Casella Waste Systems

IThe business

Casella picks up garbage, moves it to a transfer station, and buries what is left in its own landfills. That is the short version. It is a regional company, strongest in New England and upstate New York, headquartered in Rutland, Vermont, where it started in 1975 with one truck.¹⁵ It now has about 5,600 employees, 86 collection operations, 72 transfer stations, 32 recycling and processing facilities, eight landfills for household waste and one for construction debris, in eleven states.³ About 270 employees, or under 5%, are covered by union contracts.³

It is the largest public company in Vermont by market value, which is the reason I picked it. It is not a company I screened as undervalued, and I do not think it is cheap, as the valuation will show.

The money comes from three sources. A customer pays a monthly fee for pickup. A truck or a transfer station pays a per-ton fee, called a tipping fee, to dump at a landfill. And a recycling plant sells paper, cardboard, plastic and metal. Collection was 65% of 2025 revenue, disposal 14%, and the Resource Solutions segment, which processes recycling and brokers waste for national accounts, 20%.³ The business is seasonal: construction debris slows in winter, so revenue is higher from late spring to early fall.³

A truck that picks up trash from a Vermont restaurant, drops it at a Casella transfer station, and ends up in a Casella landfill pays Casella three times. That is what “vertically integrated” means, and it is why the company talks about internalization: getting as much of its own collected waste as it can into its own landfills. In the second quarter of 2026, landfill revenue was $62.4M before intercompany sales, and 53% of it, $33.1M, was sold to Casella’s own trucks and stations.² The rest, $29.3M, came from outside customers. So about half the landfill line is a transfer between Casella pockets, and the outside half is where the pricing shows.

Four segments, one weak

Casella reports three regions and Resource Solutions. The Eastern and Western regions are the old business, and they earned $72.6M and $121.2M at the operating line in 2025, after depreciation and the amortization of acquired intangibles. The Mid-Atlantic region is the new one. It came from a purchase of four subsidiaries of GFL Environmental on 30 June 2023 and a string of tuck-ins after it, and it lost $20.4M at the operating line in 2025 on $341.1M of revenue, after losing $19.6M in 2024.³⁸ Part of that is the cost of buying companies and amortizing what was paid for their customer lists. Part is slower integration: in 2024 the company wrote up the unpaid bills it inherited from GFL, because it could not chase them during the handover period.³ It says it has now finished moving the Mid-Atlantic businesses onto its own systems and has started route changes.² I can see the loss. I cannot see how fast it will close.

Landfills: the part that matters

The landfill is the business inside the business. A collection route can be copied by anyone with a truck. A landfill needs a permit, land, neighbors who don’t sue, and decades of regulator patience. The 10-K says that over the last ten years, disposal capacity has tightened in the Northeast as sites have closed and more waste leaves the region by rail for out-of-state sites.³

The numbers are specific. At the end of 2025 Casella had 48.5 million tons of permitted capacity left, and another 49.7 million tons it calls permittable: land it controls where it expects to win a permit.³ In 2025 it filled 3.7 million tons.³ At that pace the permitted space lasts about 13 years, and the permitted plus permittable space about 27 years C (calculated). Remaining permitted space fell by 4.4 million tons in the year, as 3.7 million tons were used and engineering revisions took 0.8 million.³ Total capacity, permitted plus permittable, has fallen from 107.2 million tons at the start of 2023 to 98.2 million at the end of 2025, so the company is using space faster than it is adding it.³

Two of those landfills are about to stop taking waste. On 1 October 2026 Casella said that NCES, in Bethlehem, New Hampshire, will take its last municipal waste on 31 December 2026, a year earlier than it had said before.¹⁶ And the county that owns the Ontario County landfill in New York, which can take 0.9 million tons a year, voted in December 2024 to close it when the 25-year agreement ends in 2028.³ That is a good fact and a bad one. Casella has not said where NCES’s tons will go, only that it has options in its transfer and disposal network and third-party sites.¹⁷ It tightens supply in the region, which supports price. It also takes volume and a high-margin asset off Casella’s own books, and it starts a closure bill.

The company has tried to add capacity and has been refused. On 3 April 2025 New Hampshire’s environmental regulator denied the application for the Granite State Landfill, a project with $13.5M of capitalized development cost on the books, and the company is in court.³ In Maine a court sent the public-benefit approval for the Juniper Ridge expansion back to the state in January 2026.³ New Hampshire has since passed a law that sets up a review committee for new and expanded landfills and bars expanding capacity until at least 1 July 2027, and in June Casella withdrew a key wetlands application for a proposed landfill in Dalton.¹⁷ In New York it has won an expansion permit for Hakes and is seeking one for Hyland.¹⁷ In Pennsylvania the McKean landfill has taken rail deliveries since 2024, and the company has put $4.5M of rail construction into it in the first half of 2026 alone.³²

Price against volume

The cleanest way to see this business is price against volume. From 2019 to 2025 solid waste price rose between 3.9% and 7.5% a year, about 45% compounded, and volume was down in five of seven years, about 10% compounded C (calculated).^5181920873 In the second quarter of 2026 price was up 5.5%, with landfill pricing up 4.5% on household and construction waste, and volume was down 0.6%.²

This is how a mature utility-like business grows. It does not move more tons. It charges more for the same tons. That is good as long as customers can’t leave, and it is the reason the stock trades at a premium. It is also the reason I look hard at costs: if price is 5% and the margin moves only a point or two in six years, then cost inflation is taking most of the price.

