Cummins Share Price : So The Order Book Starts Talking

Cummins shares closed 16 September almost 29% below their June high. On 15 September, two of the people responsible for explaining the company to investors walked into Morgan Stanley’s Laguna Conference and delivered a remarkably concentrated tour of everything the market might presently want to hear: improving truck conditions, 2027 emissions costs passed through to customers, higher-priced new engines, data-centre demand described as “extremely strong”, QSK95 orders stretching into the second half of 2028 and Accelera spending being cut while capital moves towards businesses Cummins says offer stronger revenue and profit growth. MarketBeat turned the conference into a financial-news story. Yahoo carried it. Nobody needs a conspiracy. The order book knows exactly when to clear its fucking throat

There is a lazy version of this story in which somebody at Cummins headquarters notices the share price, presses a large red button marked GOOD NEWS and Yahoo Finance obediently produces a reassuring headline before lunch. That is not what the evidence shows, and pretending otherwise would actually make the story less interesting.

How The Message Travels

What happened is considerably more ordinary. Cummins appeared at Morgan Stanley’s 14th Annual Laguna Conference on 15 September, represented by James Hopkins, its new Vice President of Financial Planning, Capital Management and Investor Relations, and Nick Arens, Executive Director of Investor Relations. Their audience was the market, their subject was the business and their fucking job was investor relations.

Two days later, MarketBeat published Cummins Sees Truck Demand Rebound, Data-Center Power Orders Stretching to 2028, which Yahoo Finance then syndicated to a much larger audience. MarketBeat itself says the article was generated using narrative-science technology and financial data. There is therefore no evidential basis for inventing a secret Cummins-to-Yahoo pipeline, a paid headline or some shadowy bastard carrying a Power Systems brochure into a car park.

The mechanism is simpler and, in 2026, arguably more revealing. Cummins talks in a room full of investors, the investable points enter the financial-information stream, software identifies the material most likely to matter to markets, a publisher packages it and Yahoo gives it a larger shop window. Nobody needs to push the story in the conspiratorial sense because the machinery is perfectly capable of doing the pushing itself.

The Share Price Gives The Conversation Some Context

Cummins is not staggering around with smoke coming through the bonnet. That distinction matters because this is not a story about a distressed company desperately waving its arms to distract investors from collapsing operations. Second-quarter revenue hit a record $9.5 billion, full-year revenue guidance was raised to growth of 10% to 13%, and Power Systems was particularly muscular, producing $2.3 billion of sales, up 19%, with $552 million of EBITDA at a 24.5% margin.

Data-centre demand was already helping drive that performance. Yet the market had become considerably less enthusiastic about Cummins shares. The stock reached a 52-week high of $737.76 on 18 June and closed on 16 September at $527.50, roughly 28.5% below that peak. Earlier in September, Zacks had already noted that the shares were down about 15.3% in the month following the Q2 release, despite the revenue beat and higher guidance, because adjusted earnings had missed consensus while incentive compensation, R&D, freight and product-coverage costs pressured profitability.

That gives the Morgan Stanley appearance its proper context. Record revenue and raised guidance were already in the books, while Power Systems was producing serious cash, but the share price had still given back a substantial amount of altitude. Into that setting walked two investor-relations executives carrying a truck-cycle update, a data-centre backlog and several reasons why shareholders might prefer looking through the windscreen rather than staring at the fucking rear-view mirror.

The order book had arrived to testify.

First Up: The Truck Business Is Feeling Better

James Hopkins described the North American truck market as soft over the previous year but improving over roughly the last six months. Fleets were making more money, he said, which generally supports vehicle purchases, while greater clarity around 2027 emissions requirements had removed some of the uncertainty hanging over customer decisions.

That regulatory point matters because the 2027 transition has been one of the obvious questions around the North American truck cycle. Cummins has already outlined a phased product strategy under which selected current-generation engines can remain available while customers and manufacturers navigate the tighter regulatory framework. Under the non-conforming penalty mechanism discussed at Morgan Stanley, manufacturers can continue selling products that do not meet the incoming standard provided the applicable penalty is paid.

For Cummins, the financial interpretation is wonderfully uncomplicated. Hopkins said the company expects to fully pass the NCPs through the market, meaning the selling price of the existing product rises enough to offset the penalty. Profit dollars can therefore remain broadly flat even though the higher headline selling price mechanically reduces the percentage margin.

There is no need for corporate poetry here. Regulation gets more expensive, the customer meets the extra cost at the checkout and Cummins keeps moving. The rule changes, the invoice changes with it, and nobody at the company appears particularly interested in eating the fucking bill.

And The New Engines Cost More Too

Morgan Stanley then moved to the obvious next question. Existing engines can continue under a penalty structure, but what happens to pricing on the new technology?

Hopkins’ answer was essentially that Cummins has seen this film before. The company has spent decades introducing increasingly complex engines in response to tighter emissions rules, and historically those products have carried higher prices and, over time, additional margin. Cummins expects the next emissions cycle to follow the same broad pattern.

