Inside the vertical software consolidators: what 116 acquisitions in the last six months reveal
We put 318 AI agents to work on finding every serial acquirer of vertical software, pulling each acquisition announced in the last 6 months and reading what the buyer said about the company it bought and why.
For twenty years vertical software was one of the safest corners of the industry. Its customers change systems once a decade and nobody was going to rebuild the software that runs a dental practice, an auction house or a driving school for a market of a few hundred buyers.
This year that safety is in question. AI makes a niche product cheap to rebuild and the seat-based pricing most of vertical software runs on shrinks when agents do the work the seats used to do. Public markets priced the fear in February, when software stocks sold off on exactly that argument.
The people who buy vertical software for a living kept buying. Constellation Software, Valsoft, Banyan, Visma, TeamSystem, the private equity platforms and a handful of new AI-era holding companies announced deal after deal through the spring and summer and explained themselves in writing every time. We put 318 AI agents to work on those announcements: finding every serial acquirer of software, pulling each acquisition announced between 16 March and 16 September 2026 and reading what the buyer said about the company it bought and why. The result is 116 acquisitions by 67 acquirers, coded one by one against the same questions.
Most of the buying, 65 of the 116 deals, is done by holding companies such as Constellation and Valsoft that acquire software businesses and never sell them, so a typical exit for a niche software company is now a permanent owner rather than a competitor or a fund. Those holding companies are buying at a slower pace in the last 6 months compared to 2025, while the smaller groups that are still assembling their portfolios have sped up. AI is given as a reason for the purchase in about one deal in four, and hardly ever by the holding companies that account for most of the deals.
The buyers fall into five groups. Permanent-capital holding companies buy and never sell: Constellation's groups, Valsoft's groups, Banyan, Everfield, Software Circle. Sector strategics are operating software companies buying inside their own vertical: Visma, TeamSystem, Nemetschek, Cegid, Ideagen, Asseco. Private equity platforms make add-on acquisitions for a fund that will sell in a few years: Aptean, Orisha, Nexus, Alpine SG. The two remaining groups are two getting a lot of attention recently: distressed and legacy buyers such as Marlin and the AI-era holding companies such as Circeus and Anaconda, ten deals between them analysed in the last six months.
What reasoning does each type of buyer give for their acquisitions
If you go through a hundred acquisition announcements in a row, the buyer types separate on their own. The holding companies talk about permanence. 58% of their announcements pitch a long-term home for the company while none of the strategics do. A third of holdco announcements promise that the team stays on, against just 6% for strategics. Banyan's release on HMM, a German eHealth platform, is a good example: the team stays in Moers, the relationships stay intact and the investment goes into the product and the people.
When holdcos describe the company they bought, they praise its position and its people. Nearly half compliment the team or the reputation, a third call the product mission-critical or deeply embedded and the most specific reasons they give are about how hard the target would be to displace. Examples are Biodiv-Wind's "deep integration within its customers' systems" and Aucxis's "specialist solutions deeply embedded in day-to-day operations".
The strategics talk about their own product. 61% cite cross-selling or integration and their reasons come with numbers because they have a roadmap to defend. Visma bought MaisMei, a Brazilian compliance tool with 3.7 million micro-entrepreneur customers, as a "graduation path" into its cloud ERP. Nemetschek explained HCSS with a 75% win rate in US highway contracts and a 40% margin. AI is given as the reason to buy in 23% of strategic announcements and just 9% of holdco ones and where a holdco mentions it, it is a promise about what it will invest in after the deal rather than a reason to buy.
Private equity platforms talk about the customer list. Six of the ten announcements cite distribution or access to the target's customers. Aptean bought ROTOR for its relationships with 1,000 machinery dealerships in German-speaking Europe. Orisha bought Soneka for the 25,000 property assets it manages across 13 countries. Ten deals is a small base, but the pattern matches what a platform is for: a buyer assembling a company to sell in four years needs revenue it can put on a slide.
