Cummins Inc. (CMI) - Deep Dive Research Report

Prepared 2026-06-28. Most recent reporting period: Q1 2026 (results released 5 May 2026). Six earnings calls used: Q1 2026 (5 May 2026), Q4 2025 (5 Feb 2026), Q3 2025 (6 Nov 2025), Q2 2025 (5 Aug 2025), Q1 2025 (5 May 2025), Q4 2024 (4 Feb 2025).

1. What the Company Does

Cummins makes the engines and power equipment that move heavy things and keep the lights on. When a Class 8 long-haul truck hauls freight across an American interstate, when a mine in Chile runs a 400-ton haul truck, when a hospital or a hyperscale data center loses grid power and has to ride out the outage on its own, there is a very good chance a Cummins engine or generator is doing the work. The company designs, builds, sells, and then services for decades the large internal-combustion powertrains that the global economy still runs on, plus the turbochargers, fuel systems, filtration, and exhaust after-treatment that go around them.

The business was founded in Columbus, Indiana in 1919 by Clessie Cummins, a self-taught mechanic, with financial backing from local banker W.G. Irwin. Cummins's bet was on the diesel engine - then a clumsy, industrial-only technology - as the future of mobile power. The company spent its first two decades proving diesels could be reliable and compact enough for trucks, and the rest of the century becoming the largest independent (non-captive) maker of diesel engines in the world. That word "independent" is the whole strategy: unlike Daimler or Volvo, which build engines mainly for their own trucks, Cummins sells engines to almost everyone, including direct truck-maker competitors. A PACCAR Peterbilt or a Stellantis Ram can be ordered with a Cummins engine inside.

The core value proposition is uptime. A heavy-duty engine is a capital asset that has to run hundreds of thousands of miles, and the cost of it failing - a stranded truck, a dark data center, a stopped mine - dwarfs the price of the engine itself. Cummins sells reliability backed by the densest service and parts network in the industry, so that wherever the equipment breaks, a Cummins dealer can fix it fast. This is why the company's most profitable, most stable businesses are not the engines themselves but the aftermarket: parts, filters, rebuilds, and service that recur for the entire 15-to-30-year life of the installed base.

Concretely, walk through what Cummins does for a fleet operator. The fleet buys 500 trucks from a truck OEM and specs them with the Cummins X15 engine. Cummins supplies the engine, the turbocharger (its own Holset brand), the fuel system, the electronic controls, and the emissions after-treatment as an integrated package. Over the next decade, every one of those 500 engines needs filters, injectors, and eventually a mid-life rebuild, all flowing through Cummins's Distribution arm. When EPA emissions rules tighten, Cummins engineers the next-generation engine that meets them, and the fleet upgrades. The customer is locked into a relationship that lasts far longer than a single truck purchase.

The newest layer is the energy transition. Cummins runs a zero-emissions unit called Accelera (electrolyzers, fuel cells, battery-electric powertrains) and is building a fuel-agnostic engine platform called HELM (Higher Efficiency, Lower emissions, Multiple fuels) so the same engine block can run on diesel, natural gas, or hydrogen. The thesis is that decarbonization will be gradual and messy, and Cummins wants to sell power through the whole transition rather than be disrupted by it.

2. Business Segments

Cummins reports in five segments: Engine, Components, Distribution, Power Systems, and Accelera. The first four are large, profitable, and interlocking; Accelera is a small, loss-making strategic option on the energy transition.

Engine

The Engine segment designs and builds the mid-range and heavy-duty engines that go into on-highway trucks (heavy-duty Class 8, medium-duty), buses, recreational vehicles, light-duty pickups (the Ram/Stellantis relationship), and a range of off-highway industrial equipment (construction, agriculture). This is the historical heart of the company and its most cyclical part, because it rises and falls with North American truck production. In Q1 2026, Engine segment EBITDA margin fell to 10.4% from 16.5% a year earlier as North America truck volumes dropped roughly 20% (Q1 2026 concall, 5 May 2026), a swing that shows how violently this segment moves with the freight cycle.