Recycling, the commodity piece

Recycling is the part of Casella that moves with the price of cardboard, paper and plastic. It is smaller than it sounds. Casella says a change of $10 a ton in recycled commodity prices moves annual operating income by about $1.0M.³ It protects itself in four ways: processing fees where it can, a floating fee on collection customers that rises when commodity prices fall, plant upgrades, and long-term sales to industrial buyers.³ In 2025 recycled commodity prices cut Resource Solutions revenue by $20.4M, or 18.9% of the related line, and higher tipping fees added back $14.2M.³ I treat recycling as a nuisance and not a thesis.

Fuel is the other pass-through. A 40-cent move in diesel changes fuel cost by about $5.9M a year and the fee charged to customers by about $5.6M.³ In 2026 diesel rose, and the fee covered the cost but cut about 0.4 points from the second-quarter margin, because it adds revenue without profit.²

Acquisitions, and the FCC Environment question

Casella buys companies, and it buys a lot of them. It has acquired 76 collection, transfer and recycling businesses since the start of 2018, with over $925M of annualized revenue.³ The big ones were the four GFL Environmental subsidiaries on 30 June 2023 and Twin Bridges, near Albany, on 1 September 2023; all seven businesses bought in 2023 cost $846.6M.⁸ In 2024 it bought eight for $467.9M, in 2025 nine for $229.8M, and on 1 January 2026 it added Mountain State Waste, which took it into West Virginia.³ On 1 April 2026 it bought Star Waste Systems, which the company says has about $100M of annual revenue.²¹ In the first six months of 2026 it paid $399M in cash for four businesses with about $150M of annual revenue, which is 2.7 times revenue C (calculated).⁶²¹

I looked for a Casella deal with FCC Environment and did not find one. The name is not in the 2025 Form 10-K, the 2025 and 2026 earnings releases, the proxy statement or the 2026 current reports I read, no Casella 8-K since the start of 2025 carries Item 2.01 for a completed acquisition, and web searches for Casella with FCC Environment or FCC Environmental Services returned the GFL deal and Casella’s smaller deals, nothing else.³⁶²¹⁴ Casella announces some deals only by press release, so a filings search is not proof. I treat it as not a Casella transaction, and I will correct this page if one is shown. One announced deal I did find is Waste Reduction, Recycling & Transfer, a northeastern Pennsylvania hauler with a transfer station. Waste Dive reported on 4 September that Casella had signed to buy it; price and revenue were not given, and it has not closed. I leave it out of my numbers.²²

A tuck-in is simple. Casella buys a small hauler, cuts overlapping routes, runs its waste through its own transfer station and landfill, and the same revenue earns more. That works when the price is low. The price paid and the earnings bought are not disclosed deal by deal; the company gives annualized revenue only. If the businesses earn a 20% to 25% EBITDA margin before savings, $399M for $150M of revenue is about 11 to 13 times EBITDA E (estimate). The company counts the capital to bring trucks up to standard as part of the price: acquisition-related capital spending was $66M in 2025 and is guided to $76M in 2026.⁴²

The shares trade at about 16 times 2026 Adjusted EBITDA C (calculated). Buying earnings at 10 to 13 times and valuing them at 16 times is a gain on paper for as long as the market keeps paying 16. It is not a gain in cash. I come back to this in the valuation.

IIThe moat

Does Casella have a lasting edge? In one place, yes. In others, less than the stock price suggests.

Where it is real: landfills

The edge is the permit. A new landfill in New Hampshire was refused. The expansion in Maine is back in front of regulators. NCES is closing early. When a competitor can’t add capacity, whoever has it can raise the tipping fee. Casella’s landfill price rose 4.0% in the second quarter on the whole, with household and construction waste up 4.5%.² I can see the lever working, and I can see it in the 10-K’s own words: the company says that if more sites close it sees room for higher prices and more of its own waste going into its own landfills.³

A second part is density. A collection route is cheaper to run when the next stop is close. Casella has been buying inside its existing areas for this reason, and Star Waste, it says, overlays the New England footprint.²¹ Dense routes are an advantage for the company with the most trucks in an area, and a trap for a company without them.

Where it is thinner

Look at who owns the ground. The State of Maine owns Juniper Ridge, and Casella runs it under a 30-year agreement that began in 2004. Ontario, Chemung and Clinton counties own their landfills, and Casella runs them under agreements; Ontario’s ends in 2028.³ Of the seven landfills with an annual limit stated in tons, the three county-run ones account for 1.6 million of 4.8 million tons a year C (calculated). The permits are valuable, but some come with end dates: the edge is less a wall than a set of licenses.