6 out of 10 deals in our analysed sample went to buyers who never sell
The permanent-capital holding companies did 65 of the 116 acquisitions we analysed in our sample. The strategics did 31, the PE platforms 10. The two groups that get the most attention when people talk about software M&A, the distressed specialists and the AI-era holdcos, did four and six.
For anyone who owns a niche software company, selling to such a holdco often has consequences for price, because a buyer who never sells does not necessarily need a story about the exit or price a potential strategic fit. However, these groups run acquisitions as a production line and move faster than anyone else.
Within the holdcos, two families dominate. Constellation's operating groups did 37 deals, with Volaris alone at 11 and Vela at 7 and Valsoft's groups did 12. Together that is 42% of every software acquisition we recorded. The names on the announcements are Volaris, Jonas, Vela, Harris, Perseus, Vesta, TAG, Lighthouse, Edelweiss and Aspire, which is why founders do not always realise how concentrated the buying is. Two Canadian balance sheets sit behind roughly two in five deals.
The biggest buyers are slower than last year, but the smaller ones are faster
We checked each active acquirer's stated 2025 count against its pace this year. The largest ones have slowed. Volaris completed 30 acquisitions in 2025 and 11 in our six months, about 70% of last year's monthly pace. Visma did close to 30 in 2025 and 7 this year, half its pace. Jonas went from 20 to 5. Constellation as a whole deployed 1.7 billion dollars on acquisitions in the first half of 2026, nearly three times the same period in 2025, while its groups bought fewer companies. The money went into fewer, larger deals.
There is however an acceleration one tier down. Vela doubled its pace, buying seven companies against six in all of 2025, most of them through its Juniper travel group. TeamSystem in Italy matched a full year in six months. Software Circle in the UK doubled from a small base. Banyan and Everfield held their 2025 pace. The groups still building a large portfolio are buying and the groups that already have one are pickier.
Part of the gap is the summer. Announcements per month ran between 15 and 23 from April to June and dropped to between 9 and 14 from July, which is the usual pattern and also reflects deals not yet announced. But a full-year 2025 figure already contains a summer, so seasonality explains the July dip, not the year-on-year gap. Our best reading is that the largest buyers really are slower than in 2025, digesting a record year and pricing AI uncertainty into what a niche is worth.
How does the split look like in terms of locations?
Of the deals with a stated target country, 28 were in the United States and 14 in the United Kingdom. France had 9, Germany 6, and Italy, the Netherlands, Australia, Canada and Brazil 4 each. Europe accounts for 45 of the 96 deals with a country, so almost half.
The more useful number is how far the buyers travel. 80% of holdco acquisitions were cross-border, with the acquirer headquartered in a different country from the target. For strategics the figure is 65%. The Canadian groups in particular buy everywhere and nowhere in particular: a Hungarian insurance software firm, a Croatian public sector ERP, a Spanish payroll vendor, a Swiss learning platform, a Brazilian communications company.
What is getting bought and what each sector's buyers talk about
Software for the finance function (Office of the CFO) is the most acquired category, with 18 deals, followed by healthcare with 16, HR and payroll with 10, then marketing and compliance and legal with 9 each. Below that the sample gets thin and 17 deals could not be classified from the announcement. Within the big sectors, the announcements read very differently and the differences are worth knowing if you sit in one of them.
Finance and accounting software is the one large sector where strategics outnumber holdcos, 10 deals to 7 and half of all announcements cite geographic expansion, the highest share of any sector. The reason is that tax and invoicing rules are national, so accounting software is bought country by country: Visma's Italian e-invoicing and Brazilian compliance deals, TeamSystem's entries into Turkey and France, Cegid's French payroll acquisition. If you sell finance software, your most likely buyer is a strategic that wants your country. AI appears in four of the 18 announcements, about the sample average, so it is not what makes a finance software company attractive to these buyers as much as the country.
