CSU.TO · Constellation Software 2,787.13 CAD -42.77 (-1.51%) 2026-SEP-18 12:47 EST

Constellation buys hundreds of small software companies that serve niche industries and keeps them for good. In March 2026 fears that AI would make software obsolete crushed the whole sector, and Vinall used the sell-off to buy.

What he likes is its people: the company is run by managers who think like owners (the new CEO founded the first business the group ever bought). Unlike many software firms, it doesn't hand out so many stock options that shareholders' stakes get quietly diluted.

Constellation buys small, boring software companies — the kind that sell scheduling systems to golf clubs or billing software to municipalities — and then holds them forever, using the cash they throw off to buy more of the same. Mayer calls it out twice, for two different reasons.

First, it's the company that changed his mind about growth-by-acquisition. When he wrote 100 Baggers he was, like most value investors, suspicious of acquirers; Constellation is the proof that a disciplined, repeatable, small-deal acquisition machine can be one of the best businesses there is.

Second, and more specific: its share count barely moves. That matters because every share a company issues (to pay staff, to fund a deal) hands a slice of your ownership to someone else. "That's why I love and admire these companies that have share counts that are unchanged over a long period of time. Constellation Software is an obvious one." This is a description of a model, not a buy call — he is not pitching the stock here.

Constellation buys small software companies that sell essential, unglamorous systems to niche industries, and never sells them. The engine is simply the rate at which it can keep buying, and this update quantifies it through RBC's estimate: $646 million deployed in the second quarter, $1.63 billion so far this year, and an expected $3.12 billion for 2026 — comfortably above the previous record of $2.46 billion set in 2023.

That matters because the shares are down 14% this year on fears that AI will make small software businesses obsolete. The acquisition figures say the opposite is happening operationally: the company is putting more money to work than ever.

The valuation gap is the largest in the portfolio. On the fair-value calculation the shares are worth $4,565 against a $1,909 price — more than double. The forward multiple is 15.8 against a five-year average of 31.5, so the market is paying half what it habitually paid for the same business. And the reverse calculation says the price only requires 7.5% annual growth, against the 15% expected. All three measures agree, which is the condition this archive requires before acting.

Constellation is the parent of that model and the original — more than a thousand small vertical-market software businesses bought and held, never sold, each one boring and each one critical to whoever depends on it.

Like Topicus, it appears here only in the overlap sentence naming the five companies Slegers and Akre both own. Its relevance to the profile is structural: Akre's stated search is for "companies that are able to reinvest their capital at high rates of return for long periods of time," and Constellation is the clearest listed example of exactly that — a business whose entire purpose is redeploying its own cash flow into more of itself.

Worth reading alongside the archive's other entries on the name, where it is a Very Strong conviction holding at what Slegers calls the cheapest valuation it has ever traded at.

Constellation is the largest discount in the whole book: the firm's fair-value estimate is more than two and a half times the traded price, and the shares change hands at about 16 times expected profits against a five-year average of 32. In other words, the business is not being valued differently because it earns less — it earns more — but because the market has decided software companies are threatened by artificial intelligence.

The reason the archive is comfortable is the shape of what Constellation owns: hundreds of small, dull, essential programs, each one cheap for the customer relative to the cost of replacing it. That is also what makes it the name a well-known outside investor was reported buying four days earlier, in the deepest share-price fall in the company's history.

Constellation buys small software companies — the unglamorous programs that run a bus depot, a golf club, a municipal water utility — and never sells them. Each one is tiny, each one is essential to the customer that uses it, and the licence fee is small enough that nobody notices it on the budget. That last detail is the whole argument here: when a company goes looking for costs to cut, it does not start with the software that would take two years and enormous risk to replace and that costs less than the coffee budget.

The market has decided that artificial intelligence will make writing software cheap, and therefore that companies selling software are worth less. That fear is what put the shares into the deepest fall in the company's history. Mohnish Pabrai — an investor who built his career on openly copying Warren Buffett — is named as a buyer at these prices, and the article uses him as the evidence that other serious people see the same thing.

The track record is the reason anyone is looking: roughly 30% a year since the 2006 listing, and a hundredfold return within fifteen years of going public.

Constellation buys small software companies that serve one specific industry — the systems a bus operator, a golf club or a public library runs on. These businesses are dull, nearly impossible to switch away from, and generate steady cash. Constellation buys dozens of them a year and has compounded at about 30% annually doing it.

The share price has halved over the past year while the business went the other way: sales up 20%, profit up from $136 million to $367 million, and the number that matters most for this kind of company — the cash actually available to shareholders after everything is paid — up 44% in the latest quarter alone.

The market's reasons for selling are listed here without being argued against: that AI will make this kind of software obsolete, that the company is now too large to keep finding enough acquisitions, and that it cannot repeat the record without Mark Leonard, its founder.

The counter-argument is simply arithmetic over time. If earnings and share prices eventually converge, and the earnings are rising while the price falls, the gap has to close in one direction.