Look at volume. It fell in five of the last seven years. I don’t know how much was chosen and how much was lost; the filings do not split it. But a company with a strong grip on supply would not be expected to lose tons year after year, and I read the pattern as price being bought partly with volume E (estimate).⁵³

Look at the margin. After seven years in which price rose 45%, the Adjusted EBITDA margin was 21.1% in 2019 and 23.0% in 2025, so two points C (calculated).²³⁴ Costs rose almost as fast as price. In 2025 direct labor rose from 15.3% to 16.2% of revenue and maintenance and repair from 8.7% to 8.8%, while direct costs fell from 23.3% to 23.1%.³ A company with a wall around it takes more of its price increases to the bottom line than this.

Where it is a cost, not an edge

The landfill is also a liability. Every ton buried creates a future bill to cap, close and watch the site for 30 years. I give this its own treatment in the financials: the accrued liability was $44M in 2016 and is $205M now, the 10-K’s undiscounted payments are $412M, and most of the cash falls after 2030.³⁶¹³ Whoever owns a landfill in the long run owns that bill. An honest measure of the edge has to be taken after it.

And PFAS, the long-lasting chemicals the US Environmental Protection Agency has just named hazardous substances under the Superfund law. The company’s own risk language says they will likely raise compliance and remediation costs and litigation risk.² I can’t size it, and neither can Casella. I leave it out of the base case and name it in the risks.

IIIThe competitive advantage

The last section asked what protects Casella. This one asks how it stands against the companies it competes with, and the answer is by size and margin, not by quality of place. In its larger markets it competes with Waste Management, Republic Services and Waste Connections, any of which, the 10-K says, may have greater economies of scale.³

The cleanest comparison I can build from filings is a crude one: operating income plus depreciation, depletion and amortization, over revenue, for the twelve months to 30 September for the three big companies and to 30 June for Casella, because Casella has not yet reported the third quarter. It is not Adjusted EBITDA for any of them. It is the same cut for all four, which is what matters.

Three things to say about the table. First, Casella has about one-tenth the revenue of Republic and one-thirteenth that of Waste Management C (calculated). It has no scale advantage and doesn’t claim one. Second, its crude margin is 7 to 11 points lower, and Casella’s own Adjusted EBITDA margin of 22.7% in the second quarter, which adds back accretion and landfill-lease depletion, is still well below the 27% to 31% the others earn on a plainer measure.² Some of the gap is the Mid-Atlantic and the cost of owning many new businesses at once. I can’t tell how much of the rest is structural. Third, the share price asks for more than theirs: 19.5 times against 15.2 to 18.7 times C (calculated).

The speed of growth is where Casella stands out. Revenue rose 16.9% in the second quarter, of which acquisitions added 11.2 points of solid waste revenue and price 5.5.² That is a roll-up and not a compounder of a single wheel. Casella has never paid a dividend, and it has issued shares four times in ten years: $100M in 2019, $145M in 2020, and $496M in each of 2023 and 2024.¹³

What would a competitive advantage look like here? It would be a margin that stays in step with the large three while price keeps running ahead of cost. I don’t see that yet. I see a real asset base, with a worse margin, a higher multiple, and a plan that needs the multiple to hold.

IVManagement

John Casella started the company in 1975 with one truck, and was chief executive for more than 30 years. On 1 January 2026 he became executive chairman and Ned Coletta, who had been president since July 2022, took over as chief executive.¹⁴ His brother Doug is vice chairman and runs the operating subsidiary. The two brothers own all 988,200 Class B shares, which carry ten votes each and are 13.8% of the votes.¹⁴ The board has ten members in three staggered classes, with Joseph Doody, a former vice chairman of Staples, as lead director.¹⁴ It is a family-led company with a classified board, and I do not read that as a defect or as a virtue. It makes it harder for an outsider to force a change, and it means the family’s judgment is the one that counts.

Turnover in the operating team

Two changes in 2026 matter. Sean Steves, the chief operating officer for solid waste, resigned on 6 March to work outside the industry; the company said it was not a disagreement.²⁴ The board then hired Damian Ribar from Waste Connections as chief operating officer of the whole company, due to start about 20 July.²⁵ A chief executive of 50 ten months into the job, a new operations chief from a competitor, and a 75-year-old founder as chairman: that is a handover, and I can’t tell from the outside how it will go.

How they are paid

Follow the pay. The 2025 bonus was based on Adjusted Operating Income, Adjusted Free Cash Flow, and two goals on safety and turnover. The long-term awards, 75% of the stock given to executives, pay on Adjusted Free Cash Flow and Adjusted EBITDA in the third year, adjusted up or down by how the share price did against the Russell 2000.¹⁴ These are the company’s own definitions. They leave out the cost of acquiring and integrating businesses, a good part of the capital that new businesses need, and, in the free cash flow measure, nothing of the equipment bought through finance leases.⁴³ I read that as a plan that pays people to grow EBITDA and keep the measured cash up, and I think it is a good plan for a roll-up. It does not pay anyone on net income, return on invested capital, or the closure bill, which are the things I am worried about.