The core capability is decades of accumulated combustion, durability, and emissions engineering. Building an engine that meets ever-tightening EPA and CARB emissions standards while still delivering fuel economy and a million-mile service life is genuinely hard, and the certification process is long and expensive. That regulatory engineering moat is also a recurring source of risk and product timing: in Q1 2026 management delayed the new B-series engine launch to January 2028 because of uncertainty over future EPA rules (Q1 2026 concall, 5 May 2026).

It exists as its own segment because its economics (cyclical, volume-driven, OEM-customer) are distinct from the recurring aftermarket and the high-horsepower industrial businesses. Within trucks Cummins competes against the captive engine programs of its own customers (PACCAR's MX engine, Daimler's Detroit Diesel, Volvo). It wins on being the spec customers ask for by name and on aftermarket support; it loses share when OEMs push their captive engines. Management frames Engine as a cash-and-cycle business: ride the downturns, hold share, and use pricing and aftermarket to protect margins.

Components

Components makes the things that bolt onto and around the engine: turbochargers (Holset), exhaust after-treatment and emissions systems, fuel systems, electronics, and - since the 2022 Meritor acquisition - axles, brakes, and drivetrain. The Meritor deal ($3.7 billion, completed August 2022) was the largest in company history and pushed Cummins from "the engine" toward "the integrated powertrain," giving it the driveline content to sell complete propulsion systems across both combustion and electric platforms.

The core capability is precision sub-system engineering tied tightly to emissions regulation - after-treatment in particular is a regulatory-driven business that grows in content every time emissions rules tighten. It exists separately because much of it is sold to other engine and equipment makers, not just Cummins's own Engine segment, and because the Meritor axle/brake business has its own customer base and competitive set (Dana, ZF, BorgWarner, Bosch on the components side). Components, like Engine, is exposed to the truck cycle but carries higher structural content per vehicle as emissions rules escalate. Management treats it as a content-growth story riding on regulation.

Distribution

Distribution is the parts, service, and power-generation distribution arm: it sells and services the entire Cummins range through company-owned and independent channels, and it is where the recurring aftermarket revenue lives. In Q2 2025 it posted record results, with revenue up 7% to about $3 billion and EBITDA margin at a record 14.6% (Q2 2025 concall, 5 Aug 2025), and it kept setting records through Q1 2026 at 14.2% margin (Q1 2026 concall, 5 May 2026).

The core capability is the network itself - the global footprint of branches, technicians, and parts inventory that no new entrant could replicate quickly. This is the structural counter-cyclical ballast of the company: when truck sales fall, the installed base still needs servicing, so Distribution holds up while Engine collapses. It exists separately precisely because its economics (recurring, service-driven, less cyclical) are the opposite of the OEM segments. Increasingly it is also a direct beneficiary of the data center boom - roughly half of Cummins's data-center-related revenue flows through Distribution as power-generation equipment and service (Q3 2025 concall, 6 Nov 2025). Management talks about Distribution as the steady compounder and the tailwind beneficiary of every engine and genset Cummins puts into the field.

Power Systems

Power Systems builds high-horsepower engines (the large 50-to-95-plus-liter range) and power generation equipment - the diesel and gas gensets that provide backup and prime power for data centers, mining, oil and gas, marine, and standby applications. This has become the standout growth and margin engine of the company. In Q1 2026 it delivered record results at a 29.5% EBITDA margin, up from 23.6%, driven by data center demand that grew 23% in North America and 84% in China (Q1 2026 concall, 5 May 2026). Power generation revenue growth guidance for 2026 was raised to 15-25% on accelerating data center demand (Q1 2026 concall, 5 May 2026).

The core capability is high-horsepower engineering and the ability to deliver mission-critical, certified, reliable power at scale - a data center operator cannot afford a genset that fails to start when the grid drops. Lead times for the largest units (the QSK95) had stretched to around 18 months, prompting capacity expansion. It exists as a separate segment because the customers (hyperscalers, miners, utilities), the product (industrial-scale power), and the economics (high margin, project-driven, currently supply-constrained) are entirely distinct from on-highway trucks. Within data center gensets its principal rival is Caterpillar; the two together hold over a third of the global market. Management has made Power Systems the clear strategic priority and the primary vehicle for capturing the AI-infrastructure power buildout.

