Sunbelt Rentals

IThe business

Sunbelt rents equipment. A builder who needs an aerial lift for three weeks, a hospital that needs backup power for a month, or a film crew that needs lighting for a shoot pays a day, week or month rate, and Sunbelt delivers, services and collects the machine. It has 1,611 stores: 814 in General Tool and 614 in Specialty across North America, and 183 in the UK.² In fiscal 2026 rentals were 93% of revenue, sales of used rental equipment 4% and sales of new equipment, parts and consumables 3%.² About 800,000 US customers rent from it, the ten largest are under 10% of revenue, and none is above 1%.²

This page is about a company that changed its name, country and accounting in 2026 while the business stayed the same. The next section explains it plainly, because it decides which numbers you can compare.

Ashtead, Sunbelt, and which accounting is which

For nearly all of the period on this page the company was Ashtead Group plc, a UK company listed in London that owned Sunbelt Rentals. It reported under IFRS, the international standards, in dollars from its 2022 report and in pounds before that.⁸⁹ Its financial year ended on 30 April. On 27 February 2026 a court-approved scheme made a new Delaware company, Sunbelt Rentals Holdings, Inc., the parent of Ashtead. Each Ashtead shareholder received one new share for each old one. Trading began on 2 March 2026 on the New York Stock Exchange as the primary listing, with London as a secondary listing, under the same ticker, SUNB.² Ashtead became a private subsidiary.²

Three things follow. First, the year end did not change. The fiscal year still ends on 30 April, and “fiscal 2026” is the year to 30 April 2026. The next one, fiscal 2027, ends on 30 April 2027.²³ Second, the accounting did change. The company now reports under US GAAP, and its Form 10 recast Ashtead’s results for fiscal 2023 to 2025 on that basis, and its first annual report carries fiscal 2024 to 2026, audited by PwC.⁶² Third, the earlier years are still IFRS and nobody has recast them: the ten-year table in the 2025 Ashtead report is the last place they appear.¹⁰ I show both, and I label each row. Fiscal 2017 to 2022 are IFRS. Fiscal 2023 to 2025 are shown both ways. Fiscal 2026 and the first quarter of fiscal 2027 are US GAAP only.

The gap between the two is not small on some lines. Revenue is the same on both bases. Adjusted EBITDA is not: for fiscal 2025 it was $5.02B under IFRS and $4.75B under US GAAP.¹⁰⁴ The likely reason is rent. IFRS moves the cost of leased property out of operating costs into depreciation and interest, which lifts EBITDA. US GAAP keeps most of it inside operating costs. The size matches: the cash-flow statement shows $288M of operating lease expense for fiscal 2025, against a gap of $270M C (calculated).⁶¹⁰⁴ I did not find a line-by-line bridge in the filings. The cash flows differ too. Ashtead’s IFRS “cash generated from operations” is struck after paying for rental fleet and counting used sales, while the US GAAP operating cash flow of $3,844M for fiscal 2025 is before fleet spending, so the two must not be set side by side.¹⁰²

For the rest of this page I use US GAAP figures whenever they exist, and say so when I must lean on IFRS ones. One consequence is that the long record of returns I quote, Ashtead’s return on investment of 14.5% to 19.2% over nine years, is an IFRS number, and the recast figures for fiscal 2023 to 2025 are 19%, 17% and 15% on the company’s US GAAP basis.¹⁰⁶ That is close enough for me to treat them as one series.

Three segments, with very different returns

North America General Tool rents the common equipment on a job site: aerial work platforms (39% of that fleet at cost), forklifts (25%), earthmoving machines (17%), lighting and small tools.² Specialty rents products that need expertise, in power and HVAC (43% of that fleet), scaffold, pumps, film and TV equipment, flooring, trench safety, temporary fencing and structures. The UK segment does both.²

I find this table the most useful on the page. Specialty earns about 24 cents of segment profit a year on each dollar of fleet at cost, General Tool about 15 and the UK about 4 C (calculated). The UK is the weak part, and a restructuring in December 2025 and a goodwill test that passed with little room are the company’s own acknowledgement.²⁴ General Tool is the base: 58% of revenue, and the segment where profit fell 8% in fiscal 2026 on 2% more rental revenue.⁴ Specialty’s profit rose 3% on 6% more rental revenue, and then in the first quarter of fiscal 2027 its rental revenue rose 25%.⁴⁵ Part of that was acquisition and part was a one-off event, and I take it apart in the section on rate.

The rental-versus-ownership cycle, and what fleet spending does to cash

The case for rental is simple, and the company argues it in its annual report. Rental makes up around 55% to 60% of the North American market for this equipment, and it sees room for the share to pass 60% over time. The reasons it gives are replacement-cost inflation, changing technical requirements, and health, safety and environmental rules, all of which make renting cheaper and easier than owning.² I can’t test that claim, because the rental share is a management estimate.

