Software, Data & AI

In brief: Profitable software companies in the German SME market are currently valued at 6.1-10.6x EBITDA, small caps in the 7.5-9.5x band (based on DUB KMU-Multiples Q3/2026, analysis SourcingClub). Buyers are software serial acquirers such as Main Capital and Volaris, PE platforms and international strategics. TMT accounted for 34% of all DACH PE deals in 2025 (PwC 2026). In my experience, a structured sale takes 6 to 9 months.

Market overview: software, data & AI in the DACH region

Software is the most active M&A sector in the region. In 2025, 591 software and IT transactions were announced in the DACH region, after 578 the year before, with a disclosed volume of €33.9bn (M&A Review, January 2026). The median deal value was €37.8m (M&A Review, January 2026). Most transactions therefore take place in the lower mid-market, not in the billion-euro deals that make the headlines.

On the private equity side the picture is even clearer. In 2025, TMT accounted for 34% of all DACH PE transactions and 26% of the volume. In total, the region recorded 557 PE deals with a volume of €88.3bn (PwC Private Equity Trend Report 2026). This makes software the most acquired sector in DACH private equity.

To be clear on scope: this page covers the product and data business, meaning product software and vertical SaaS, data engineering and analytics service providers, and AI implementation and AI products. IT operations, managed services and IT system houses have their own page, IT services. Data and AI service providers with project business belong here, because buyers value them as part of the data and AI market. Industry software for medical practices, clinics and care services also touches the market covered under Healthcare & medical technology.

The vendor landscape is highly fragmented. Alongside a few large firms there are thousands of owner-managed software companies, many of which have grown in an industry niche over two decades. Many come from the founder generation of the 1990s and 2000s, whose owners are now reaching handover age. Around 109,000 Mittelstand companies a year are aiming for a succession up to 2029 (KfW Nachfolge-Monitoring Mittelstand, 2025). There are no succession statistics specific to software, but the age cohort of software founders points to a growing number of handovers.

The market environment in 2026 is split. The median multiple of listed SaaS companies fell from 5.6x ARR in December 2025 to 3.2x in June 2026, discussed in the market as the "SaaSpocalypse" (SaaS Capital Index, 06/2026). Buyers use this anchor in negotiations. PE software dealmaking slowed in early 2026, because AI risks and weaker revenue growth pushed buyers' and owners' price expectations apart (S&P Global Market Intelligence, 02/2026), and DACH mid-market PE was slower overall in the first half (Roland Berger, H1/2026). The software serial acquirers kept buying all the same, and German SME multiples are more stable than the public market bands (based on DUB Q3/2026).

Who is buying right now

The buyer side ranges from serial acquirers to PE platforms and international strategics. Publicly documented examples:

• Main Capital Partners: number 1 among platform acquirers in the DACH software market (AGC ranking). Tracxn counts 84 direct platform acquisitions. Most recently, SMS Stiewi joined the MACH group as its sixth add-on (Main Capital, 04/2026). Focus: enterprise software with buy-and-build.
• Volaris Group (Constellation Software): 14 acquisitions in the first half of 2026 alone (Volaris Mid-Year Update, 07/2026). Buys profitable niche software vendors with a permanent hold model. Brand, team and location stay, with no resale.
• Hg: invested in Diamant Software in Bielefeld in November 2025 (Tracxn, 06/2026), alongside the $6.4bn take-private of OneStream (Marketscreener, 04/2026). Hg portfolio company Septeo acquired the Karlsruhe legal tech provider stp.one for an estimated €560m (PrivSource, 11/2025).
• Bregal Unternehmerkapital: building a European platform for accounting and ERP software, with majority acquisitions of Uniconta and SnelStart in October 2025. Around €7bn in assets under management (VC Magazin, 10/2025).
• LEA Partners: Karlsruhe-based software investor whose platform NEXOWARE brings together software for managed service providers (LEA Partners, press release 2025).
• KKR: acquired the IT service provider Datagroup for €598m as a buy-and-build platform for managed services in the Mittelstand (Handelsblatt, 04/2025).
• Warburg Pincus: take-private of PSI Software for €595m (M&A Review, 01/2026), evidence of continued PE interest in German industrial software.
• Strategics: NICE acquired the Düsseldorf AI agent provider Cognigy for €819m, and Proofpoint acquired the Hanover security provider Hornetsecurity for €893m (both 2025, M&A Review 01/2026).
• Horváth: acquired the Stuttgart AI start-up Studicon and brings all its AI solutions together under the Horváth.AI brand, an early example of consolidation in the data and AI consulting landscape (consulting.de / BDU, 2026).