Accelera

Accelera is the zero-emissions unit: electrolyzers (for green hydrogen), fuel cells, battery-electric powertrains, and eMobility components. It is small (revenue in the low hundreds of millions) and loss-making, with 2026 losses guided down to $270-300 million (Q1 2026 concall, 5 May 2026). The core capability is a portfolio of early-stage clean-power technologies and the engineering to integrate them; its competitors are the hydrogen and electrification specialists - Plug Power, Ballard Power Systems, Nel, Siemens Energy on electrolyzers, and a crowded field on battery-electric.

It exists separately because it is a venture-style strategic option with completely different economics from the cash-generative core, and management has been candid that it is not on its original path. The electrolyzer market, in CFO Mark Smith's words, "dried up faster than anything I have seen in my career" (Q3 2025 concall, 6 Nov 2025), forcing impairment charges and a sharp narrowing of focus. Through 2025-2026 management has been actively shrinking the losses: reorganizing the unit, pacing electrolyzer investment, and in Q1 2026 selling the low-pressure fuel cell business to Alstom. The strategic role is a hedge - keep a seat at the decarbonization table without letting the losses become material to the group.

Segment summary

3. Products and Business Detail

The product catalogue spans from a 2.8-liter pickup engine to a 95-liter data center genset.

On-highway engines. The flagship is the X15, the heavy-duty engine specced into Class 8 long-haul trucks across multiple OEM brands. Below it sit the medium-duty B and L series engines used in delivery trucks, vocational vehicles, buses, and RVs. The B6.7 and B-series mid-range engines also power Ram pickups through the long-running Stellantis relationship. The forward roadmap is the HELM platform - a fuel-agnostic block engineered so the same base engine can be configured for diesel, natural gas, or hydrogen combustion, letting fleets decarbonize without re-architecting their vehicles. The new B-series under HELM was delayed to January 2028 over EPA rule uncertainty (Q1 2026 concall, 5 May 2026).

High-horsepower and power generation. The QSK and QSK95 high-horsepower engines anchor the Power Systems range, powering mining haul trucks, drilling, marine vessels, and large gensets. The data center genset products combine these large engines with alternators and controls into mission-critical standby and prime-power systems. Demand here has run ahead of capacity: lead times on the biggest units stretched to roughly 18 months, and Cummins announced a roughly $150 million expansion of its Fridley, Minnesota high-horsepower facility (early 2026) to lift QSK95 output by about 30%, plus $200 million of investment across U.S., England, and India plants for large data-center engines flagged at Q4 2024 (Q4 2024 concall, 4 Feb 2025).

Components. Holset turbochargers, exhaust after-treatment and selective catalytic reduction systems, fuel systems, electronic controls, and - from Meritor - axles, brakes, and drivetrain. Emissions content per vehicle ratchets up each regulatory cycle, which is the structural growth mechanic of this segment.

Accelera. PEM and alkaline electrolyzers for green hydrogen, hydrogen fuel cells, battery-electric powertrains, and eMobility traction systems. Notable references include a 100-megawatt electrolyzer order for bp's German green-hydrogen project, the largest Cummins had assembled (Q1 2025 concall, 5 May 2025).

Manufacturing and process knowledge. Engines are built across a global plant network with deep machining, casting-tolerance, and assembly know-how; the hard part is meeting emissions certification (EPA, CARB, Euro standards) while preserving durability, which requires years of validation testing before a product launches. This is why product timing is regulation-bound and why a delayed EPA rule cascades directly into a delayed engine launch.

Geographies. Cummins is genuinely global. North America remains the largest market and the source of the truck cyclicality. China is now a major and fast-growing market: in Q1 2026 China revenue rose 19% to $2.1 billion, with power generation up 84% and truck sales up 14% (Q1 2026 concall, 5 May 2026). India is both a manufacturing base and a growth market (through Cummins India, a separately listed entity). Europe rounds out the manufacturing and demand footprint.

Corporate-structure milestones that shaped today's company. The 2022 Meritor acquisition added driveline content. The 2023 IPO and 2024 full split-off of Atmus Filtration Technologies exited the filtration business and sharpened the portfolio. The rebranding of the zero-emissions unit as Accelera consolidated the clean-power bets. And the 2026 sale of the low-pressure fuel cell business to Alstom continued the pruning of Accelera's losses.

4. Customers

Cummins's customers fall into four broad groups, each with a different buying logic.