The cost of the model is fleet. A rental company buys machines, rents them out, then sells them. Depreciation of rental equipment was $1.85B in fiscal 2026, and most fleet types are depreciated over 5 to 10 years to a salvage value of 10% to 15% of cost.² Spending on fleet is not like spending on a factory. It is a dial the managers can turn. In good years they buy more, to serve demand and replace old units. In slow years they cut hard, the fleet ages, and cash piles up. The 10-K says exactly that: traditionally these companies produce cash only in downturns, and Sunbelt believes it has changed that.²

Look at the record. In the pandemic year, fiscal 2021, Ashtead spent $0.82B on rental equipment and its free cash flow on its own definition was $1.82B.⁹¹⁰ Two years later, with demand strong, it spent $3.02B in fiscal 2023 and $3.76B in 2024, and free cash flow fell to $0.44B and $0.13B.⁶ In fiscal 2025 it spent $2.25B, and in fiscal 2026 $1.84B, and free cash flow came back to $1.68B and $2.06B.² Nothing about the business improved between 2024 and 2026 in the way those cash figures suggest. It bought less. The fleet shows it: the average age of the fleet rose from 49 months in April 2025 to 53 months in April 2026.⁴

The chart shows the same thing as a ratio. When the fleet is new, its net book value is a high share of its cost. In April 2024 it was 65%. By April 2026 it was 58%, about where it stood in fiscal 2021 and 2022 C (calculated).²⁶⁹ Now the company has turned the dial again. Its raised outlook for fiscal 2027 is $2.75B to $3.15B of gross rental capex and $2.4B to $2.8B net of used sales, against $1.84B and $1.42B last year.⁵ In the first quarter alone it spent $759M, nearly double the $394M of a year earlier, and free cash flow was $70M.⁵³ So fiscal 2026 was a year of spending below the need, and fiscal 2027 is a year of catching up and growth. Neither year shows what a normal year looks like. That is the problem I try to solve under owner cash in Financials.

Utilisation and rate: what drives the revenue

A rental business earns revenue from three things: how many machines it owns, how often they are on rent, and what each rental pays. Sunbelt publishes a ratio that mixes the last two, called dollar utilization: a year of rental revenue divided by the average cost of the fleet.⁴ It was 60% in fiscal 2023, 56% in 2024, 54% in 2025 and 55% in 2026.⁶⁴ General Tool fell from 53% to 47%. Specialty fell from 84% to 74% and has since climbed back to 75%, and 77% in the latest quarter.⁶⁴⁵

Read this carefully. A falling ratio does not mean machines sit idle. Fleet is counted at cost, and the cost of new machines has risen, so every added dollar of fleet earns a lower ratio even at the same rental price. United Rentals publishes the effect: it assumes inflation takes 1.5 points a year off the revenue productivity of its fleet.¹¹ The management of Sunbelt says General Tool time utilization is among its best two or three years, and that supply and demand are tight.¹² But Sunbelt does not publish a rental rate. Its chief financial officer said on the call that the company does not comment on rate, and its documents say only that rates were “stable” in the fourth quarter and that rate improvement offset part of higher fuel costs in the first.⁴⁵¹² So here is something I can’t know: how much of any year’s growth came from price, how much from time on rent, and how much from mix. The sensible reading is that volume and mix did most of the work, and price kept up with cost inflation.

The first quarter of fiscal 2027 is the strongest in years, and needs to be read with its helpers taken off. Rental revenue rose 12.5%. The acquisition of Reliant Asset Management, sold under the Aries brand, added about one point.⁵ The FIFA World Cup added about 250 basis points, by the company’s estimate.⁵ On the call, management said about 75% to 80% of that went to Specialty and 20% to 25% to General Tool.¹² I turn that into dollars: 2.5% of $2.60B is about $65M of World Cup revenue, so about 6 points of Specialty’s 25% growth, and 1 point of General Tool’s 7.4% C (calculated). Add the 3 points from Reliant in Specialty, and the underlying growth was about 16% in Specialty and 6% in General Tool E (estimate).⁵³¹² Both are good. They are not the headline figures.

Selling used equipment

When a machine is done, Sunbelt sells it. This is where the books test depreciation. If the company sells for more than the book value, it depreciated too fast, and if less, too slowly. The gain on used sales was only $25M in fiscal 2021, in the pandemic. It reached $185M in 2023 and $223M in 2024 when equipment was scarce, then fell to $81M in 2025 and $65M in 2026.⁹⁶² Used sales are 4% of revenue.²

In 2026 the gain was 14% of sales, against 26% in 2024 and 30% in 2023. That cannot be read as sloppy depreciation, because the high years were unusual: the company deferred some sales from fiscal 2023 into 2024, and 2024 volume was large.⁶ But it is a warning against assuming the used market always pays well. Sunbelt’s first-quarter used sales were $85M with a $15M gain, a margin of 18% C (calculated).³ Management said recovery rates were rising and expects margins to improve as a new retail channel grows.¹² United Rentals reports a gross margin on used sales of 44.8% in its third quarter and recovers 51.4% of original cost on what it sold.¹¹ I won’t compare the two. They depreciate on their own policies, and a high used-sale margin can mean a company writes assets down faster. What I can say is that if Sunbelt’s depreciation were one year too long, the 10-K puts the cost at $251M a year of extra expense, about 10% of operating income.² That is the size of the question.