Family offices and Mittelstand holding companies also buy, usually without a press release. Which buyer type suits you shapes the price, the structure and your role after the sale. The differences are set out in the comparison PE investor vs. strategic buyer.

EBITDA multiples in the sector

The valuation anchor for the German SME market is the DUB KMU-Multiples. Growth SaaS and service providers follow their own logic, so the table shows them separately:

Data basis: DUB KMU-Multiples (dub.de/kmu-multiples), Q3/2026, sector software development & SaaS. Subsegment classification and analysis: SourcingClub. The EBITDA bands smooth out quarterly swings and are therefore not identical to the raw DUB values in every segment (methodology, in German). Full ranges by size class: EBITDA multiples DACH. The rows on ARR valuation, data engineering service providers and AI vendors are not covered by DUB. They rely on the transaction sources named in each case.

The ARR bands for growth SaaS come from global private SaaS transaction data (Livmo, L40, Windsor Drake, Axial, 2026). Realistically, DACH SMEs sit at the lower end of the band. The service provider bands are based on Aventis Advisors (more than 600 IT services deals from 2015 to mid-2025, median 10.2x EV/EBITDA) and CT Acquisitions (2026), and the AI row on FE International, Development Corporate and Valutico (2026). Whether the EBITDA or the ARR logic applies in an individual case depends mainly on growth.

What raises the multiple and what lowers it

Valuation example: a vertical SaaS vendor worked through

A fictitious but typical example for the company size in this sector: a vendor of industry software for skilled trade businesses.

The EBITDA figure already assumes adjustments for a market-rate managing director salary and one-off effects. This bridge from the P&L to adjusted EBITDA is exactly the first point of discussion in a process. The valuation anchor is the small cap band for software development and SaaS based on the DUB KMU-Multiples Q3/2026 (analysis SourcingClub), i.e. 7.5-9.5x EBITDA. With €1.1m adjusted EBITDA, that gives an enterprise value of €8.3-10.5m. The Rule of 40 score of 32, the 80% recurring share and the vertical focus (25-30% premium over horizontal software, Axial 2026) justify the upper half of the band, indicatively around €9.4-10.5m.

A cross-check using the ARR logic supports the result: €9.4-10.5m corresponds to 1.8-2.0x ARR, which is below the global private SaaS bands of 3-5x for growth under 20% (Livmo, L40 2026). For the German SME market, where the EBITDA logic dominates, this discount is the norm. Typical structure: most of the purchase price is paid at closing (cash-free/debt-free with a working capital target), and the rest as an earn-out over 24 months on EBITDA targets (SRS Acquiom 2025) or as a reinvestment in the buyer's HoldCo. The biggest lever in the example would have been customer retention: with 120% NRR instead of 105%, market data suggests 30-50% higher multiples would be achievable (FISART / Walden, 2025/2026).

The sale process in the sector

In my experience, a structured software sale takes 6 to 9 months: preparation and data room, confidential approach to suitable buyers, indicative offers, due diligence, negotiation and closing. Three features set the sector apart from other industries.

First, code and IP due diligence. Buyers review the architecture, samples of the source code, the deployment pipeline, the security history and the legal ownership of the code. Development by freelancers without clean IP assignment clauses is a classic deal killer (Papermark / CT Acquisitions, 2026). Open source deserves particular attention: only 2% of the codebases audited in 2025 contained no open source code, and copyleft licences such as the GPL can contaminate the IP position (Black Duck, 2026). A software composition analysis scan before the process starts takes that argument away from the buyer. The preparation pays directly into the purchase price: technical DD leads to renegotiation in 30-40% of software-heavy deals, with typical reductions of 5-25% for material findings (PitchBook 2025 via CT Acquisitions).

Second, the new AI review layer. Since 2025/2026, buyers systematically check four areas: dependency on third-party models, the strength of the company's own data moat, how far AI agents could replace the core function, and the concentration of AI talent (Valutico, 2026). One in five strategics abandoned a deal within a year because of AI concerns (Bain, 2026). Regulation comes on top: since 2 August 2026, extended obligations under the EU AI Act apply, with fines of up to €35m or 7% of worldwide revenue (M&A Review 01/2026; Valutico 2026). AI vendors are best off putting the compliance documentation into the data room from the start.