Truck and equipment OEMs. The most important on-highway relationship is with truck makers - PACCAR, Stellantis (Ram), and other OEMs that offer Cummins engines as a spec option, plus off-highway equipment makers. The buying decision is split: the OEM has to qualify and integrate the engine, but the end fleet often demands Cummins by name because of resale value and serviceability. Sales cycles are long (a new engine program is designed in years ahead of production), and once an engine is designed into a vehicle platform it stays there for the platform's life. This is real lock-in: re-engineering a truck around a different engine is expensive and slow.

Fleet operators and the aftermarket. The fleets that actually run the trucks are the recurring-revenue customers. They buy filters, parts, injectors, and rebuilds through Distribution for the entire life of the engine. The decision criterion is uptime and total cost of ownership, and the switching cost is the installed base itself - you service what you own. This is the stickiest, most predictable revenue in the company.

Data center and hyperscale operators. The fastest-growing customer set. Hyperscalers and colocation operators buy large gensets for backup and increasingly prime power. The buyer here is a data-center infrastructure and facilities team, and the criteria are reliability, certified performance, lead time, and the ability of the supplier to deliver at scale on a build schedule. With lead times around 18 months and supply constrained, the relationship currently favors the supplier. Switching costs are high because the genset is integrated into the facility's power architecture and the operator standardizes on a vendor for service.

Mining, oil and gas, marine, and standby power. High-horsepower industrial customers who buy for harsh-duty reliability and a global service network. These are project-driven, relationship-based purchases.

Concentration and contract structure. No single customer dominates the way one buyer can dominate a pure-play supplier, but the on-highway OEM relationships (especially Stellantis/Ram and the major truck makers) are individually large and concentrate the truck-cycle risk. The revenue mix blends OEM program business (cyclical, won years in advance), recurring aftermarket (predictable, the ballast), and increasingly multi-year data-center supply commitments (the new growth backbone). The aftermarket layer is what makes the overall revenue base far steadier than the truck cycle alone would suggest.

5. Competitive Landscape

Cummins sits in several distinct competitive arenas at once, and its position differs sharply by segment.

In on-highway truck engines, the central structural fact is that Cummins is the largest independent engine maker selling into an industry where its biggest customers also build their own engines. PACCAR (the MX engine), Daimler Truck (Detroit Diesel), and Volvo Group all have captive engine programs and want to migrate volume to them. Cummins wins when fleets demand the X15 by name for resale value and serviceability; it loses share when OEMs successfully push captives. This is a slow, grinding share battle inside a cyclical market.

In data center and high-horsepower power generation, the duopoly-like reality is Cummins versus Caterpillar, who together command over a third of the global data-center backup-generator market (per market research, Cummins ~16%, Caterpillar ~18%). This is currently the most attractive arena Cummins competes in: demand is exploding with AI infrastructure, supply is constrained, and both incumbents have pricing power. Generac competes at the smaller commercial and residential end. Increasingly, Chinese players are intensifying competition within China itself, which management has flagged.

In components, Cummins competes against Dana, ZF, BorgWarner, and Bosch on driveline and turbo/fuel-system content - a more fragmented, competitive arena where regulation-driven content growth is the prize.

In Accelera (zero-emissions), the field is crowded and currently brutal: Plug Power, Ballard, Nel, and Siemens Energy on hydrogen, plus a wide field on battery-electric. This is where Cummins is weakest and where the market itself contracted sharply.

Barriers to entry are high in the core: emissions certification, the multi-decade service network, the installed base, and the capital intensity of high-horsepower manufacturing. They are far lower in clean-tech, which is exactly why Accelera has struggled.

Where Cummins is strong: the aftermarket network, high-horsepower power generation into data centers, and the independent-supplier position. Where it is exposed: the truck cycle (deeply), captive-engine share erosion at OEMs, China competition, and the structurally weak clean-tech segment.

6. Industry

Cummins operates at the intersection of three industries: commercial trucking, power generation, and industrial off-highway equipment.

Demand drivers. On-highway demand is driven by freight volumes, fleet replacement cycles, and emissions-regulation timing (which pulls demand forward into "prebuys" ahead of new rules). Power generation demand is driven by electrification, grid reliability concerns, and - the dominant new force - the AI data center buildout, which requires enormous quantities of reliable backup and prime power. Off-highway demand tracks mining, construction, and commodity cycles.