IIThe moat

The source of any protection here is local scale. A rental store needs trucks, mechanics and a yard, and it earns its return when many machines rent from the same cluster of stores. The 10-K says that as clusters mature they reach non-construction customers and get better margins and returns.² The three largest companies in North America hold about 31% of the market together and Sunbelt holds about 11%. The rest is small independents, and more than 40% of US revenue sits with companies that have five or fewer stores.² Those are company estimates.

Evidence it holds

The first sign is the customer base and its stability. Sunbelt added about 126,000 new credit account customers in the first two years of its current plan, after about 118,000 in the previous plan.² The second is Specialty: dollar utilization of 75% to 77% is far above anything in General Tool, and it did not fall in the weaker local market of fiscal 2026.⁴⁵ The third is how small an equipment seller’s hold on it is. No customer is above 1%.²

Evidence it is thin

Returns tell the real story. The company’s adjusted return on investment, which is adjusted operating profit over average net assets before tax and debt, was 19% in fiscal 2023 and 14.2% in the fourth quarter of 2026.⁶⁴ At a tax rate near 25% that is about 11% after tax, close to the 10% I ask for. A company with a strong edge shows high returns that hold through the cycle. This one earned 14.5% to 19.2% a year over nine years of IFRS reporting, averaging 17%, and today is at the bottom of that range C (calculated).¹⁰ The fall came from buying a lot of fleet at higher prices, while rental revenue for each dollar of fleet slipped.

The prices are also not set by Sunbelt. The company says this plainly: rates in General Tool were “approximately flat” in the fourth quarter, and it does not say if rates rose last year.⁴ Its biggest rival reports that fleet productivity rose 6.5% in its latest quarter on better rates, utilization and mix.¹¹ So there is a competitor showing a lift and Sunbelt not showing one. That is not a decisive fact, because the two measure differently. It is the reason I call the protection modest.

IIIThe competitive advantage

The honest picture is a company with a real advantage in one place, a narrower one in another, and none in a third. Here are Sunbelt, the larger United Rentals and the smaller Herc side by side, as each reported. The three do not define their measures in the same way, so I show them as printed.

Where Sunbelt wins

It wins on Specialty. Power and HVAC gear is 43% of that fleet, and the fourth-quarter growth was led by load banks, which test backup power for data centers.²⁴ Fleet sits in a niche with high utilization and enough know-how that small shops can’t copy it quickly. Specialty profit per dollar of fleet is about 24%, more than General Tool’s 15% by 9 or 10 points C (calculated).

It also wins on a cost of debt. In July it sold $750M of notes due 2036 at 5.650%.¹⁴ On 29 October, United Rentals sold notes due 2036 at 6.200%.¹⁵ The securities differ, and the dates are three months apart, so I don’t read a lot into 55 basis points. It does say lenders are comfortable.

Where it loses

It loses on size and on margin. United Rentals is about 1.6 times larger by quarterly revenue, and made an adjusted EBITDA margin of 47.6% against Sunbelt’s 42.2%.¹¹⁵ The company explains its own margin slide with the mix: more ancillary revenue, such as delivery and fuel, and more Specialty revenue, with higher repair and repositioning costs.⁴⁵ That is true, and it also leaves a margin that fell from 44.0% to 41.9% in a year when revenue grew 3%.⁴

It loses on growth in the biggest cycle. The mega-project pipeline is the story of this cycle. The 10-K cites Dodge Construction Network: mega projects, those over $400M, are expected to total more than $1.5 trillion of starts in 2026 to 2028, against about $765B in 2023 to 2025, led by data centers.² On the call, management said data centers are about 13% of the mega-project universe it tracks and semiconductors about 3%.¹² United Rentals grew rental revenue 17.8% and Herc 12% in the latest quarter, both with that demand.¹¹¹³ Sunbelt grew 12.5% with the help of the World Cup. That is the comparison I would watch.

And it loses in the UK, which is 9% of revenue and earns about 4% on its fleet, in a market the company calls subdued.²

IVManagement

Sunbelt is run by people who came up inside it. Brendan Horgan, 52, joined nearly thirty years ago, became chief operating officer of North America in 2003 and chief executive of Ashtead in May 2019.⁷ In the seven years from fiscal 2019 to 2026, revenue went from $5.87B to $11.15B, up 90%, adjusted EPS from $2.27 to $3.72, up 64%, and the company’s return on investment from 17.8% to 14.2% C (calculated).¹⁰⁴ Revenue is on the same footing on both accounting bases, and the earnings figures are IFRS at the start and US GAAP at the end, so read the second pair as rough. It is a record of growth bought at falling returns on the capital it used.

The finance chief is fairly new. Alex Pease joined Ashtead as finance chief designate in October 2024, took over in March 2025, and is now chief financial officer of Sunbelt.⁷ The general counsel gave notice in October, stepped down from the role at once, and stays on until November 2027 under a retirement agreement; the company is searching for a successor.¹⁷ Neither fact worries me alone. A finance chief in the role since March 2025 and a general counsel now stepping down mean the control team is in transition just after a change of country, exchange and accounting.