Third, the customer analysis. Buyers group customers by start quarter and measure retention over 12-24 months. A blended churn of 10% can look healthy while the most recent cohort churns at 30%, and that is exactly what DD uncovers (FE International, 2026). A source code escrow is also common: a neutral third party holds the source code, and modern agreements also name key people and set out their availability (Escode / PRAXIS, 2025).

Typical deal structures

Earn-outs are common in the sector, but less common than often assumed: in 2025, around 18% of all M&A deals included an earn-out component, down from 26% in 2023 (SRS Acquiom 2025 Deal Terms Study). PE buyers used earn-outs in 28% of platform deals in 2024, mostly EBITDA-based with a median term of 24 months. Strategics used them in 18% of deals, more often revenue- or milestone-based and running around 36 months (SRS Acquiom via Harvard Law School Forum, 07/2025). The key figure for your calculation: across all deals, only 21% of promised earn-out amounts are actually paid out (SRS Acquiom, 2025). In addition, 75% of agreements now explicitly exclude acceleration on a resale (SRS Acquiom via Peony, 2026). An earn-out is therefore upside, not a secure part of the purchase price, and you should be able to influence the metrics yourself.

The alternative in PE platform deals is the rollover, i.e. reinvestment: the founder exchanges part of the purchase price for real equity in the buyer's HoldCo. Unlike an earn-out, this is not a conditional payment but a stake in the second exit, common for founders who keep working in the business (CT Acquisitions / L40, 2026). Both earn-outs and rollovers typically tie founders to the company for 24-36 months (SRS Acquiom via Harvard Law School Forum, 07/2025). If you want neither, the serial acquirer model along the lines of Volaris offers a third option: usually a 100% acquisition with a permanent hold approach. Brand, team and location stay, and there is no pressure to resell (Volaris Mid-Year Update, 07/2026).

The standard mechanics of a software deal remain similar across all buyer types: a cash-free/debt-free purchase price with a working capital target, escrow or holdback for warranty claims, and liability caps (Aventis Advisors, 2026). These levers decide the net proceeds, not just the multiple.

For PE investors: the buy-and-build thesis in brief

The multiple arbitrage in the sector can be read directly from the DUB bands: add-ons in the micro cap segment cost 6.1-7.6x EBITDA, while a mid cap platform trades at 8.3-10.6x (based on DUB KMU-Multiples Q3/2026, analysis SourcingClub). On paper, every integrated add-on is therefore revalued at the platform multiple. How closely this playbook is already being run in the DACH region is shown by Main Capital's MACH group with six add-ons (04/2026) and Bregal Unternehmerkapital's accounting platform (10/2025). Demand is structurally broad: 2,698 SaaS transactions worldwide in 2025, an all-time record with 28% growth, with private equity involved in around 58% (FE International / Livmo, 2026).

The second thesis is only just emerging: the consolidation of data and AI service providers. The German market for data and AI services is growing at double-digit rates and much faster than the IT market overall (Lünendonk, 2026), while the consulting market as a whole grew by only 0.5% in 2025 and, according to BDU, is entering a consolidation phase (BDU, 2026). Acquisitions such as Horváth/Studicon mark the beginning. In the adjacent managed services field, KKR's Datagroup platform (€598m, Handelsblatt 04/2025) follows the same logic. Anyone who builds a platform with a high managed services share early can currently still buy add-ons in the 6-12x bands of the lower mid-market (Aventis / CT Acquisitions, 2026).

The bottleneck is the same in both theses: access to owner-managed targets before they enter a process. The mechanics behind this are explained in What is deal origination, and my services for funds on the page for investors.

From my sourcing work

Software is the sector where preparation makes the biggest difference. In practice, three points cost money most often: a net revenue retention figure that nobody can report cleanly, unclear rights to the code, and a business that does not work without the founder. Since 2025 there is also the AI question, and it cuts both ways. Buyers want to know whether a product has proprietary data or real workflow depth, and whether standard models could attack it. If you have settled these points before the first conversation, you negotiate from a different position, whether the sale is one year or five years away.

If you are thinking about a sale in the next one to three years, an early, discreet assessment is worth it: which metrics your company reaches today, which multiple band is realistic and which buyer type fits. A confidential buyer match is a non-binding first step for this, without your name going into a broad process.