Two details are worth a look. The bonus gate is $158.3M of Adjusted Free Cash Flow, which is exactly the 2024 result.¹⁴¹⁵ So the first dollars of the 2025 bonus depended on beating last year. And the plan lets the committee raise the targets when the company buys something big, so that executives don’t profit from acquired earnings that were not in the budget.¹⁴ That is sensible and it is also an open door: the pro forma models behind the adjustment are internal and I can’t check them.

The chief executive’s total pay for 2025 was $4.41M, or 55.7 times the median employee’s $79,050.¹⁴ I don’t think that is out of line for a company this size. John Casella’s family also benefits in other ways: Casella Construction, owned by his son and Doug’s son, was paid $8.56M in 2025 for landfill development and closure work, and two of John’s children are employed by the company.¹⁴ The board approves these under a bidding procedure and I have no evidence of a problem. I just note that the company that builds and closes its landfills is partly a family company.

Skin in the game

The founders are rich in stock but small against the company. John Casella’s 514,426 shares, counting Class B, are worth about $51M at the close and Doug’s 595,478 about $59M. With the rest of the directors and officers that is about $140M of a $6.3B company C (calculated).¹⁴¹ The company prohibits hedging and pledging and has ownership guidelines.¹⁴ In 2026 the chief executive sold 12,500 shares in May at about $90, Doug Casella sold 16,500 in August at about $91, and a director sold three small lots, while in the 29 Forms 4 I read no officer or director bought a share with cash.^26272829

An activist arrives

On 20 October 2026 Jana Partners, an activist fund, said it had a new position in Casella, and the shares rose 7.5% that day, from $86.54 to $93.07.³⁰¹ Investing.com reports that Scott Ostfeld of Jana said so at an investor conference in New York, and that the fund wants better margins and returns on invested capital, pay tied to shareholder value, a public commitment to a 27% Adjusted EBITDA margin by 2029, and a consulting firm to improve performance.³⁰ I found no Schedule 13D, no 8-K and no reply from Casella, so the stake size and the demands rest on that one news report, and I do not know how large the stake is. Two things stand out. The demands match what I would ask of this plan: the pay plan has no measure of return on capital. And the family’s Class B shares carry 13.8% of the votes, so an activist with a few percent of the shares has less leverage than usual. I can’t tell whether the board will act, or whether 27% is reachable, and I give it weight in the bull case, not the base.

What they did with capital over a full cycle

Judge a capital allocator by what a decade added up to. From 2016 to 2025 Casella took in $1.8B from operations, spent $1.3B on equipment and landfills, $2.0B on acquisitions, and raised $1.2B of new stock.¹³ Debt rose by $643M. Shares outstanding, on the diluted weighted measure, rose 54%.¹³ Per share, Adjusted EBITDA rose 2.3 times and the company’s Adjusted Free Cash Flow 4.3 times C (calculated). The share price rose 7.9 times to the end of 2025.¹³³¹⁴¹ So shareholders owned a smaller piece of a much bigger company, and were paid mostly in a higher multiple.

That was fair in a rising market. Equity raised at $100 to buy a company at 11 to 13 times EBITDA looks like good value when the stock is valued at 17 times E (estimate). The test of a roll-up is what happens when the stock stops rewarding it. In 2022 the shares fell 7%, and in 2023 the company bought $852M of businesses, paid for with $496M of new shares and $439M of net new borrowing.¹¹³ Whether that was a wise use of the next dollar, I leave to the Mid-Atlantic results, which at the operating line are two years negative.³

VFinancials

Ten years of figures, with the company’s own measures labelled as theirs. The adjusted lines are company definitions, not accounting standards. Casella has not yet reported the third quarter of 2026; it will do so after the close on 9 November.³² The latest filings are the second-quarter release and 10-Q of early August.

Read the table from the top. Revenue and Adjusted EBITDA went up almost every year, and the EBITDA margin went from 21.3% to 23.0%. Operating income did not follow. It was $60M in 2020 and $64M in 2025, because depreciation and amortization went from $91M to $307M, a result of buying businesses and the intangibles that come with them. Net income of $8M in 2025 compares with Adjusted Net Income of $80.6M. The bridge is about $76M of amortization of acquired intangibles and $24M of deal costs added back, less $29M of tax on those items.⁴ Net income in 2020 includes a $52.8M tax benefit, so it is not a clean number.¹³

Adjusted EBITDA is not owner cash

Here I can say something specific. In 2025 Adjusted EBITDA was $423M. Cash from operations was $330M. The company’s Adjusted Free Cash Flow was $180M, and net income was $8M. These are four different answers to “what did the business earn?”, and each is real in its own way. The one I care about is the cash an owner could take out without making the business smaller.