How they are paid, and what that pays them to do

Horgan’s pay for fiscal 2026 was $14.63M, up from $12.13M. Of that, $1.21M was salary, $11.08M was stock awards at accounting value and $2.22M was cash bonus.⁷ The largest share is stock, and 82% of his annual stock grant is in performance units, with maximum value of 850% of salary.⁷ The performance units pay on four measures: relative total shareholder return, compound growth in adjusted EPS, and adjusted return on investment, each 30%, and carbon intensity, 10%.⁷ Those are the right kinds of measure. Return on investment is the one that matters most here, and it is the one that has been falling.

Two features of the pay I do not like. First, the test was waived at the relisting. When the company assumed the Ashtead stock awards in the move, it dropped the performance conditions on outstanding performance units, subject to each holder’s consent, and deemed them met at 85.5% of maximum, a figure it says matches the five-year average. It also dropped the holding periods and the board’s discretionary underpin on restricted units.⁷ The board’s reasons are in the proxy: US practice, the cost of re-testing under new accounting. They are fair reasons. But it means a large part of the pay already granted no longer depends on the next three years’ returns.

Second, the bonus rewards cash after fleet spending in the same year. Half of the annual cash bonus is adjusted pre-tax profit and half is free cash flow before interest, tax and “capital allocation decisions”.⁷ In fiscal 2026 the profit half paid 68% of maximum, on $2,017M against a target of $1,890M, and the cash half paid 94.7%, on $3,105M against a target of $2,972M.⁷ Rental capex had just fallen by $0.4B.² I am not saying anyone held back spending to earn a bonus. The targets were set before the year and the result beat them. But a measure that pays more when the fleet is bought less will, over time, favour a fleet that gets older, and I would want the board to say how it guards against that.

How much of their own money is in

More than at most companies. Horgan holds 443,379 shares, about $35.8M at the close, or about 30 times his salary, against a guideline of six times.⁷¹ All 14 current officers and directors together hold 652,219 shares, 0.16% of the company and about $53M C (calculated).⁷ That is real, but it is small next to the outside holders. Dodge & Cox reported 53.1M shares, 12.8% of the company, for 31 March and has filed no amendment since.¹⁸ Vanguard Capital Management reported 22.2M shares, 5.4%, for 30 September, down from 38.4M, 9.4%, at 30 June, and BlackRock 21.1M, 5.2%, down from 36.7M, 9.0%.¹⁹²⁰ Two index managers cutting by about 40% in one quarter is large, and their filings do not say why.

What they have done lately is quiet. The Forms 4 since March show awards, shares withheld for tax, one sale and one tiny purchase.²¹ John Washburn sold 6,702 shares in July at $70.00 and $72.50, about $0.48M C (calculated). Kyle Horgan bought 25 shares at $73.30 on 1 October, about $1,900. Nobody bought in size, and nobody sold much. Closes since the listing have ranged from $63.09 to $86.06.¹ I read the insider record as neutral.

One more thing, from the proxy’s related-person section: Kyle Horgan, who runs Specialty, is the chief executive’s brother. His pay for fiscal 2026 was about $1.08M in cash and $0.49M of vested stock.⁷ He has been with the company since 1998 and runs the segment with the best returns. It is disclosed, and the board knows. I mention it because the segment that carries the growth case is run by the chief executive’s brother.

What they did with capital over a full cycle

The record from fiscal 2021 to the first quarter of fiscal 2027, six and a quarter years, is below. In that time Sunbelt spent $14.2B on rental equipment, $4.4B on acquisitions, $2.6B on dividends and $2.6B on buybacks.⁹⁶²³

Three things stand out. They bought fleet hard into the boom: $8.5B of rental equipment in the three years to April 2024, and the fleet at cost went from $11.9B in April 2021 to $17.7B in April 2024.⁹⁶¹⁰ They spent about $1B a year on acquisitions from fiscal 2022 to 2024, then paused for two years. And they bought back stock only a little until the last two years, then $1.41B in fiscal 2026, 4.2% of today’s market value C (calculated).² After the listing the company bought 3.69M shares in March and April at an average of $69.35, about 14% below today’s close C (calculated).² The shares outstanding fell from 430.7M to 410.3M, down 4.7%, in fiscal 2026.² That was a good use of money if the shares are worth more than $70, and a poor one if they are worth less. On my base they are worth less, so by my numbers the buyback paid more than the shares were worth. If the market is right, it was cheap.

The acquisitions

Sunbelt describes its acquisitions as “bolt-on”, meaning small businesses that add stores and fleet in places it already serves. The long record fits: 13 deals in fiscal 2026 for $224M, five deals for $141M in fiscal 2025.² The newest is different. On 1 May it bought Reliant Asset Management, a modular space rental company that trades as Aries, and one other business, for $669M in total.⁴³ The accounts value its fleet at $265M, intangibles at $80M and goodwill at $319M.³ In other words, for each dollar of fleet it paid about $2.50, and about 48 cents of every dollar went to goodwill.

