. Q2 and diminishing further in Q4, with some drag persisting into 2026. Q4 is set up favorably given the absence of last year's $1.6B dealer inventory destock and the continued pull-through from merchandising.

Resource Industries (RI) — Mining Truck Cycle Building

• Sales: $3.1B, -4% YoY; in line with expectations.
• Segment operating profit: $537M, -25% YoY; segment margin 17.4% (-500bps YoY).
• Tariff impact: ~230bps drag on margin; RI bears ~20% of the Q3 tariff impact.
• Mining detail: Sales-to-users -3%, in line. Mining slightly worse on off-highway truck delivery timing; heavy construction + quarry/aggregates slightly better. Full-year sales-to-users expected lower as customers maintain capital discipline. Coal price weakness driving an increase in parked trucks ⇒ lower rebuild activity in 2H.
• Backlog dynamics: Order rates strong, particularly in large mining trucks and articulated trucks. Coal exposure low single-digits of revenue and diminishing. Customer fleet utilization remains high; fleet age remains elevated. Autonomous solutions seeing continued demand acceptance — the long-cycle electric/autonomous truck story is still pre-revenue but in the order book.

Energy & Transportation (E&T) — The Engine

• Sales: $7.8B, +7% YoY; now CAT's largest segment by sales and profit.
• Segment operating profit: $1.6B, +4% YoY; segment margin 20.2% (-60bps).
• Tariff impact: ~110bps drag on margin; E&T bears ~25% of the Q3 tariff impact.
• Application breakdown (sales):
  • Power Generation: +28% YoY. Driven by reciprocating engines for data-center applications. Sales-to-users in PG up 19%; turbines and turbine-related services for PG slightly down on timing.
  • Oil & Gas: +2% YoY. Solar Turbines + turbine services growth offset by reciprocating engine softness in well servicing (customer capital discipline + consolidation). Gas compression reciprocating engines showing positive momentum.
  • Industrial: +1% YoY. Electric power applications driving growth from a low base.
  • Transportation: -7% YoY. Marine softness + international locomotive delivery timing.
• Solar Turbines: "Almost unprecedented interest" per Creed; the new Titan 350 platform is being well-received. Backlog remains strong; healthy order and inquiry activity.
• Capacity: Large reciprocating engine capacity additions are in process — Creed framed the bigger step-change as end-of-2026 / heading-into-2027, "not a cliff event" but continuous monthly improvement. Currently CAT is "not operating at the most efficient level" while balancing capacity build and demand fulfillment.

Financial Products (Cat Financial) — Healthy Credit, Tailwind from Merchandising

• Revenues: ~$1.0B, +4% YoY.
• Segment profit: $248M, +9% YoY.
• Credit quality: Past-dues 1.62% (-12bps YoY) — the lowest second-quarter level in over 25 years. Allowance rate 0.94%, near historic lows.
• Activity: Retail credit applications +5% YoY; retail new business volume +5% YoY (highest Q2 in over 10 years), reflecting attractiveness of merchandising programs.
• Used equipment: Inventory levels remain low; conversion rates remain above historical averages as customers buy at end of lease term.
• Read-through: Cat Financial is functioning as the demand-shaping lever for CI right now — the low-rate financing programs are pulling forward retail purchase decisions and ~50% of the merchandising program cost is recovered through interest margin over the financing life.

Key Topics & Management Commentary

Power Generation: Multi-Year Secular Pull, Capacity-Constrained Through 2027

Power Gen is the headline story this quarter and the principal driver of the Power Gen / data-center thesis investors are watching. Q2 sales-to-users in Power Gen grew 19%; total E&T Power Gen sales up 28%. The detail buyside investors will care about: CAT is taking orders "pretty extended" for data-center demand — planning with the largest data-center customers years in advance — with full visibility to schedules even into the post-capacity-expansion timeframe.

“We're taking orders pretty extended when it comes to data centers. ... We're planning with the largest data center customers years in advance. So we have line of sight to their schedules out into the time. ... When it comes to Solar, we continue to increase production there. We're pleased with the interest and uptake in the new Titan 350 platform. We're continuing to see, I would say, almost unprecedented interest in power generation when it comes to Solar Turbines.”
— Joe Creed, CEO

Read-through: The Power Gen pull is multi-year and pre-booked. The bottleneck is capacity, not demand. Capacity additions are linear monthly improvements with a step-change end-of-2026 / 2027. This is exactly the shape of cycle a long-cycle industrial wants — revenue visibility and limited downside — but it is also visible enough to the buyside that the multiple is unlikely to expand further on this story alone.