Industry size and trajectory. The data center generator market is the high-growth pocket: estimates put it at roughly US$7.6-10 billion in 2025-2026 and growing at a mid-single to high-single-digit CAGR through the early 2030s, with some forecasts pointing toward roughly US$17-20 billion by 2034. Diesel gensets still dominate (around 81% of data-center generator share in 2025), with hydrogen and HVO-ready units the fastest-growing niche. AI clusters that can draw up to 100 kilowatts per rack are forcing operators to redesign backup-power architecture, which is structurally favorable to high-horsepower suppliers. Cummins and Caterpillar together hold over 35% of this market.

Supply chain position. Cummins sits upstream of the truck and equipment OEMs (it supplies the propulsion) and is a tier-one prime mover supplier to data centers and industrial customers. In power generation it is close to the end customer; in trucks it sells through OEMs.

Regulation. Emissions regulation is the defining structural force. EPA and CARB standards (the EPA 2027 rule and beyond) drive engine redesign cycles, content growth in after-treatment, and the timing of prebuys. Regulatory uncertainty is double-edged: it can pull demand forward (prebuy) or, as in 2025-2026, stall product launches when the rules themselves are in flux. Management's working view through this period was that "EPA27 regulations will stay in place," with challenges more likely aimed at 2030+ greenhouse-gas standards (Q4 2024 concall, 4 Feb 2025), but the uncertainty was enough to delay the B-series launch to 2028.

Cyclicality. The on-highway business is highly cyclical and swings hard with North American truck production. Through 2025 that cycle turned sharply down - heavy- and medium-duty volumes fell 25-40% in some quarters, with CFO Mark Smith calling recent order activity "amongst the weakest 3-or-4-month periods we've had in the last 20 years" (Q2 2025 concall, 5 Aug 2025). The aftermarket, Distribution, and Power Systems are far less cyclical and provide ballast through these troughs.

Tailwinds and headwinds. The dominant industry tailwind is data center power demand. The dominant headwinds are the truck-cycle trough, tariff-driven cost and demand uncertainty (a recurring theme through 2025), and the collapse of the green-hydrogen/electrolyzer market that hit the clean-tech sub-industry.

7. Growth Triggers

All items below are drawn directly from the six concalls and attributed.

• Data center power generation ramp. Repeated across all six calls and accelerating. Power generation guidance was raised to 15-25% growth for 2026 from a prior 10-20% (Q1 2026 concall, 5 May 2026). Earlier, management guided data-center-related revenue up 30-35% for 2025 (Q3 2025 concall, 6 Nov 2025) and confirmed a roughly $2 billion data center sales target for 2026 (Q2 2025 concall, 5 Aug 2025).

  "U.S. and China have been areas where we've seen a lot of growth" in data center applications (Q1 2025 concall, 5 May 2025).

• High-horsepower capacity expansions to clear the backlog. A ~$150 million Fridley, Minnesota expansion to lift QSK95 output ~30% (flagged early 2026), plus $200 million across U.S., England, and India plants for large data-center engines (Q4 2024 concall, 4 Feb 2025). These commissionings convert an 18-month order backlog into shippable revenue.

• China strength in both power generation and trucks. China revenue up 19% in Q1 2026, with power gen up 84% and truck sales up 14% (Q1 2026 concall, 5 May 2026) - a repeated growth theme.

• North America truck recovery / EPA 2027 prebuy. Management raised its 2026 North America heavy-duty truck unit forecast to 230,000-250,000 from 220,000-240,000 (Q1 2026 concall, 5 May 2026), expecting a second-half improvement, having earlier framed a prebuy ahead of EPA 2027 (Q4 2024 concall, 4 Feb 2025).

• HELM multi-fuel engine platform and new B-series. New engine platforms targeting EPA 2027 compliance, with about $1 billion invested in engine plants (Q2 2025 concall, 5 Aug 2025); the B-series launch is now set for January 2028 (Q1 2026 concall, 5 May 2026).

• Accelera loss reduction. 2026 Accelera losses guided down to $270-300 million from a prior $325-355 million projection, helped by the low-pressure fuel cell sale to Alstom (Q1 2026 concall, 5 May 2026).