The same ratio shows up in fiscal 2026: $224M paid, of which $89M was rental equipment and $122M goodwill.² A bolt-on deal has cost about two and a half times what its fleet is worth. A new store costs only the fleet and the lease, though it takes years to fill. The company has opened 99 of a planned 300 to 400 new stores, and I would like to see the return on those compared with the return on the bought ones.² It does not publish either. Reliant added about one point to first-quarter rental growth, which is roughly $100M of rental revenue a year on my arithmetic, so the price is about six times revenue E (estimate).⁵ The company says the deal will add to earnings per share in its first year and keep leverage inside its target range.⁴ I accept that and still think it is a high price. Management called its deal pipeline “very strong” on the call, so there will be more.¹²

Governance, briefly

Three points. The company disclosed a material weakness, in how it classified debt between current and long term, in the Form 10 and its January quarter. It reported the weakness fixed by April 2026.² The first annual report did not include the usual audit of internal controls because new registrants are given a year.² And the auditor, PwC in Charlotte, flagged one critical audit matter: the tax-free treatment of the reorganization.² None of these is a red flag. They are the loose ends of a company that changed accounting, country and auditor office in under a year.

VFinancials

Nine years of Ashtead, then Sunbelt, show a business that doubled in size and earned a lower return as it did. The table keeps IFRS and US GAAP on separate rows. Where a row has only one basis in a column, the other is blank.

Revenue has compounded at 11.7% a year since fiscal 2017 C (calculated). Adjusted operating profit has compounded more slowly, because the margin fell: from 28.2% in fiscal 2017 to 22.4% in 2026, six points, though 2017 to 2019 were before IFRS 16, so the start is not like for like.¹⁰⁴ The best comparison is within one basis. On US GAAP, adjusted operating margin went from 26.9% in fiscal 2023 to 24.2% in 2025 and 22.4% in 2026.⁶⁴ Revenue grew in two of those three years and slipped 0.6% in fiscal 2025. The margin fell in each.

What an owner could really take out

The company reports free cash flow of $2,055M for fiscal 2026. It defines that as operating cash flow less net spending on rental and other equipment.⁴ My number is lower, and the difference is the most important thing on this page. I start with the same operating cash flow, $3,784M, which already pays interest, tax and rent. Then I take off what an owner has to pay to keep what he owns.

Three deductions. Stock pay: operating cash flow adds back $80M of stock compensation, but staff were paid in shares that cost owners money.² Non-rental equipment: trucks, buildings and software, $352M spent less $41M from sales, or $311M.² Fleet to stand still: the hard one. Depreciation is the cost of machines wearing out, at what they cost years ago. Replacing them costs more. I take depreciation, $1,851M, add 3% of the fleet’s net book value of $11.2B to cover the rise in what new machines cost, $337M, and take off the $65M gain on used sales. That is $2,123M E (estimate).² The company’s own sensitivity says one year’s change in useful life is worth $202M to $251M of depreciation, so I am working in a range, not a point.²

Fleet spending in fiscal 2026 was $1,842M gross and $1,418M net of used sales. That is $705M less than my stand-still figure, which is why the reported free cash flow is higher than mine.² Fleet age rose from 49 to 53 months in the same year.⁴

So owner cash for fiscal 2026 is about $1.3B in the middle, or $3.08 a share on today’s 412M diluted shares E (estimate). Net income was $1.33B, $3.15 a share. Adjusted net income, using the company’s $3.72 of adjusted EPS, was about $1.56B. My range is $1.09B at the low end, where I add 10% to depreciation, to $1.97B at the high end, where last year’s level of fleet spending is enough. In the middle, owner cash is close to reported net income. The gap with the company’s free cash flow is $0.78B, or 38%, and with its adjusted earnings about $0.29B C (calculated). I have not added back the $134M of restructuring and relisting costs in the year, which were mostly one-off.² Doing so adds roughly $0.1B after tax.

Quality of earnings

Five things deserve a footnote. First, fiscal 2026 stock pay was $65M of expense or $80M in the cash flow statement, against income of $9M in fiscal 2025. Before the move the stock awards were accounted for as liabilities that follow the share price, which fits that swing.²⁷ That swing explains part of the 15% fall in net income. Second, fiscal 2025 included a $28M release of a provision for a customer that went bankrupt, and the fourth-quarter margin of fiscal 2026 fell 290 basis points, not 400, once that is removed.⁴ Third, deferred tax: the income statement carried deferred tax expense of $281M, $236M, $42M and $93M in fiscal 2023 to 2026, which means cash tax ran below book tax. A US law of 4 July 2025 permanently restored 100% bonus depreciation, which lets a company deduct the cost of new fleet in the first year.² That helps cash while the fleet grows, and it reverses when growth stops. I use a cash tax rate of 23% against a book rate of 25% to 26%, so I give some of it credit. Fourth, depreciation: see the used-sale gains above. Fifth, the UK reporting unit passed its goodwill test with the company saying the margin was not large, on $3.5B of group goodwill.²

The balance sheet

Debt at 31 July was $8,556M and cash $32M, so net debt was $8,524M, or 1.8 times adjusted EBITDA, against a stated range of 1 to 2 times.³ It was 1.6 times at 30 April.⁴ Of the debt, $7,343M is senior notes and $1,213M is a secured revolving credit facility of $4,750M that runs to November 2029, priced at the benchmark rate plus 1.25 to 1.375 points. At 30 April 81% was fixed rate and the weighted cost of committed facilities was about 5%.³²⁵ A single covenant, a fixed-charge ratio, applies only if unused availability falls below $475M; it was $3,750M.³ In July the company sold $1.2B of notes at 4.950% due 2030 and 5.650% due 2036, and in August it repaid $550M of 1.500% notes.¹⁴³ That swaps cheap old debt for dearer new debt; at the difference of about 3.5 points on $550M it adds roughly $19M a year of interest E (estimate).