That is useful because it strips away some of the ceremonial bullshit surrounding emissions transitions. Cleaner engines require engineering, engineering requires investment and new technology creates additional value that somebody eventually prices into the product. Hopkins also acknowledged that warranty accruals can be somewhat higher during launch, which is why Cummins intends to introduce the new engines initially at lower volumes, gather real-world data, fix whatever needs fixing and expand once the technology has accumulated field experience.

None of that is unusual industrial practice. What is more interesting for shareholders is that the 2027 transition offers Cummins two monetisable routes. Customers can keep buying familiar products and absorb the penalty through price, or move towards newer technology carrying its own additional price and margin potential.

Different road, same fucking till.

Then The Conversation Reaches The Server Room

This is where the presentation stops clearing its throat and starts singing.

Morgan Stanley’s Angel Castillo made clear that data centres could not be treated as some minor side topic because investors would quite rightly be pissed off if he skipped over them. Power Systems has become one of the more important parts of the Cummins story, and the explosion in AI infrastructure has dragged standby generation out from behind the building and straight onto Wall Street’s favourite PowerPoint.

Nick Arens described data-centre demand as “extremely strong”. Cummins has renewed a multi-year agreement with one of its large hyperscale customers, while the company is now taking QSK95 orders into the second half of 2028. That is not some airy 2030 aspiration floating beneath a stock photograph of a server rack. Those are orders being taken now.

Supply is tight enough that customers unable to secure the 95-litre product are shifting into 78-litre, 60-litre and even 50-litre alternatives. There is something beautifully industrial about that. The fashionable end of AI involves chips, tokens, models and executives saying compute until somebody confiscates the microphone; at the other end, Cummins is apparently telling customers that the enormous fucking engine is sold out and asking whether they would care to try the merely gigantic one.

That is the physical underside of the digital boom, and Cummins is standing underneath it with a catalogue.

The $9 Billion Target Says Hello

Cummins also reiterated its Analyst Day target of more than $9 billion of data-centre exposure by 2030. Precision matters here because the company did not say it already has $9 billion sitting in the order book. The $9 billion-plus figure is a forward target for exposure to the data-centre market by 2030.

What is already sitting in the order book stretches far enough out that Cummins says it is taking QSK95 orders into the second half of 2028. Arens also said the 2030 opportunity remains largely underpinned by the diesel standby story, which tells you something important about where the supposedly frictionless digital future eventually meets the physical world.

The cloud may live in cyberspace, but the bastard still needs a basement.

Cummins is not relying solely on diesel either. The company continues developing natural-gas prime-power solutions and Battery Energy Storage Systems, while broader high-horsepower capacity plans span standby generation, mining and prime-power applications. The logic is increasingly obvious: Cummins does not need every customer to choose the same technology so long as it remains somewhere inside the power architecture.

If AI requires diesel standby, Cummins can sell diesel. A development needing natural-gas prime power creates another product opportunity, while batteries offer yet another route into the same infrastructure spend. The fuel mix can move around while the electricity bill keeps getting fucking bigger.

Competition Has Also Found The Buffet

Morgan Stanley did not spend the entire session gently stroking Cummins’ order book. The obvious competitive question came up as well because other manufacturers have noticed that data centres require mountains of power and increasingly obscene amounts of capital are flowing into that infrastructure.

Arens acknowledged that the mid-to-high-single-digit pricing Cummins has achieved in recent years is likely to moderate as competition intensifies. That matters because it prevents the conference from becoming some childish EVERYTHING UP FOREVER exercise. More capacity is entering the market, competitors can see the same opportunity and eventually pricing tends to remember that customers have alternatives.

Cummins’ counterargument is that its position rests on more than simply having large engines in a catalogue. It points to long-standing hyperscaler relationships, global distribution, service capability, installed experience and a relatively limited group of manufacturers able to supply very high-horsepower diesel systems at the scale these projects require.

Other people have smelled the food.

Cummins believes it already has its arse on one of the better tables.

Then Accelera Hands Some Of The Wallet Back

Eventually, the conversation reached capital allocation and Accelera, where the contrast becomes considerably sharper.

Hopkins said Cummins has made what he described as difficult but necessary decisions to reduce Accelera’s losses, including lowering research and development spending and exiting businesses such as electrolyzers. The purpose, he said, was to redirect capital towards other parts of Cummins offering stronger revenue and profit growth while retaining core capabilities in technologies the company still expects to matter.

There is no need for TCAP to decorate that too heavily. The money went looking for returns.

At Q2, Accelera produced $145 million of sales and recorded a $69 million EBITDA loss, whereas Power Systems generated $2.3 billion of sales and $552 million of EBITDA. One side contains much of the zero-emissions future Cummins has spent years presenting to the world; the other is selling industrial power into one of the most capital-hungry infrastructure booms on earth.

You do not need a finance degree to work out which spreadsheet currently looks prettier.

Cummins is not abandoning zero-emissions technology. Hopkins specifically described retaining important capabilities while pacing investment according to the rate at which markets actually adopt them. That is a rational capital-allocation position, but it is also considerably more informative than the sustainability wallpaper. Investor conferences have a habit of turning lofty transition language back into ordinary questions about margin, adoption, capital and where the fucking money is actually being made.