Company Adjusted Free Cash Flow is a reasonable measure of what the business generates before it grows by acquisition, and the company says so: it adds back acquisition capital spending, deal costs and the McKean rail line.⁴ I agree with leaving out the purchase price. I would go two steps further. First, stock pay: $14.2M in 2025 is added back in operating cash flow because it is not cash, but it is a cost, paid in shares that dilute everyone. Second, equipment bought on finance leases. In 2025 the company took $45.3M of equipment on finance leases, which never shows up as capital spending or in the free cash flow, though the principal is repaid in financing cash flow.³ In April 2026 it signed a master lease with Huntington National Bank for vehicles and equipment, with no stated limit on the amount.³⁸ A truck on a lease is still a truck the business needs. I deduct both.

That gives owner cash of $120M for 2025, 67% of the company’s measure. It has grown from $63M in 2021 at 17.6% a year, which is good C (calculated). For 2026 I start from the guide midpoint of $205M and take off the same two items to get about $141M. Against a market value of $6.3B that is a yield of 2.2% C (calculated). If I also count the acquisition-related capital spending and deal costs, the cash after every dollar spent was $26M in 2025.

Capital intensity, and where it is going

Casella spends a lot to stand still. Replacement capital spending, which renews trucks, containers, facilities and landfill cells, was $162M in 2025, 8.8% of revenue, and $82M in the first half of 2026, 8.2%.⁴² Add finance-leased equipment and growth items and total capital was 15.8% of revenue in 2025 C (calculated), against an Adjusted EBITDA of 23.0%: about two-thirds of it, before interest, tax or closure.

The year so far

The first half of 2026 looks fine on the surface: revenue rose 13.4% to $1.00B and Adjusted EBITDA 12.5% to $220M. Under it, the margin slipped from 22.2% to 22.0% for six months and from 23.5% to 22.7% in the second quarter, with fuel recovery fees explaining about 0.4 points.² Net income was a loss of $1.8M. The company cut its net income forecast twice, from $16M to $22M in February to $0 to $6M in August, while raising revenue guidance by 6%.⁴²¹² More revenue, less profit: the cost of buying companies.

Debt

Casella is not over-borrowed, but it has used its room. Debt principal was $1.17B at the end of 2025 and $1.36B on 30 June 2026, after spending $400.8M on acquisitions with $25.5M of cash left, including $185M drawn on the revolver.²⁶ The August release counts five 2026 deals and the 10-Q four by 30 June, so one closed in July: about $15M of annual revenue, in the guide but not in June’s debt, and I add nothing for it C (calculated).²²¹ The lender’s ratio of net debt to its own version of EBITDA was 2.34 times at the end of 2025 and 2.70 times in June, against a limit of 4.00, and it counts $67.5M of estimated annual earnings from acquisitions as if owned for the full year.³⁶ On the $447M the company reported over the last twelve months, the same debt is 3.0 times C (calculated). The honest range is 2.7 to 3.0 times, and I use the higher number when I compare.

The rate is mostly fixed; a $15M bond was reset on 1 June at 4.3% for ten years.³⁹ Of the debt, $273.5M is tax-exempt bonds at fixed rates between 2.9% and 5.1%, $515M of the variable loan is swapped at an average of about 3.6% before the margin, and the rest floats; a one-point rise in rates adds about $2.9M a year to interest.³ The term loan matures in September 2029, and $842M of the debt falls due in the three-to-five-year bucket.³ Interest is guided at $67M for 2026, or 7.1 times Adjusted EBITDA C (calculated).² That is comfortable. The risk is not interest. The risk is that the next few deals are bought with more debt at 2.7 to 3.0 times.

The closure bill

A landfill has to be capped when full and watched for 30 years after. Casella records a liability for it: $193M at the end of 2025 and $205M on 30 June 2026, up from $44M in 2016.¹³⁶ It grows each year as the company buries more waste ($8.4M of new obligations in 2025), as it accrues interest on the liability ($14.5M), and when engineers revise costs ($29.5M in 2024, $4.6M in 2025).³ Cash paid out against it is small so far: $7.0M in 2025.³

The balance sheet number is a present value. The 10-K’s table of what the company expects to pay, undiscounted, is $412.3M: $9.8M in 2026, $56.9M across 2027 and 2028, $33.7M across 2029 and 2030, and $312.0M after that.³ So the cash cost is about to rise from $7M to something like $28M a year in 2027 and 2028, and it keeps rising as landfills fill. I suspect the end of the Ontario County agreement in 2028 is part of it, though the filing does not say so. These payments are already inside operating cash flow, so the company’s Adjusted Free Cash Flow is after them. The point is that the 2025 base year flatters the next ten, and I model the schedule in the valuation.