Operating leases are not in net debt. They are $2.86B of liabilities at April, mostly property, and the rent is already in operating costs, so adjusted EBITDA is after it.² I do not count them as debt, and I do not add back the rent. Under IFRS this was handled the other way, which is another reason to keep the two bases apart.

How safe is 1.8 times? Cash flow rises in a downturn because fleet spending is cut, as fiscal 2021 showed. In my bear case, EBITDA drops about 8% from its fiscal 2027 level by fiscal 2030 and net debt falls. If EBITDA dropped 25% with no cut in spending, leverage would be near 2.4 times E (estimate). That is a risk to dividends and buybacks, not to solvency.

VIValuation

First, how this goes badly wrong

Before the numbers, I asked what would most likely ruin the business or the investment, and worked back. Three things could, and none needs a surprise.

• The cycle turns after the fleet is bought. Sunbelt’s fleet at cost went from $13.5B to $20.1B in four years.⁹³ In the three years to April 2026 rental revenue for each dollar of fleet went from 60 cents to 55. Each point of it is about $0.2B of rental revenue on a $20B fleet, and almost all of that is profit C (calculated). If demand for non-residential and mega-project building pauses, revenue for each dollar can fall faster than fleet can be sold, and fixed costs and debt stay. Bear is built on that.
• The cash is better than the business. Fiscal 2026’s free cash flow of $2.1B was produced by spending $0.7B less than it takes to stand still, by my estimate, and a fleet that aged four months.²⁴ That cash is real, and it will be spent. The raised guide for fiscal 2027 already does: $2.75B to $3.15B of gross rental capex. Anyone who values the business on the cash of the last two years is valuing a fleet that is wearing out.
• Paying a high price, with debt. Net leverage rose from 1.6 to 1.8 times in a quarter, mostly on a $669M purchase at about two and a half times its fleet.³ A bigger deal on the same terms, struck at the top of a cycle, with fleet that falls in value and a 1-to-2-times target to defend, is how a good company becomes a stressed one.

All three are versions of one question: does Sunbelt earn its cost of capital through a whole cycle, or only in the part where demand is strong? They set my bear. I keep a fourth outside the model because I can’t size it: that the mega-project boom, led by data centers, stops suddenly rather than fading. Management says some localities have placed moratoriums on data centers, and that many projects started before they took effect.¹²

What the price implies

At $80.73 the market values Sunbelt at $33.3B on 412.4M diluted shares, the 409.8M outstanding on 4 September plus 2.6M of stock units to vest C (calculated).¹³⁷ Add $8.5B of net debt and the enterprise is $41.8B, which is 8.7 times trailing adjusted EBITDA and 8.3 times the $5.02B middle of the raised fiscal 2027 range C (calculated).³⁵ The price is 20.6 times trailing adjusted earnings of $3.92, and 24.1 times GAAP earnings of $3.35 C (calculated).⁴⁵ The dividend of $0.30 a quarter is 1.5% of the price.⁵ The owner cash yield for fiscal 2026 is 3.8% E (estimate), against the 4.95% ten-year Treasury.²²

What does it pay for? Not for ordinary growth. With margin flat at 42%, the price needs the fleet and revenue to grow 17% a year for nine years. Or, on my base growth, it needs a fiscal 2036 EBITDA margin of 56%, where the best year on record is 44.0%. Or dollar utilization of 76% where the best year in this record is 60%. The reason is a fact about this business. Fleet that earns about 11% after tax creates little value when you charge it 10%, so growth is worth little at this rate and the price must come from the cash the existing fleet earns. The simpler way to say what the market is doing: the price equals my base cash flows at a discount rate near 7.4%.

What I can’t judge

Two inputs decide this valuation and I can’t know either. One is how much fleet it takes to stand still. Depreciation says one thing, the aging fleet says another, and the company does not disclose how much of its spending is replacement and how much is growth. On the call management drew the line between the two but did not size it.¹² The other is rate. Sunbelt does not publish rental rates, so I can’t tell how much of any year’s revenue came from price. Here is the range, not a point. Read my $42 with the grid beside it.

The discount rate

I use 10.0%. The ten-year Treasury yields 4.95% and Damodaran’s implied equity risk premium was 4.41% on 1 November.²²²³ A beta of about 1.15 gives 10.0%. I do not have a beta I trust. Sunbelt has traded in New York for eight months, the earlier US history in my data is thin over-the-counter trading, and the London history is in pounds. So I take 10.0% by judgement, the same rate I use on my GE Vernova page for another business whose results follow a capital-spending cycle, and I do not lower it for the 5% to 6% the company pays on its debt, because the equity carries the cycle E (estimate). At 8.5%, my base value is $59; at 9.0%, $53.