This Is What Investor Relations Is Supposed To Do

That distinction deserves to be made clearly because TCAP is not accusing Cummins of improper market manipulation. Investor Relations is supposed to explain the investment case, address perceived risks, clarify uncertainty, highlight operating strengths and make sure shareholders understand whatever management believes the market may be undervaluing or misunderstanding.

Hopkins and Arens did precisely that. If the truck cycle looks uncertain, Cummins can point towards improving fleet economics and greater regulatory clarity. Concerns that 2027 emissions rules destroy profitability can be answered by explaining that penalties are expected to pass through while newer technology carries additional price and margin potential. Questions about whether the data-centre boom is already running out of road meet QSK95 orders stretching into H2 2028, while worries about competition are answered with hyperscaler relationships, distribution and installed scale.

Accelera gets the same treatment. Investors wondering whether Cummins will keep pouring money into zero-emissions technologies regardless of commercial adoption are told that spending is being resized and capital redirected towards areas with stronger revenue and profit prospects.

That is not sinister.

It is fucking efficient.

Then The Machine Writes The Headline

This is where the MarketBeat and Yahoo part becomes properly interesting.

MarketBeat explicitly says its report was produced using narrative science technology and financial data to provide rapid reporting. In other words, the modern financial-news machine consumes an investor-relations appearance and identifies the material most likely to become an investable headline.

The billboard becomes truck demand rebound and data-centre orders stretching to 2028. Yet the same conference also contains Accelera R&D cuts, an electrolyzer exit, pricing moderation as competition increases, the mechanics of passing 2027 emissions costs through to customers and other details that are considerably less photogenic. Those points are not hidden, and some appear in MarketBeat’s own story. All are available in the underlying conference record.

They simply do not get the marquee.

Recovery and backlog do.

Yahoo then syndicates the MarketBeat article, and Cummins’ appearance in a conference room becomes a clean financial-news headline sitting in front of a much larger audience. Again, nobody needs to “push” anything in the conspiratorial sense. The modern information system does that perfectly well on its own: say something sufficiently investable in the right room and software will carry the fucking bags.

TCAP understands that model rather well.

Then Generac Hands The Story Some Accidental Context

The market subsequently supplied Cummins’ remarks with a particularly neat piece of context.

On 17 September, investors were digesting Generac’s major agreement with Amazon for data-centre generators, involving about $2.4 billion of expected initial deliveries during 2027 and 2028. Generac shares jumped sharply as the market acquired another shiny way to trade the AI power-infrastructure boom.

The chronology matters. Cummins spoke at Morgan Stanley before that Generac/Amazon news reached the market, so there is no basis for claiming Cummins saw Generac’s announcement and rushed out some defensive response.

What the Generac deal does is make Cummins’ comments from two days earlier look exceptionally well timed. While investors watch another generator manufacturer celebrate a giant hyperscaler win, the Cummins story circulating through Yahoo reminds them that Cummins already has QSK95 orders running into H2 2028, has renewed a multi-year hyperscale agreement and still expects more than $9 billion of data-centre exposure by 2030.

Different company, different deal, same enormous fucking appetite for electricity.

The Share Price Slips, So The Order Book Starts Talking

That leaves the question sitting in the headline: did Cummins go to Morgan Stanley because its share price had fallen?

TCAP cannot establish that, and there is no reason to force the chronology into an emergency-response narrative. Investor conferences are scheduled events, and management teams attend them whether their shares are soaring, sinking or wandering sideways looking bored.

What can be established is more useful. Cummins arrived at Morgan Stanley with its shares materially below their June peak, after a period in which record revenue and raised guidance had failed to prevent a substantial pullback. In that setting, two investor-relations executives delivered a highly concentrated answer to the questions hanging over the investment case, covering truck demand, pricing, margins, 2027, data centres, 2028 and capital discipline while reiterating the $9 billion target and explaining why Accelera spending was being trimmed.

An automated financial-news system subsequently distilled the discussion into exactly the sort of headline likely to make a shareholder stop scrolling.

That is not proof of orchestration.

It is proof that Investor Relations understands its fucking job.

Follow The Money, Not The Wallpaper

Cummins spends considerable energy explaining what it wants the future of power to become. Investor conferences tell you something slightly different because they are less interested in the wallpaper and considerably more interested in what presently pays.

The truck business remains enormous. Existing powertrains can potentially remain saleable under the expected 2027 framework, with Cummins intending to pass penalties through to customers, while new emissions technology should command additional price and eventually improved margin. Data-centre diesel demand is booked years into the future, natural-gas prime power creates another route into the same infrastructure problem and battery storage offers another. Accelera investment, meanwhile, is being resized according to commercial adoption while capital moves towards businesses producing stronger revenue and profit.

That is a considerably more useful picture of Cummins than a brand campaign. The company is not betting the house on one future; it is positioning itself to get paid through several of them. Right now, with the shares well below their June high, Cummins also has one particularly persuasive bastard sitting quietly in the accounts department.

The diesel order book.

Turns out it talks beautifully.

Lee Thompson – Founder, The Cummins Accountability Project