Tariffs: $1.3–1.5B Net for 2025, "No Regrets" Mitigation Posture

The defining variable on margins this year and likely 2026. The full-year 2025 net tariff impact estimate was raised to $1.3–1.5B (from prior framework that implied $2.6B gross before mitigation). The Q3 tariff cost is ~$400–500M with Q4 larger; CI bears ~55%, RI ~20%, E&T ~25% of the Q3 impact. Bonfield was direct that the mitigation taken so far is "no regrets" — cost controls, USMCA certification, limited dual-sourcing — not the longer-cycle moves (sourcing changes, footprint relocation) that would require committing capital under uncertainty:

“We've learned a little bit more, obviously, in the last week or so in some of the announcements that are out. We're still working to understand the implications of those in the details. ... Some of the actions are quicker for us to implement and quicker for us to see results versus moving a footprint requires some investment and can take a significant amount of time.”
— Joe Creed, CEO

Importantly, Creed and Bonfield both declined to characterize any portion of the tariff drag as "structural" or "permanent" — with Creed: "We're not thinking in those terms right now." The implicit posture: ride the uncertainty, take "no regrets" actions, defer the bigger moves. That is the right posture for shareholders if you believe the tariff regime is itself negotiable; it is the wrong posture if the tariff regime is durable and competitors with different supply chains are not equally exposed.

Backlog Coverage at Historically Unprecedented Level

The Raso (Evercore ISI) question on the call was the analytically richest moment: backlog coverage of next twelve months shipping is over 80% of implied 2H sales, vs. a historical average of ~55%. The follow-up was whether CAT would reprice the backlog to capture margin. Creed's answer was nuanced — CAT has flexibility on backlog pricing depending on segment/product, but the focus is on dollar OPACC (operating profit after capital charge) rather than margin alone, and the merchandising lap will reduce the price headwind through Q3 ⇒ Q4 anyway:

“We're going to get the balance right. ... When it's merchandising programs with Cat Financial and rates over time, that's beneficial for us. ... We haven't made any decisions definitively at this point in time. ... I'd like to see how much we can mitigate on tariffs, specifically other ways before we use pricing as a lever.”
— Joe Creed, CEO

Bonfield's add-on was the most thesis-relevant comment of the call: “What it does mean is we have good momentum going into 2026, which is the most important thing. ... Operating leverage is always the most optimal way to manage margins for a company like Caterpillar.” The implication: 2026 setup is for operating leverage as price headwinds lap, capacity comes online, and tariff drag (presumably) stabilizes. Whether that translates into a Power-Gen-priced multiple expansion or a more muted cyclical normalization is the question we cannot yet answer.

Construction Industries: Cyclical Reset Lapping in Q3/Q4

The Feniger (Bank of America) and Fisher (UBS) questions clarified the CI lap. Bonfield: Q3 expected strong with inventory loading + sales-to-users growth; Q4 absent the prior-year $1.6B destock should be the strongest setup in the segment. Price headwind halves in Q3, narrows further in Q4, with residual 2026 drag. Rental fleet loading is expected to normalize after a period of below-trend loading — Bonfield: “Rental revenue has been rising. ... Dealers tend to do it at points in time, depending on particularly around their heavy rents. ... We expect now to start to see that come back to a more normalized level as we move into the second half of this year.”

Capital Allocation: $1.5B Returned, Investor Day November 4

$800M of buybacks + dividend in Q2; June dividend +7% (fifth consecutive year of high-single-digit increases). Net debt in ME&T $5.2B with cash + marketable securities of $6.6B. New $2B bond issuance at attractive rates. ME&T FCF guide of ~$7.5B for 2025 implies continued capital return capacity. Creed: “I look forward to sharing more about our strategic priorities and the incredible growth opportunities that lie ahead during our upcoming Investor Day in November.” November 4 will be a meaningful catalyst — the multi-year algorithm framework, mid-cycle margin targets, and Power Gen capacity cadence are all candidates for the agenda.

FY 2025 Guidance & Outlook

Q3 specifics: Sales to grow moderately YoY, volumes higher in all three primary segments. Q3 net tariff cost ~$400–500M (CI 55% / RI 20% / E&T 25%). Q3 enterprise adj. operating margin ex-tariffs similar to prior year; including tariffs lower YoY. CI segment margin lower YoY incl. tariffs; RI lower YoY incl. tariffs (price + SG&A/R&D); E&T similar YoY incl. tariffs (volume + price offset by tariffs).