Three cases

The model is explicit about the fleet. I set the growth in fleet at cost, the revenue for each dollar of it, and the EBITDA margin for each year from fiscal 2027 to 2036. Depreciation is 9.8% of average fleet. The fleet’s net book value, as a share of cost, sets the cash: net rental capex is depreciation plus the change in net book value, less the gain on used sales and the fleet bought in acquisitions. That identity is how the accounts work, so cash cannot get ahead of it. Stock pay of 0.7% of revenue stays a cost. Non-rental capex is 2.8% of revenue. Working capital takes 10% of added revenue. Cash tax is 23% of adjusted operating profit after stock pay. I count nothing from future acquisitions, buybacks or dividends. What is left is owner cash before interest, and I take off net debt of $8.5B.

My base fiscal 2027 sits on the company’s raised guide: revenue of $12.1B against growth of 6% to 9%, EBITDA of $5.05B against $4.92B to $5.12B, and gross rental capex of $3.12B against $2.75B to $3.15B E (estimate).⁵ So the near year is not where I differ from management. The disagreement is over what a mature rental fleet earns and what it costs to keep.

The price sits 92% above my base and 20% above my bull. That is the plain result. But why might the market be right? Because it can see what I can, and weights it differently. Sunbelt raised guidance within three months of giving it, Specialty grew 25%, and United Rentals and Herc both raised theirs.⁵¹¹¹³ Mega projects are a real change in the kind of work rental companies do: bigger and longer, and in the 10-K’s account they offset a weaker local market.² If a business like this is valued at 8.0% or 8.5%, near the 7.4% the price implies on my base cash flows, then my base is worth $68 or $59, and my bull $105 at 8.0% E (estimate). If the cycle is long enough, the market is right. My bull carries that, and it returns 9.1% at this price, still short of my 10%.

Other uses of the same dollar

The ten-year Treasury pays 4.95%.²² Against that, my base here returns 7.4% a year and my bear 6.0% E (estimate). The closest comparison on this site by type is GE Vernova, a maker whose orders follow capital spending and whose page requires the same 10.0%. There the base cash flows return 5.9% at $857.87, a shortfall of 4.1 points. Sunbelt’s is 2.6 points. Costco returns 6.2% against 8.5% required, a 2.3-point shortfall, and Visa pays 8.2% against 9.0%. Those pages use their own rates and methods, so I treat gaps of a point as noise. By that test Sunbelt is priced more richly than Visa and a little more richly than Costco, and less richly than GE Vernova. Nothing here is cheap by my numbers.

What the last real pause looked like

My bear is harsher than anything in Ashtead’s record that I opened. In the pandemic year, fiscal 2021, IFRS revenue still grew 3.7% to $6.64B, adjusted operating profit fell 3% and the return on investment was 14.9%.¹⁰ Capital spending fell by half, from $1.88B to $0.95B.¹⁰ I did not open filings from earlier downturns, so I can’t check a deeper one against the record. The bear is a shock arriving after a large fleet purchase. I mark it as an estimate with no precedent in what I opened.

The margin of safety

• Price$80.73Close, Q4 2026 [1]
• Base value$42My estimate E (estimate)
• Margin of safety below$32set from my bear case C (calculated)

I set the margin of safety from the bear case: at $32, my $25 bear value is at least three quarters of the price paid, and the price is 24% under my base value. Why the bear? Because the danger in this business is the cycle turning against a big fleet, and a discount to base wouldn’t protect against that. At $32 my bear returns 9.0% and my base 11.3% a year E (estimate). Closes since the New York listing have ranged from $63.09 to $86.06, so the stock has not been near $32 since it moved.¹ At $80.73 the price is 152% above that level. There is no margin of safety on my numbers. The stress case is a stronger test: revenue falling from $11.6B to $9.6B and margin to 35%, which gives $13.

Two different questions

How good is the business, and how good is the price? They are separate questions. As a business Sunbelt is solid: the second-largest in its market, a record of growth through a pandemic, a segment (Specialty) earning about 24% on its fleet, leverage inside its own range and a stable list of customers. Its flaws are these: a return on investment that fell from 19% to 14% in three years, a General Tool business whose revenue for each dollar of fleet keeps falling, and a free cash flow figure that is flattered by a year of light spending. As for the price, at $80.73 it is above my bull case, so it asks you to believe in a long cycle and a lower discount rate than I use. This looks like a good business at a full price. That is a different thing from a good business at a bad one, and I do not rate it. It does say that the next five years have to go well for the price to be earned.

VIIThe case for the price

What most people miss

Three things cut in Sunbelt’s favour and get little attention. First, Specialty earns about 24% a year on its fleet at cost, and without the World Cup and the Reliant deal its first-quarter rental revenue still grew about 16% E (estimate).²³¹² Second, net leverage was 1.6 to 1.7 times at each April year end from 2023 to 2026, while the fleet at cost grew from $15.8B to $19.2B.¹⁰⁶⁴ Third, the 10-K cites Dodge Construction Network: mega-project starts, over $400M each, are expected to top $1.5 trillion in 2026 to 2028, against about $765B in 2023 to 2025.²

The bull case

In plain words: renting is taking share from owning, large projects and non-construction work are growing, and Sunbelt has the stores, Specialty lines and balance sheet to take more than its share. The clearest statement is the company’s own. In September the chief executive pointed to momentum in revenue and profit and to a free cash flow platform that works through the cycle, and the finance chief said the fleet spending is tied to committed customer demand.⁵¹² The 10-K adds 300 to 400 planned new stores, 99 opened, and S&P Global Market Intelligence’s forecast of 4% a year US rental growth to 2028.²

Dodge & Cox reported 53,101,847 shares, 12.8% of the company, in a Schedule 13G for 31 March 2026, with no later amendment found.¹⁸ I found no published view from it on Sunbelt and did not verify any analyst rating, so I leave those out.