Healthcare is the opposite. Ten of its 16 buyers were holdcos, 44% of announcements pitched permanence, 44% called the product mission-critical and 38% quoted a customer count, from 430,000 patients on a correctional health platform to 3,000 clients for a Polish pharmacy chain system. Three of the eight healthcare deals with a known seller were private equity exits, which makes healthcare the clearest example in our sample of PE selling to permanent capital. A fund builds the platform and the holdco takes it off their hands.
HR and payroll is where AI is the reason to buy. Four of the ten announcements mention it, the highest share of the large sectors and it is also the only large sector where the distressed and legacy buyers appear, with two of the ten deals. That combination says something about the sector: the newest HR products are being bought for their AI and the oldest are being bought for their cash.
Compliance and legal software seems like more of a closed club. Not one of the nine deals was a buyer entering the sector for the first time, 89% were described as deepening an existing position, Ideagen and Volaris between them did most of them and every seller whose identity we know was a founder. If you own a compliance product you will sell to someone already in compliance and they will describe your product as the piece they were missing.
On the other hand, construction and property is where new shelves are opening. Half of its six announcements were buyers entering the vertical for the first time, the highest share of any sector and half quoted customer numbers. The holdcos opened several other shelves this year that nobody would have predicted: driver training in Germany, karting venue management, salary packaging in Australia, bird and bat detection for wind farms, auction software in the Netherlands, financial broker software in Ireland. Each was described as the leader in a niche too small for anyone else to build a business around. Manufacturing software, with only four deals, had the most consistent announcements in the sample: every one praised the technology and three of four promised the team would stay.
AI is in the announcement one time in four and rarely from the buyers doing most of the buying
Across all 116 deals, 30 announcements mention AI somewhere. The AI-era holdcos mention it every time, the strategics in 29% of their deals, the holding companies in 17%. Roper's purchase of CentralReach, a clinical platform Roper described as "first out of the gates in terms of AI adoption", is a rare case of a permanent-capital buyer naming AI as the reason.
The holdcos know what AI is doing to software. Their model does not depend on it. A buyer who prices a niche software company on its cash flow and holds it forever does not need the product to win the AI race, it needs the customers to stay. The strategics, who have to defend a roadmap, are the ones who must buy AI capability or explain why they did not. For founders who have spent the last few years building AI features, its more likely that this is the audience that will pay for them.
Which of them buy small?
From the acquirers we looked at, Constellation looks at businesses from 1 million and above, Banyan buys from 2 to 30 million, Alpine SG from 1 to 50 million of ARR, the micro acquirers such as saas.group and SureSwift from 1 to 10 million. Lumine is the only group in our sample whose floor, around 20 million dollars, excludes a typical mid-sized European software company.
Where vertical software goes from here
The floor under vertical software valuations is permanent capital. Holding companies that never sell took 65 of 116 deals and Constellation alone put 1.7 billion dollars to work in the first half, nearly three times its 2025 pace in money if not in deal count. As long as those balance sheets treat a niche software company as a cash flow to be owned for decades, the public market's fear about seat pricing sets the discount, not the existence of a buyer.
The buying moves across borders and into smaller niches. 80% of holdco acquisitions crossed a border, and the verticals opened this year, driver training, karting venues, salary packaging, bird detection for wind farms, are markets of a few hundred customers each. The buyers have run out of large niches and are working through the small ones, which is where most European software companies sit.
If the trend from the last 6 months sustains, the names doing the buying (especially for smaller companies) will be changing. The largest operating groups, Volaris, Jonas, Visma, bought fewer companies than in 2025 and spent more per deal, while Vela, TeamSystem, Software Circle and Everfield doubled or held their pace. The next two years of vertical software consolidation could be run less from Toronto and Oslo and more from the second-tier compounders that are still building a large portfolio.
Subscribe to the newsletter so that new editions land directly in your inbox.
And if you are thinking about selling a vertical software company, I run Eilla, an AI-native M&A advisory. Message me.