Analyst Q&A — Notable Exchanges

Q&A was unusually tariff-heavy and CI-cyclical-heavy this quarter, reflecting where the buyside is uncertain. Notable threads:

• Tami Zakaria (JPMorgan) opened with the long-term tariff structuralization question. Creed's framing: U.S. footprint already large (50,000+ employees, 65 locations across 25 states, exports +75% since 2016, hourly workforce +29%), supply chain optimized for global competitiveness; "all options on the table" but not yet committing capital under uncertainty. Read: long-term tariff drag treated as negotiable, not structural.
• David Raso (Evercore ISI) drew the most analytically rich exchange: backlog coverage at 80%+ of next-12-month shipping is historically unprecedented (vs. ~55% average). Asked if CAT would reprice the backlog. Creed declined to commit but acknowledged flexibility; Bonfield's "good momentum into 2026 ... operating leverage" was the closest thing to a 2026 pre-announcement on the call.
• Jamie Cook (Truist) on E&T capacity additions and the trajectory through 2026/2027. Creed framed capacity additions as continuous improvement with a step-change end-of-2026 heading into 2027; current operations not at most efficient levels while building capacity; throughput continuing to increase.
• Robert Wertheimer (Melius) probed Solar Turbines specifically — the orders being placed for expanded recip capacity and the Solar capacity expansion. Creed: data-center orders pretty extended; Solar production continuing to increase; "almost unprecedented interest" in Solar Turbines for power generation.
• Kristen Owen (Oppenheimer) on remaining tariff uncertainty. Bonfield: only a limited number of country agreements made; Section 232 / 302 investigations still pending; situation remains very fluid.
• Chad Dillard (Bernstein) on the dealer inventory build assumption (flat YoY). Bonfield: planning assumption based on current order rates from dealers; close to the period where new dealer orders fill by end-of-year. Creed: "different seasonal pattern" given strong sales-to-users; absence of 2H drawdown is itself a relative tailwind.
• Mig Dobre (Baird) probed long-term tariff "permanent" margin question and 2026 setup. Creed politely declined to commit to a 2026 view; "great operational performance now ... great momentum ... will update everyone on 2026 later in the year."
• Steve Volkmann (Jefferies) on competitive PINS (population-in-service) dynamics. Creed: focused on customer service; merchandising programs getting strong response especially in CI; sales-to-users continuing to be up in NA and globally for CI despite global construction softness.
• Michael Feniger (Bank of America) on CI exiting the year — given the prior-year $1.6B Q4 destock and current STU strength. Bonfield confirmed strong Q4 CI setup; price headwind halving in Q3 and narrowing further in Q4 with some 2026 drag.
• Steven Fisher (UBS) on rental fleet loading and the merchandising program logic. Bonfield: low-interest financing recovers ~50% of program cost through margin over the financing life; rental loading expected to normalize in 2H.
• Kyle Menges (Citigroup) on RI backlog visibility into 2026 and coal exposure. Creed: STUs softer 2H but order rates healthy and backlog up, particularly in large trucks and articulated trucks; coal revenues low-single-digits of total and diminishing.

What They’re NOT Saying

• No 2026 framework. Multiple analysts probed; Creed declined to commit, even directionally. Bonfield's "good momentum into 2026 ... operating leverage" was the closest the call came. The November Investor Day will be the natural setting for a 2026 framework reveal — until then, treat the multi-year picture as analytically open.
• No quantification of how much pricing power exists in the backlog. Raso's question was not answered directly. The implication that CAT could reprice the existing backlog to capture margin (effectively passing tariff costs through) was politely deferred — meaning either CAT does not yet believe it has the unilateral pricing power vs. competitors, or it does not want to signal that capacity until the merchandising programs lap. Either way, this is upside optionality not yet priced.
• No specific capacity targets for 2026 / 2027 reciprocating engine capacity. Creed gave qualitative shape ("step-change end-of-2026 / heading into 2027 ... not a cliff event") but no quantification of incremental shipment capacity. The November Investor Day is the natural venue for this disclosure.
• No commentary on autonomous truck revenue runway. "Continued demand and customer acceptance" was the framing; no quantitative pipeline. The autonomous and electric mining truck cycle is in early innings and not visibly contributing to the algorithm yet.
• No commentary on China beyond "about flat." Notable given the China data-center capex theme on the buyside. Creed: China was about flat YoY but below internal expectations after a stronger Q1; APAC outside China remains soft.
• No mention of buyback pacing change. $800M repurchases in Q2 — not above or below recent quarters. The implicit signal: continued balanced capital return. Investor Day will reveal whether the Power Gen capacity cycle pulls capital intensity higher (and buybacks lower) at the margin.