• Distribution aftermarket compounding on a growing installed base. Management describes installed-base expansion as "a tailwind for the distribution business" (Q3 2025 concall, 6 Nov 2025).

8. Key Risks

The truck cycle is severe and Cummins is deeply exposed to it. The Engine and Components segments swing hard with North American truck production, and the 2025 downturn was one of the worst order environments in two decades. Engine segment margin collapsed from 16.5% to 10.4% year over year in a single cycle turn (Q1 2026 concall, 5 May 2026). The mechanism is direct: freight weakness and uncertainty cut fleet orders, OEM build rates fall, and Cummins's engine and component volumes fall with them. This is a high-probability, recurring drag rather than a tail risk - it will happen again every cycle.

EPA regulatory uncertainty stalls product launches and distorts demand. Emissions rules drive Cummins's entire product cadence. When the rules are in flux, launches slip - the B-series moved to January 2028 explicitly because of EPA uncertainty (Q1 2026 concall, 5 May 2026). The same uncertainty makes the prebuy thesis unreliable: if EPA 2027 timing shifts, the demand pull-forward Cummins has been counting on may not materialize as expected. This is a moderate-probability, moderate-impact risk that is largely outside management's control.

Data center concentration cuts both ways. Power Systems is now the margin and growth engine, and it is increasingly tied to one demand source: hyperscale data center capex. If the AI-infrastructure buildout slows, normalizes, or shifts toward alternative power architectures, the segment carrying the company's record margins would decelerate just as the truck cycle remains weak. This is currently a low-probability but high-impact risk - the very strength that is masking the truck trough could reverse.

Tariffs and trade policy. Tariffs were a persistent 2025 headache - roughly $22 million net EBITDA hit in Q2 2025 (Q2 2025 concall, 5 Aug 2025) - and severe enough that management withdrew full-year guidance entirely at Q1 and Q2 2025. CEO Jennifer Rumsey framed it plainly:

"We are entering unchartered territory as the trade tariffs start to have a more significant impact beginning in the second quarter" (Q1 2025 concall, 5 May 2025). The direct cost impact has been manageable, but the second-order demand and planning uncertainty is the larger danger.

China competition. China is a major growth source, but management has flagged intensifying local competition in data center power there. A Chinese market that grows but commoditizes would deliver volume without the margins Cummins earns in North America.

Accelera as a persistent cash drain and strategic misjudgment. The electrolyzer market collapse - "dried up faster than anything I have seen in my career" (Q3 2025 concall, 6 Nov 2025) - forced impairments and showed that Cummins's clean-tech bets can destroy value. The risk is moderate in size (losses are being cut toward $270-300 million) but it is a reminder that the energy-transition hedge is not free.

9. Walk the Talk

The six concalls used: Q4 2024 (4 Feb 2025), Q1 2025 (5 May 2025), Q2 2025 (5 Aug 2025), Q3 2025 (6 Nov 2025), Q4 2025 (5 Feb 2026), and Q1 2026 (5 May 2026). The most recent is within 90 days of today.

Start with Q4 2024 (4 Feb 2025), where management laid out a detailed 2025 plan: revenue down 2% to up 3%, EBITDA margin 16.2-17.2%, Power Systems margin expanding to 19-20%, North America heavy-duty trucks at 260,000-290,000 units, and Accelera losses narrowing to $385-415 million. This was a conservative, wide-banded guide that explicitly flagged truck-cycle and EPA-timing uncertainty.

Then reality intervened. At Q1 2025 (5 May 2025), management withdrew full-year guidance entirely, citing the tariff shock. This is the kind of move that tests credibility, but it was handled honestly rather than papered over - they named the cause and did not pretend to a precision they no longer had. Crucially, the operating business held: Power Systems hit a then-record 23.6% EBITDA margin even as truck volumes fell 21%. The promise that Power Systems was a durable margin engine was being kept even as the macro guide was abandoned.

At Q2 2025 (5 Aug 2025), management still declined to reinstate guidance - a sign of genuine uncertainty rather than convenient vagueness - and the truck market deteriorated further, with Smith describing orders "amongst the weakest 3-or-4-month periods we've had in the last 20 years." But two things they had consistently promised continued to deliver: Power Systems set another record (22.8% margin) and Distribution set a record (14.6% margin). And they backed up the capital-return commitment by raising the dividend 10% to $2.00 per quarter, the 16th consecutive annual increase - a concrete kept promise.