Balance sheet quality

Goodwill was $1.12B at the end of 2025, 34% of total assets, plus $291M of intangibles, against equity of $1.57B. Take those away and tangible equity is about $158M C (calculated).³ Impairments come late and arrive in a batch.

VIValuation

First, how this goes badly wrong

Before the numbers, I asked what would most likely hurt a holder of this stock, and worked backward. Three things could, and none needs a shock.

• Cost catches price while the tons leave. Price has risen 5% a year and the margin has risen 1.9 points in six years. NCES stops taking waste on 31 December and Ontario County ends in 2028, which takes two landfills and their volume off Casella’s books.¹⁶³ In a year like 2020, volume fell 7.2% with price rising 4.5%.¹⁸ If price slows to 3.5% and costs keep rising at 4%, the margin falls and a price-led business has no growth left. This sets my bear.
• The capital and the closure bill are larger than reported earnings suggest. Standing still costs 11% to 12% of revenue in equipment and landfills. Closure cash is $7M a year and the 10-K’s own table has it at $28M a year in 2027 and 2028, then more.³ PFAS rules could add a cost no one has priced.² None of this is in a headline number such as Adjusted EBITDA. It is in my owner cash.
• The multiple that pays for the deals goes away. The business trades at about 16 times EBITDA, and it is buying at an estimated 11 to 13 times E (estimate). Each turn of EBITDA multiple is worth $7.50 a share; at 12 times 2026 EBITDA, the shares would be worth $69, 30% below the close C (calculated). The three large companies trade at 15 to 19 times crude EBITDA, so I do not think the multiple is odd, but it is not a number the company controls.^9101112

The first two set my bear and base. The third I do not put in the model. It is the part of the price that is not about cash.

What the price implies

At $98.88 the market values Casella at $6.30B on 63.7M diluted shares: 63.64M outstanding in two classes on 15 July plus 0.07M for share units C (calculated).⁶¹ Add net debt of $1,338M on 30 June for an enterprise value of $7.64B. That is 16.0 times the 2026 guide midpoint and 17.1 times the earnings reported over the last twelve months C (calculated).² It is 78 times the company’s Adjusted earnings a share for 2025 and 52 times my owner cash a share C (calculated).⁴

What does the price pay for? Hold the margin on my base path and the price needs revenue to grow 12.7% a year for ten years, to $7.0B in 2036 C (calculated). The company grew revenue 14% a year over the last nine, but with $2.0B of acquisitions and a 54% larger share count. Or take my base growth, and the price works only if the Adjusted EBITDA margin is 35% for ten years, against the 20% to 23% Casella earns now and the 27% to 31% the three large companies earn on a plainer measure C (calculated). Or take my base cash flows and ask what the market must pay for the business in 2036: 17.4 times EBITDA, roughly the multiple today C (calculated). That last one is the clearest reading. The price says: nothing about the multiple changes, and Casella grows its EBITDA 5% a year for ten years, and then you earn 8%.

What I can’t judge

Four inputs decide this, and I can’t know them. The first is the normal margin of the Mid-Atlantic region, which the company reports only after depreciation and amortization, at a loss; I can’t see its EBITDA. The second is how much of the next ten years’ closure and capping cost is already recorded; the company revises these estimates by $5M to $30M in a year.³ The third is what Casella pays for its acquisitions; it tells me revenue and cash, not earnings. The fourth is PFAS, which no one can price yet. The shares also rose 23% between 30 September and the Q4 2026 close. One part I can explain: 7.5% came on 20 October, when Jana Partners said it had a new position.³⁰¹ Another 7.4% came on 22 October, and I found nothing that explains it. So I give a range and not one number. Read my $45 with the grid next to it.

The discount rate

I use 8.0% on the cash before interest, which is the cost of the whole business and not of the shares alone. The ten-year Treasury yielded 4.90% in early November, and Damodaran’s implied equity risk premium was 4.41%.⁴¹⁴² A beta measured on 105 weeks of returns against the S&P 500 comes out at 0.17, which I don’t use; a number that low for a company with this much debt and this many deals says the data are noisy, not that the business is risk-free C (calculated).¹⁴³ I use a judgement of 0.9, which gives a cost of equity of 4.90% plus 4.41% times 0.9, or 8.9%. The debt costs about 5.2% before tax, since the term loan was 5.19% on 30 June, or about 3.9% after tax at 25%.⁶ Equity is 82% of the enterprise value and debt 18%. The mix gives 8.0% E (estimate). It is lower than the 9.0% a purely equity-funded business of this kind would ask, and the sensitivity table shows both ends. One consequence the market could fairly point to: at 8.0% and 3.0% growth my base values the business after 2036 at about 7.5 times that year’s EBITDA E (estimate), because about 36 cents of each dollar of EBITDA reaches the owner once equipment, landfill closure and tax are paid. The market pays 16 times today. If the true discount rate is nearer 7% and growth 3.5%, the base is $69, still under the price, so the gap does not close on rates alone; it needs the multiple to hold.