Against reality

Much of that case is already in the numbers. Fiscal 2027 is guided to $4.92B to $5.12B of EBITDA, 5% to 10% above last year, and the first quarter grew EBITDA 8.7% on rental revenue up 12.5%.⁵ My base takes the middle of it. If Sunbelt grows rental revenue about 10% through fiscal 2028, lifts revenue for each dollar of fleet to 57% or 58%, holds an EBITDA margin of 44% or better, and lets the fleet age without a catch-up bill, my bull of $67 is the right picture, and the price of $80.73 is possible at a discount rate of about 9.1% E (estimate). The price at my 10% needs more than that.

The distance today is in three places. General Tool dollar utilization is 47%, where the bull wants the whole company at 57%.⁵ The EBITDA margin is 42.2% against a bull of 44% and above.⁵ And the fleet-age bill has not arrived: spending was light in fiscal 2026 and has only just risen.²⁵

Multiply it out. My base needs five things: fiscal 2027 on the raised guide (85%), dollar utilization holding at 55% to 56% (60%), margin near 42% (65%), no downturn before fiscal 2030 (65%), and no bigger fleet catch-up (60%). Those are my guesses, and they give 13% E (estimate). My bull needs mega-project and Specialty growth carrying rental growth of 9% or more through 2028 (45%), dollar utilization near 58% (40%), a margin of 44% to 45%, above any year on record (30%), a fleet that can age with no catch-up (60%) and no value-destroying deal (85%): 3% E (estimate). The items overlap, so call the odds a few times higher. They are still small. The price is a bet on the bull.

What a 10x would take. Ten times $80.73 is $807, or about $333B of market value on today’s shares. At 20 times earnings that needs about $16.6B of net income, 12.6 times fiscal 2026. At the current net margin of 11.9% it needs about $140B of revenue, 12.6 times this year’s. The company’s own figures put its share of the North American rental market at 11%, which makes that market about $87B C (calculated).² So a tenfold price would need more revenue than the whole North American rental market, unless the multiple expanded a great deal. By 2036 it needs the price to compound at 26% a year for ten years, against adjusted EPS growth of 12% a year from fiscal 2017 to 2026, on mixed accounting bases C (calculated).¹⁰⁴ Buying back 3% of the shares a year cuts those sums by about a quarter, and the revenue is still above the market. I know of no equipment renter that has risen tenfold from this size in a decade, and I did not build a list to check. For Sunbelt, tenfold is close to impossible, and I would not build a case on it.

Where I come down

The Specialty returns moved me: a quarter of the fleet earns around 24% and is growing fast. The last two years’ cash flow did not, because it came from spending less than a fleet needs. Dollar utilization climbing back toward 57% without a jump in fleet spending would raise my values most, and a split of replacement and growth capex would help more than any other disclosure.

VIIIRisks, and what would change my view

The three ways this goes badly wrong are at the start of Valuation. These are the rest.

• Demand concentrated in mega projects. Data centers are about 13% of the mega-project universe management tracks.¹² A pause in a few large builds could move revenue in a quarter, and the localities that have placed moratoriums on data centers show it can happen.¹²
• Local construction. The 10-K says the local commercial market is weaker than in recent years because of the long period of higher interest rates, and that it hits the small and mid-sized projects that are an important part of its customer base.²
• Equipment cost and tariffs. Tariffs could raise what new fleet costs, and the 10-K warns that they might do so by more than anticipated.² A higher cost of new fleet raises both replacement spending and the cost base for dollar utilization.
• One-off revenue. Hurricane response added $90M to $100M to fiscal 2025 and did not repeat.² The World Cup added about 250 basis points to the first quarter and will not repeat either.⁵
• Used-equipment prices. If they fall, the gain on sale goes and depreciation may have to rise. A 1-point drop in expected resale value across the fleet is about $21M a year of extra depreciation, and a shorter life of one year about $251M.²
• Refinancing and rates. At 30 April 19% of debt was floating rate and no interest-rate swaps were in place.² Cheap 1.5% notes have just been repaid and replaced by 5% to 6% money.³¹⁴
• Accounting, controls and people. A material weakness in debt classification was disclosed and reported fixed. The new US GAAP numbers are less than a year old, outside auditors have not yet attested to internal controls, and the general counsel is leaving and the finance chief started in 2025.²¹⁷⁷
• The UK. It earns about 4% on its fleet, and the goodwill test passed with little room.²
• Price paid for growth. Bolt-on deals have cost about two and a half times their fleet. The pipeline is “very strong” in management’s words, and at a price like Reliant’s many deals would dilute the return.¹²
• A larger holder. One holder, Dodge & Cox, owns 12.8%, and a change in its view could move the price on volume alone.⁷ I note that without implying it will act.