Constellation buys small software companies that sell unglamorous, mission-critical programs to narrow industries — the software a marina, a bus operator or a municipal court runs on — and then holds them permanently, using the cash they generate to buy more. It has compounded at roughly 30% a year and became a hundred-bagger within fifteen years of listing.

The valuation section is the useful part, because ordinary earnings barely describe this business: accounting rules force it to write down the price of every acquisition over time, which depresses reported profit without any cash leaving. So Slegers uses Free Cash Flow Available To Shareholders — the cash actually left over for owners — which was about $1.68 billion in 2025. Assuming a further 15% of growth this year, the shares are priced at a 4.8% yield on that cash, which he says is the cheapest the company has ever been.

The sentiment marker he offers is a good one: the market is depressed enough about it that Mohnish Pabrai, an investor who buys statistically cheap things rather than compounders, has bought the shares.

Constellation buys small software companies — the dull, essential kind that run a golf club's bookings or a water utility's billing — and then leaves them alone to keep collecting their fees. It has done this hundreds of times, and this post files it as a holding company rather than a software company, which is the more honest description.

The machine is organised into six groups, each a business in its own right: Harris (about 30% of revenue, mostly utilities, healthcare and government), Volaris (25%), Topicus (15%, spun off in 2021 so it could use its own shares to pay for bigger European deals), Jonas (15%), Perseus (10%) and Vela (5%).

The record is the strongest on the list and is not disputed anywhere in this archive: about 30% a year since the 2006 listing, a hundredfold return inside fifteen years, and a total gain of more than 18,000%. No price or valuation is attached here — that work appears elsewhere in the archive, where the stock is a top-conviction holding bought during its first serious drawdown.

Constellation appears here with no words at all — just a chart of what investors have been willing to pay for its earnings over the past decade, with today's level at the bottom of the range. That is the whole argument, and in context it is a strong one: the issue has just argued that the market as a whole is in a bubble, and this is the counter-example.

The business is a collection of hundreds of small software companies that run one specific type of organisation each and are almost never replaced. Nothing about that changed this quarter; the price did. Compounding Quality added to the position two weeks before this article was published.

Constellation buys small software companies whose products run one specific kind of organisation — a hospital department, a school district, a transit authority — and holds them permanently. Because the software is essential and the customers are institutions, revenue barely moves.

The one number that matters for a business like this is not the price-to-earnings ratio but what it earns on each new dollar it invests, because reinvestment is the entire engine. Over the past three years that figure — cash return on incremental investment — has been 29.1%: for every $100 put to work, it generates about $29 a year. As long as that holds and there are companies left to buy, the compounding continues.

The shares have fallen because investors fear AI will make niche software easy to replace, and because founder Mark Leonard stepped down as chief executive for health reasons. The post answers the first with durability and the second with a fact from François Rochon's shareholder letter: Leonard remains on the board.

Constellation buys hundreds of tiny software companies that run one specific industry's operations — golf clubs, funeral homes, transit authorities — and never sells them. The software is dull, cheap relative to the customer's budget, and impossible to rip out without breaking the business, so the revenue keeps arriving.

The signal here is who is buying it. Mohnish Pabrai is a deep-value investor: he normally buys statistically cheap, unloved assets, not high-quality compounders. His adding heavily to Constellation is therefore read as a statement about price rather than style — "Mohnish Pabrai thinks Constellation Software is way too cheap." That is a useful kind of evidence: an investor stepping outside his usual habitat is telling you something about valuation, not about taste.

This is the second Constellation purchase in three weeks, and the argument has shifted from the multiple to the yield. Turn the valuation upside down: instead of asking how many times cash flow the shares cost, ask what percentage of your purchase price arrives as cash each year. That figure is now 6.8% — higher than it has ever been in the company's history — and on 2027 estimates it reaches 9.7%.

There is a second layer. Constellation deliberately holds its reported cash flow down by ploughing money into buying more software businesses. The cash it could report if it stopped growing is higher than the cash it does report, so the true yield to an owner is better than 6.8%. Slegers' word for the sell-off is blunt: "completely unfair."

The other reason for the purchase is housekeeping rather than valuation. Constellation had drifted down to 3.8% of the portfolio — one of the two smallest positions in an eighteen-name book — despite being one of the seven highest-conviction holdings. "It deserves a higher weight, especially at these valuation levels." He added $25,000 at a CAD 2,550 limit, on top of the $15,000 added on 1 February at CAD 2,600.

Constellation owns thousands of small software businesses, each serving one narrow industry — hospital departments, fire services, bus operators, courts. None is famous; all are hard to remove once installed. The whole software sector has fallen more than 25% in six months on the belief that AI will make software easy to replace, and Constellation fell with it, down to about 16 times the cash it generates.

Slegers gives three reasons the fear is misplaced, and they are worth separating. Software for healthcare and public safety has to be right every time, and current AI is not reliable in that way. The cost and risk of ripping out software wired into how an organisation works are what actually keep customers, not the difficulty of writing code. And AI makes Constellation's own developers faster. He then borrows the most inconvenient possible witness for the bear case — Nvidia's Jensen Huang, who calls the "AI replaces software" idea illogical, on the grounds that AI is being built to operate existing software rather than replace it. The final structural point is the arithmetic of a roll-up: it is one thing to disrupt a software company, another to disrupt thousands of them, each customised to a different niche.