Three cases

The model is simple on purpose. I set revenue growth and an Adjusted EBITDA margin for each year from 2027 to 2036. From the margin I take, in order: cash tax, ramped from 8% of adjusted operating income in 2027 to 25% by 2031 as the $124M loss carryforward and bonus depreciation run out; capital at 8.6% of revenue for replacement plus 1.0% for organic growth; finance-leased equipment at 2.5%; landfill-lease cash at 0.35%; closure cash on the 10-K’s table, $28.4M a year in 2027 and 2028, $16.8M in 2029 and 2030, then rising from $22M to $37M by 2036; and working capital at 4% of added revenue.³ I start from the middle of the August guide, $2.10B of revenue and $478M of Adjusted EBITDA, plus a second-half 2026 stub of about $59M E (estimate).² I include no acquisitions beyond those closed: the company discloses price only in total, and a deal bought at its value adds nothing to mine. Net debt is the 30 June figure, $1.34B.

My 2027 revenue of $2.23B is 6.2% above 2026: about 4.7 points from price and volume, and 1.5 from the full year of this year’s deals. That second figure is the 10-Q’s pro forma: the deals added $30.8M of revenue to the first half had they been owned from the start of the year C (calculated).⁶ The company usually gives its next-year guidance with fourth-quarter results in February. Where I differ from the optimists is on what the growth costs.

Where would $2.5B of revenue in 2030 come from? From price at 4% to 5% a year, which Casella has delivered in each of the last seven, and volume near zero, better than the −1.4% average of those years, because I assume Casella’s own sites pick up some of the closing landfills’ tons. The base fades that to 3.2% by 2036. The doubt is volume: the bull needs +1%, the bear −1%, and the history favors the bear. I can’t resolve it.

The price sits 122% above my base, 26% above my bull, and 4.4 times my bear. That is the plain result. Why might the market be right? Because it sees what I see and adds a view on three things I leave out. One is the multiple: the three large companies trade at 15 to 19 times, and a regional company with the last scarce landfills in the Northeast may deserve the same. Two is acquisition: if Casella can keep buying $150M of revenue a year at 11 to 13 times and the market keeps paying 16, the paper value rises about 4% a year on its own. Three is a lower required return: at 7% instead of 8%, my base is $61, my bull $105, and the price is still above both.

On quality and price separately: I think the business is a good one. It owns scarce disposal capacity, prices ahead of inflation and has a long runway. The price is another matter. At $98.88 I get 5.7% a year on the base case against 4.90% on a Treasury, and I would have to be wrong about volume, margin and capital all at once for the price to be fair. A good business at a price with little room for error is a common story, and I think this is one.

Other uses of the same dollar

The ten-year Treasury pays 4.90%. My base here returns 5.7% a year at the price and my bull 7.2% E (estimate).⁴¹ The three large waste companies are the natural alternative, and I looked at them only as far as a crude multiple: they cost 15.2 to 18.7 times EBITDA against Casella’s 19.5, and earn higher margins on the same measure.^9101112 I have not valued them, so I can’t say which is cheaper. The nearest on this site is my Williams page: it prices a capital-heavy infrastructure business at a return of 9.0% on its cash flows. On that comparison Casella at $98.88 offers less.

The margin of safety

I set the margin-of-safety price at $29: my bear value divided by 0.75. Below it, the bear case would still leave about three-quarters of the price paid. That is a rule, not a forecast. The price is 241% above it. The shares last closed below $29 on 8 January 2019, and below my base value of $45 on 4 May 2020 C (calculated).¹

The weighted value, with 25% on the bear, 50% on the base and 25% on the bull, is $48. The weights are mine. I use them only to see that no sensible weighting reaches the price.

VIIThe case for the price

What most people miss

Three facts cut in Casella’s favor and get less attention than the acquisitions. First, price has not cost volume where it matters most. In the second quarter of 2026 landfill prices rose 4.0% and landfill volume rose 8.5%, the opposite of collection, where volume fell 1.4%.² Second, the company has not sold shares since 2024: shares outstanding rose 0.25% in 2025, from stock pay alone, and the 2026 deals were paid for with cash and borrowing.³⁶ Third, cash taxes are very low: a net refund of $0.2M in 2025, and about $124M of federal loss carryforwards going into 2026.³ I model that fading by 2031.