By Q3 2025 (6 Nov 2025), the data center thesis they had been repeating for a year was clearly materializing (30-35% data-center revenue growth guided), validating the Power Systems narrative. But the same call exposed where management had been wrong: the electrolyzer market collapse forced impairments, and Smith admitted demand "dried up faster than anything I have seen in my career." This is the one place where the multi-year Accelera story did not survive contact with reality - and to their credit, management said so directly and pivoted to cutting losses rather than defending the original plan.

At Q4 2025 (5 Feb 2026), with tariff fog clearing, management reinstated a 2026 outlook (revenue growth 3-8% initially) and reported record 2025 earnings excluding one-time items - a strong result delivered through one of the worst truck downturns in twenty years, which speaks to the ballast strategy (Power Systems, Distribution, aftermarket) working exactly as advertised.

Then at Q1 2026 (5 May 2026), they raised that fresh guidance hard - revenue growth to 8-11% from 3-8%, power generation to 15-25%, and the truck forecast up - on data center strength. They also did two honest, un-promotional things: delayed the B-series engine to January 2028 over EPA uncertainty, and took a $199 million charge to exit the low-pressure fuel cell business to Alstom.

The pattern across six calls: management is consistently accurate and conservative on the things they control (Power Systems and Distribution delivered record after record, the dividend rose on schedule, the ballast held through a brutal cycle), and they are honest about the things they got wrong or cannot control (they withdrew guidance rather than fake it during the tariff shock, they admitted the electrolyzer market collapse plainly, and they delayed product launches rather than rush them). The repeated guidance withdrawals in 2025 are a yellow flag for predictability, but they were driven by genuine external shock, not by management missing its own operational targets. This reads as a management team that does roughly what it says, errs conservative, and tells you when the world breaks its plan.

10. Shareholder Friendliness Index

Dividends. Cummins is a long-standing dividend grower with 16 consecutive years of increases. The quarterly rate progressed from $1.68 in 2023 to $1.82 in 2024 to $2.00 from mid-2025 (a 10% raise announced at the Q2 2025 call, 5 Aug 2025), putting annual dividends per share on the order of $6.72, roughly $7.00, and roughly $7.64 across 2023-2025. The dividend has grown every year through the period, including straight through the 2025 truck downturn and the tariff-driven guidance withdrawal, which signals confidence in the through-cycle cash generation rather than a payout under strain.

Buybacks and dilution. Cummins repurchases shares under a $2 billion authorization approved in 2024, following completion of a prior $2 billion program. In the most recent ~90-day window, the MoatMap data shows a single buyback filing dated 31 March 2026 with roughly $243 million of consideration; the Q1 2026 call confirmed $243 million of repurchases at an average $536.97 per share that quarter, alongside $276 million of dividends, for $519 million returned (Q1 2026 concall, 5 May 2026). For the older periods outside the 90-day window, buybacks were ongoing but modest relative to the dividend through 2024-2025 - capital return in those years leaned on dividends and debt reduction (about $969 million returned via dividends and debt paydown in 2024 per the Q4 2024 call, 4 Feb 2025). The company's stated framework is to return approximately 50% of operating cash flow to shareholders over time. Share count has been broadly flat (roughly 137-138 million shares), with buybacks largely offsetting option dilution rather than aggressively shrinking the count.

Verdict: Returns Capital - a reliable, growing dividend (16 straight years) backed by a consistent ~50%-of-operating-cash-flow return policy, with buybacks present but secondary to the dividend.

11. Insider Activities

This is a US (NYSE) listing, so the primary source is SEC Form 4 via EDGAR; the MoatMap database block (current as of 2026-06-27, US venue) is used as the spine, cross-checked against EDGAR for the most recent two weeks. The block is not marked stale. Over the trailing 12 months it records 26 transactions across 17 insiders: zero open-market buys, 11 open-market sells, and 15 "Other" (grants/equity awards/deemed corporate actions).

There are no open-market purchases by any insider in the window. That removes the single strongest signal this section can carry; there is no conviction-buying to flag.