Constellation owns more than a thousand small software companies, each one selling a program that a single narrow industry depends on — scheduling for London buses, management software for Florida golf courses, library systems in Canada. None of them is exciting; all of them are hard to replace. It buys these businesses cheaply, leaves them running, and uses the cash they throw off to buy more. Since listing it has returned over 20,000%.

What makes this month different is the drawdown. In its entire history the stock had never fallen more than 25% from a high; it is now down more than 50%, because investors fear AI will make niche business software worth nothing, and because the founder, Mark Leonard, has retired. Slegers thinks both fears are overdone: the software is embedded in operations rather than sold on features, and the successor, Mark Miller, has been inside the company since its first acquisition in 1995, owns roughly $700m of stock, and has been buying more in the open market — which is what people do when they think the price is wrong. The row uses the Toronto ticker (CSU.TO); the US over-the-counter line is CNSWF.

Constellation buys small software companies that sell to one specific industry — the software a bus depot uses to plan routes, or a fire department uses to log calls. These businesses are unglamorous and tiny, but their customers can almost never leave: the software is wired into how the organisation actually works. Constellation buys hundreds of them, keeps them running, and reinvests the cash into buying more.

The stock has halved in valuation terms — from 35 times cash flow in 2024 to 16.2 times now — because investors fear AI will make writing software so cheap that anyone can replace these products. Slegers thinks that misreads what customers are paying for. As one commentator he quotes puts it, the moat is not the cost of writing the code; it is the mission-critical data, the regulations the software encodes, and the pain of migrating away from it. If that is right, AI is a tool that makes Constellation's own developers faster and makes acquisitions cheaper. He put $15,000 in at a limit of CAD 2,600.

Constellation buys small software companies whose products run one specific industry — a marina's booking system, a bus operator's scheduling — and never sells them, using the cash they throw off to buy more.

It is currently in the worst share-price fall of its entire life as a public company, because investors fear artificial intelligence will make this kind of software cheap to rebuild. It is also the smallest position in the portfolio, down about 20%, and rated the strongest thing on the list.

Two pieces of evidence support that. The numbers: the shares cost 17.7 times next year's profits against an average of 31.5 over the past five years, roughly half the usual price for the same company. And the behaviour: the chief operating officer, Mark Miller, has just put $3.6 million of his own money into the stock — the sort of purchase that is hard to explain if the people running it believed the disruption story.

Constellation buys hundreds of small software companies that sell essential, unglamorous programs to specific industries, and holds them forever, using their cash to buy more.

What is striking about this entry is that it contains no financial argument at all. The case made is about the person running it. Mark Leonard avoids publicity, makes no promises, and has taken no pay package for three years — set explicitly against Elon Musk at the other end of the same list, whose proposed award would be worth a trillion dollars.

That is not sentimentality. In a business built entirely on buying assets cheaply and never selling them, a chief executive with no incentive to hit a share-price target and no appetite for attention is a structural advantage: nothing pushes him to overpay for a deal or to talk the stock up. The closing table's other numbers — a 6.2% profit margin — look poor and are an accounting artefact, which is why this company is normally valued on cash rather than earnings.

Constellation buys hundreds of small software companies that sell unglamorous but essential programs to narrow industries, and holds them permanently, using their cash to buy more.

The interesting detail here is the date: it was bought on 10 November 2025 and is only 3.45% of the book — the smallest weight, tied with Interparfums — while carrying a STRONG BUY. That is what building into a falling price looks like from the inside: a starter position at the top rating, with room to add, in a name the rest of the archive spends the spring arguing the market has misjudged on AI fear.

Constellation appears here as the template, not the recommendation. It buys small software companies that serve one narrow industry each — the kind of program a marina, a dentist or a bus operator cannot run without — and now owns more than a thousand of them, letting each keep operating on its own. Slegers calls it the best serial acquirer in the world.

Its relevance to the CMG case is threefold: it proves the model works at scale, its former executives are the ones now steering CMG, and its compensation design (cash bonuses that employees must use to buy shares on the open market) is being copied at CMG. He offers no valuation or stance on Constellation itself in this post — it is the yardstick against which the small Canadian imitator is measured.

Constellation buys small software companies that make niche programs for one specific industry — dental practices, marinas, public transit operators — and then leaves them alone to keep running, sending the cash they generate back up to headquarters to buy the next one. That kind of software is called vertical market software (VMS); it is boring, mission-critical, and almost never replaced.

Slegers calls it the best serial acquirer in the world, and the compounding comes from repetition: hundreds of small, cheap, high-return acquisitions rather than one big bet. He concedes the stock is expensive and then dismisses the objection in the spirit of the exercise — "when you own a stock for 50 years, it doesn't matter that much." The row uses the Toronto ticker (CSU.TO); its US over-the-counter line is CNSWF.