Market Reaction

The print landed before the open on August 5, 2025. Pre-market reaction was modestly negative on the EPS miss and the raised tariff guidance — press reports cited a 1–3% premarket dip. Through the regular session, the stock recovered some ground as the buyside digested the record backlog, the raised top-line, and the constructive Power Generation commentary — with the close roughly flat to modestly down on the day. The principal market debate was the same one we are weighing in our rating decision: how much of the Power Gen secular pull is already priced, and how durable is the tariff drag. The next data points (Q3 print + November Investor Day) are the meaningful catalysts.

Street Perspective

The bull case being made on the Street post-print converges on three planks: (1) Power Generation is a multi-year secular pull with multi-year visibility into data-center customer schedules and a binding-capacity rather than binding-demand bottleneck through 2027; (2) the record $37.5B backlog — with 80%+ coverage of next-twelve-month shipping — is a historically unprecedented level of revenue visibility; (3) the FY 2025 top-line raise plus the Q4 setup (absent prior-year $1.6B CI destock) signals a constructive 2026 setup as price headwinds lap and capacity comes online.

The bear case being articulated on the Street centers on: (1) the $1.3–1.5B 2025 tariff drag is itself a function of trade-policy negotiation, not a stable input — if the regime hardens, the structural margin level reset is bigger than today's print implies; (2) the Power Gen / data-center narrative is broadly understood and the multiple does not have meaningful room for further expansion on this story alone; (3) Construction Industries is mid-cyclical-reset; the merchandising lap is only halfway through and the rental fleet reload assumption is not yet confirmed; (4) Resource Industries demand discipline among mining customers continues, and the autonomous/electric truck cycle is still pre-revenue contribution.

Our read sides closer to the bear framing on (1) and (2) and the bull framing on (3) — and concludes that a Hold rating is the appropriate stance until tariff stabilization and the November 4 Investor Day are in hand.

Model Implications

• FY 2025 revenue: We mark the trajectory implied by Q1+Q2 + the raised "slightly higher" full-year guide as supportive of $66–67B FY 2025 sales & revenues, with E&T pulling and CI/RI offsetting.
• FY 2025 adjusted operating margin: We model the bottom-half of the target range (high-teens) inclusive of the $1.3–1.5B net tariff impact. If the tariff impact reverses lower, we have margin upside; if it deteriorates further, we have margin downside.
• FY 2025 ME&T FCF: ~$7.5B as guided; supportive of continued ~$1.5B/Q capital return cadence (split buybacks + dividend).
• FY 2026 framework (preliminary): Operating leverage as the principal margin recapture lever; we underwrite mid-single-digit revenue growth driven by E&T capacity ramp + CI normalization, with margin recovery dependent on tariff stabilization. Solar Turbines and reciprocating engine capacity step-change is a 2026 exit / 2027 contributor.
• Backlog conversion: Record $37.5B backlog with 80%+ coverage of next-12-month shipping translates into limited 2H 2025 revenue downside risk and supportive 1H 2026 visibility.
• Capital return: We model continued $1.5–2.0B/Q capital return through 2026 absent a major M&A or capacity acceleration announcement at Investor Day.

Thesis Scorecard

Overall: Three of four bull pillars confirmed. The fourth (CI inventory rebalance) is tracking but not complete. Bear case is intact — tariff uncertainty + Power Gen narrative being priced + CI cyclical reset are real and offsetting. We initiate at Hold.

Action: Initiating coverage of Caterpillar (CAT) at Hold. We expect to revisit the rating after the Q3 print and the November 4 Investor Day, which together should clarify (a) tariff trajectory stabilization, (b) CI inventory rebalance completion, (c) Power Gen capacity ramp cadence, and (d) the multi-year algorithm framework. We do not chase — nor do we short — on the current setup.

Net: A best-in-class cyclical industrial leader with a real Power Gen secular pull and a record backlog, offset by an active tariff overhang and mid-cyclical CI/RI resets. Hold pending Q3 + Investor Day confirmation.