Old, unglamorous construction software is getting acquired at multiples nobody else is paying right now. Nemetschek just bought 39-year-old HCSS for roughly 20 times EBITDA. On the podcast, Dustin DeVan called the pattern the “bathtub” theory: the workflow software with the business logic and years of data already baked in is what’s hard to displace, not the AI sitting on top of it. Here’s the case, and the numbers behind it.
Patric Hellermann noticed something odd on the latest Bricks, Bucks & Bytes and said so out loud. Software companies with the better part of a decade, sometimes several decades, on the clock are getting bought this year at 10x to 20x ARR. He couldn’t think of many other verticals paying multiples like that for software right now, vertical or otherwise. His read: call it a bullish signal for construction tech, whatever else is true about the broader market.
Dustin’s answer for why cuts against the AI narrative everyone else is running. Set aside Datagrid, the kind of company built for the AI moment. Look instead at something like HCSS, he said, an acquisition of what people in the industry perceive as “bathtubs.” The phrase is Dustin’s, and it’s doing real work: a bathtub is old, it’s not exciting, and it is very hard to rip out once it’s plumbed into the walls.
The logic runs like this. Construction still leans on workflows that are antiquated compared to more software-forward industries, which sounds like a weakness until you notice what it actually means for an incumbent. If you already own the system that holds the business logic and years of transaction data, you have something a new AI wrapper cannot replicate no matter how good its model is. The bathtub is already installed. The AI is what gets built on top of it.
The Deal That Proves the Point
What Nemetschek actually paid for HCSS
Nemetschek’s move on HCSS is the cleanest data point for the bathtub thesis available right now. Founded in 1986 and based in Sugar Land, Texas, HCSS builds estimating and field management software for the heavy civil and infrastructure sector, the trades that dig and pour rather than the ones that finish and furnish. It serves more than 4,000 companies. Nemetschek is paying roughly $2.4 billion, a figure the company itself calls the largest acquisition in its history, at slightly above 20 times HCSS’s projected 2025 EBITDA. Based on HCSS’s reported 2025 revenue of approximately $215 million, that price also works out to somewhere around 11 times revenue.
The underlying business explains why buyers are willing to pay it. HCSS has been growing its annual recurring revenue at over 20% a year, carries an EBITDA margin near 40%, and churns under 2% of customers annually. That is not a startup profile. That is four decades of entrenchment showing up as a balance sheet.
Why the Boring Software Wins
What actually makes a workflow hard to leave
None of this is really an argument against AI. It’s an argument about where AI creates value versus where it merely borrows it. A model that reads drawings well is legitimately useful, but a model with no connection to a contractor’s actual estimating history, supplier pricing, and job costing data is starting from zero every time. The company that already has that data, accumulated over a customer’s entire relationship with the software, has a running start no wrapper can match on day one.
That’s the acquisition logic showing up across this cycle. Buyers aren’t paying 20x EBITDA for a feature set. They’re paying for the switching cost, the labeled data, and the trust that took decades to build with an industry that, as more than one guest on the podcast has pointed out, does not extend trust easily. Once a piece of software becomes the system of record for how a contractor bids and gets paid, replacing it means retraining every estimator on the team and re-earning years of institutional memory. Few buyers want to take that risk on an unproven newcomer, however good the demo looks.
Deal math, decoded weekly
We break down the acquisitions everyone’s talking about and explain the numbers nobody else bothers to check.
Join 3000+ ReadersThis Isn’t a One-Deal Story
The multiple shows up across the cycle, not just once
HCSS is the clearest example, but it’s not isolated. Procore’s own $845 million purchase of DroneDeploy landed at a double-digit revenue multiple on a company with a much shorter operating history and no comparable moat, which is a different bet: paying up for reach and data breadth rather than entrenchment. Autodesk’s roughly $3.6 billion acquisition of MaintainX, its largest ever, pushes the same logic into operations and maintenance data. Read together, the message to founders is blunt. If you’re building something truly new, the AI is the pitch. If you’re trying to get acquired at a premium multiple, the data moat and the switching cost are what buyers are actually pricing in.
| Company | Years operating | Acquirer | Approx. multiple |
|---|---|---|---|
| HCSS | 39 | Nemetschek | ~20x EBITDA (~11x revenue) |
| DroneDeploy | ~13 | Procore | Double-digit revenue multiple |
| MaintainX | ~9 | Autodesk | Undisclosed, deal value $3.6B |
| PlanGrid | ~6 | Autodesk | Undisclosed, deal value $875M |
It’s Dustin DeVan’s framing, floated on the Bricks, Bucks & Bytes podcast: old, unglamorous workflow software that already holds a contractor’s business logic and years of transaction data is extremely hard to displace, regardless of how good a newer AI tool looks in a demo. The AI gets layered on top of the bathtub; it rarely replaces it. (Source)
Nemetschek agreed to acquire HCSS from Thoma Bravo at an enterprise value of roughly $2.4 billion, slightly above 20 times HCSS’s projected 2025 EBITDA. Based on HCSS’s reported 2025 revenue of about $215 million, that implies a revenue multiple of roughly 11x. (Source)
HCSS was founded in 1986 and is based in Sugar Land, Texas. It builds estimating, job costing, project management, safety, and fleet management software for heavy civil and infrastructure contractors, serving more than 4,000 companies across North America. (Source)
On the podcast, Patric Hellermann pointed out that decade-plus-old vertical SaaS companies are getting acquired at 10 to 20 times ARR this year, multiples he couldn’t find many other software categories matching currently. He read it as a bullish signal for the category, tied to the entrenchment and data moats these older platforms have built up.
Not necessarily, but it does mean the two categories are being priced on different logic. Entrenched platforms like HCSS are priced on switching cost and proprietary data built up over decades. Newer AI tools are typically priced on growth and technical differentiation, which is a harder case to make durable without a data moat of their own.
Procore’s $845 million purchase of DroneDeploy and Autodesk’s roughly $3.6 billion acquisition of MaintainX, its largest deal ever, both landed within months of the HCSS announcement. Each involves a strategic buyer paying up for a company with an established, sticky customer base rather than a pure feature set. (Source)
Related Articles
Nemetschek’s HCSS Deal Isn’t Really a $2.4B Deal
What a Top ConTech VC Thinks Founders Are Getting Wrong About AI