The bull case

In plain words: Northeast landfill space is scarce and getting scarcer, Casella owns a good share of what is left, it can raise prices every year, and it is rolling up a fragmented industry. Management makes that case. On 6 August 2026 its chief executive said pricing was strong, disposal pricing was up 4.7%, landfill volumes rose from outside sales and internalization, and the pipeline “remains very strong.”² The company has raised 2026 revenue guidance twice and Adjusted EBITDA guidance once.⁴²¹² Jana Partners, on 20 October, said it wants a 27% Adjusted EBITDA margin by 2029, a path close to my bull case, which reaches 27.0% in 2030.³⁰ Among analysts I verified one dated note: TD Cowen, on 6 June 2025, raised its target to $135 from $130, kept a positive rating, and argued consensus leaves out acquisitions, so growth looks too low and the multiple too high.⁴⁴ It is 17 months old, from a price near $117, and I could not verify a newer one by name. BlackRock (14.3%), Vanguard (10.5%), Conestoga and Wasatch (about 5.5% each) have published no case that I found.¹⁴

Against reality

The bull case needs four things at once: price stays ahead of cost, tons stay up as two landfills close, the Mid-Atlantic is fixed, and the market keeps paying today’s multiple while the company buys at lower ones. If they do, the price is possible. My bull value is $79, and the price needs more: my base cash flows and a 17.4 times EBITDA multiple in 2036, about today’s. The evidence today: pricing is real, at 5.5% in the second quarter.² The margin is not moving: 22.0% in the first half against 22.2% a year earlier.² The Mid-Atlantic lost $20.4M at the operating line in 2025.³ Owner cash was $120M, 1.9% of the market value.³⁴ Six months of acquisitions cost $401M, 2.7 times revenue.⁶

It helps to multiply it out. My base needs four things. Price at 4% or more with volume near flat: say 60%. The Mid-Atlantic margin reaching the other regions by 2029: 55%. No surprise in closure cost or PFAS: 75%. Capital intensity falling from 12.1% to 11.4%: 65%. Those are my guesses, and they multiply to 16% E (estimate). They overlap, so it is a little higher, perhaps a fifth. The price needs the same base with the multiple held, which I put at one in two: about one in ten E (estimate).

What a 10x would take. Ten times $98.88 is $989 a share, or about $63B of market value. That is more than Waste Connections’ $48B today and close to Republic’s $75B C (calculated).¹¹¹⁰¹² At 20 times EBITDA and about $3B of net debt it needs $3.3B of EBITDA, 6.9 times the $478M I expect in 2026. At the 30% margin the three large companies earn, that is about $11B of revenue by 2036, 5.2 times 2026 and 18% growth a year for ten years; Waste Connections’ revenue today is $9.9B C (calculated).¹¹ History does not help much. Casella’s own shares rose 9.0 times in the ten years to Q4 2026, from $11.03, but they started with a market value of about $0.45B, and EBITDA rose 3.5 times while the multiple went from 8.6 to 17.2 times. Waste Management’s shares rose 3.8 times, Republic’s 4.8 times and Waste Connections’ 3.9 times over the same ten years, excluding dividends C (calculated).¹⁴⁵ The multiple cannot double again on a business this size, so a tenfold move needs EBITDA to grow about 21% a year for ten years. I put that under 1% E (estimate).

Where I come down

What moved me is the pricing: 5% a year for seven years, with landfill volume up this quarter, is real, and the landfill closures will help it. What didn’t move me is the cash. Owner cash was $120M in 2025 against a market value of $6.3B, and the margin has risen two points in six years. What would raise my values is evidence on the Mid-Atlantic and on price paid for deals: a Mid-Atlantic at break-even would be worth about one point of margin, or $5 a share on my model, and deals bought under 10 times EBITDA would add value I do not model. The price needs more than that, and that is the plain result.

VIIIRisks, and what would change my view

These are the things that could be wrong with my page, with the company, or with both.

• Volume. Tons have fallen in five of seven years. If that continues, and price cannot offset cost, the margin goes down and growth ends. NCES and Ontario County are the tests.¹⁶³
• Closure and PFAS. The closure liability is $205M and the 10-K’s undiscounted table has $412M, with most of it after 2030.⁶³ PFAS is a regulatory and litigation risk that the company itself describes and cannot size.²
• Acquisition risk. The company is buying at 2.7 times revenue with debt. Integration has been slow in the Mid-Atlantic, and the company wrote up inherited receivables in 2024.³
• Leverage and rates. 2.70 times by the lender’s measure, about 3.0 on reported earnings.⁶ The term loan matures in September 2029 and $842M of debt falls due in 2029 and 2030.³
• Governance. A dual-class structure and a classified board; the family has 13.8% of the votes and 1.7% of the shares.¹⁴ A new chief executive and a new operating chief within seven months, and an activist who wants changes.³⁰
• Regulators and courts. New Hampshire refused the Granite State landfill and now bars landfill expansions until at least July 2027. A Maine court sent the Juniper Ridge approval back. The county that owns Ontario voted to close it.³¹⁷
• My model. It leaves out all acquisitions that are not yet closed and gives them no credit. If Casella’s deals earn more than I think, the value is higher. It also uses a cash tax rate that I ramp up by judgement.

Last, what I have not said. I have not said the stock will fall or that it won’t. I have said what the price needs and how much of that I can see. I have not valued the three large waste companies, so I cannot say Casella is dearer than they are on value, only on a crude multiple. And I could not verify any dated view from a named investor beyond the one note from June 2025, so the case for the price rests mostly on the company’s own words.