The "Other" cluster on 12 May 2026 is the routine annual director equity grant: eleven directors (Tsien, Stone, Quintos, Nelson, Miller, Lynch, Harris, Fisher, Di Leo Allen, Belske, plus others) each received an identical 306 shares at $0 cash cost (Form 4, 2026-05-12). Identical share counts on the same date at zero price are the hallmark of a scheduled board compensation grant, not a market signal. Similar zero-price "Other" entries for officers (Merritt, Stoner, Fetch, Boakye, Lamb-Hale) are equity-award or vesting events rather than discretionary trades.

The genuine open-market activity is all selling, and it clusters tightly in the days after the Q1 2026 earnings release (5 May 2026) - the standard post-earnings trading window when executives are cleared to transact:

The largest seller by far is Jennifer Bush, the Power Systems president, who sold roughly 5,000 shares for about $3.5 million across six tranches on 12 May 2026 - notable given Power Systems is the company's standout-performing segment, but the timing (immediately post-earnings, in tranches) is consistent with a planned diversification or scheduled 10b5-1-style sale rather than a view on the business; no specific reason is disclosed in the block, so treat it as reason not disclosed. The other officer sells (Fetch, Merritt, Jackson) are modest relative to executive compensation at this level and fit the pattern of post-vesting, post-earnings-window diversification. None individually rises to a red-flag scale.

Net assessment. Insiders are net sellers, with zero open-market buying over twelve months. The selling is broad-ish (four officers) but small in aggregate and concentrated in the routine post-earnings window, while the bulk of recorded activity is non-discretionary director grants. There is no cluster buying, no first-time CEO/CFO purchase, and no unusually large discretionary sale outside normal diversification. Read: mild-neutral. The absence of any insider buying is worth noting in a year when the stock and Power Systems margins have been strong, but the selling pattern itself looks routine rather than alarming. Net signal: neutral, with a faint negative tilt from the complete lack of buying.

12. Scenarios

Bull case. The AI data center buildout proves to be a multi-year, structural demand wave rather than a spike, and Cummins's Power Systems segment rides it. The Fridley and international capacity expansions come online, the 18-month QSK95 backlog converts cleanly into shipped revenue, and Power Systems margins hold near record levels while Distribution compounds on the swelling installed base of gensets in the field. China keeps growing in both power generation and trucks without crushing margins. Meanwhile the North American truck cycle bottoms and turns up in 2026-2027, with an EPA 2027 prebuy adding a cyclical kicker just as the new HELM B-series launches in 2028 into a refreshed product cycle. Accelera's losses shrink toward immateriality after the fuel-cell divestiture, removing a drag. In this world Cummins has quietly transformed from a truck-cycle engine company into a data-center-power compounder with a cyclical call option attached, and the market re-rates the durability of its earnings.

Base case. Management delivers roughly what it has guided. Power Systems and data center demand carry the company through 2026, with power generation growing in the raised 15-25% range and Distribution setting more records on aftermarket. The North America truck market improves modestly in the second half of 2026 (toward the raised 230,000-250,000 unit forecast) but stays well short of a boom, so Engine and Components recover gradually rather than sharply. Tariffs remain a manageable, low-tens-of-millions headwind. The dividend rises for a 17th straight year, buybacks stay secondary, and Accelera losses keep narrowing. Cummins remains two businesses in one - a cyclical truck-engine maker and a structurally growing power-systems supplier - with the strong half masking the weak half, much as it did through 2025.

Bear case. The data center demand that is currently carrying record Power Systems margins normalizes or shifts toward alternative power architectures, decelerating the one segment holding the company up - just as the North American truck cycle stays in its trough longer than expected and the hoped-for EPA 2027 prebuy fails to materialize because of continued regulatory uncertainty (the same uncertainty that already pushed the B-series to 2028). Engine and Components margins stay compressed, China growth commoditizes under intensifying local competition, and a fresh tariff escalation forces another round of demand and planning disruption. Accelera, despite the loss-cutting, remains a reminder that the energy-transition bets can destroy capital. In this scenario the cyclical and the structural businesses weaken at the same time, the dual-engine resilience that defined 2025 breaks down, and Cummins is left as a truck-cycle company in a downturn with its growth story on pause.