The quarter the platform paradigm reasserted itself

The two largest cheques written in built environment software this quarter did not come from venture capital. They came from the incumbents. Nemetschek agreed to buy HCSS for roughly 2,4 billion dollars in April. Autodesk agreed to buy MaintainX for 3,6 billion dollars in May. Both were the largest acquisitions those companies had ever made, and both landed inside ten weeks of each other. Read together, they are not two deals. They are a statement about who sets the price in this market, and what that price now rewards.

The venture story of the last two years was that AI-native startups would take the built environment workflow from the suites. The second quarter of 2026 was the quarter the suites answered. They did not answer with press releases about their own AI roadmaps. They answered with capital, at a scale that reset what the top of the market looks like, and they moved directly into the wedges the startups had been selling into. The consequence for anyone building or backing a company in this sector is not abstract. It changes what your company is worth to the only buyers large enough to pay a premium for it.

What changed: the capital concentrated

Set the quarter against the backdrop of the year that preceded it. Across 2025, ConTech investment reached a record 6,57 billion dollars over 337 transactions, per Cemex Ventures, a figure that surpassed the 2021 and 2022 peak even as the broader venture market declined for a third consecutive year. The composition matters more than the total. AI-enabled solutions took 77 percent of that capital. AI is no longer a category inside ConTech. It is the operating layer across design, construction, and operations, and a company without it is now the exception that has to explain itself.

The capital did not just concentrate by theme. It concentrated by company. The clearest read of the funding environment this quarter came from the investors themselves: the absolute volume of capital available sits at roughly 2021 levels, but it is being deployed across an order of magnitude fewer companies. The same dollars, a tenth of the recipients. That is not a market cooling. It is a market sorting, and the sort is brutal in the middle. The 250 to 500 million dollar exit band, once the reliable path for a strong single-product company, is now orphaned: too large for a tuck-in acquirer to swallow cheaply, too small to stand as a platform in its own right.

• Enhanced Productivity64%
• Other AI-enabled13%
• Non-AI23%

AI is no longer a category inside ConTech. It is 77% of the capital, and productivity took the bulk of it.

Source: Cemex Ventures, Top 50 ConTech Startups 2026

Two figures from the public and private markets show how sharply the sort now runs. In the private market, companies carrying a credible AI story command materially larger rounds and richer valuations than their peers, on the order of two to three times the round size and pre-money valuation. In the public market, over the six months to February 2026, software multiples compressed hard, with horizontal software indices down around 40 percent and vertical software down around 45 percent, back to levels last seen in the uncertainty of 2014 and 2016. The same technology story that inflates a private round is deflating a public multiple. Anyone planning an exit is now navigating between those two currents.

Private AI pre-money

Private AI round size

Public horizontal SW

Public vertical SW

The same AI story inflates a private round and deflates a public multiple. Above 1.0x, private AI carries a 2-to-3x premium; below it, public software multiples fell toward half their prior level.

Source: Houlihan Lokey, PropTech and software market updates, 1H 2026

The pattern is not confined to construction. In the adjacent PropTech market, US companies raised around 5,8 billion dollars across 2025, up roughly a third year on year, against 109 M&A transactions, up around a fifth. The same reading holds across the built environment: capital is available, it is concentrating, and the realistic exit is a strategic acquisition rather than a public listing. A founder in property software faces the same repricing a founder in construction software does, from the same class of buyer.

What happened: the deals

The two incumbent acquisitions are the spine of the quarter, and each carries a distinct strategic message.

Nemetschek's purchase of HCSS, at roughly twenty times 2025 earnings, moved the company decisively toward the delivery side of construction and into North American heavy civil. A platform whose narrative had been design-led repositioned itself, in one deal, as delivery-led. Autodesk's purchase of MaintainX, at roughly twenty-five times ARR for a business around 135 million dollars in ARR growing above 50 percent, did something structurally similar in a different direction: it extended Autodesk from Design and Make into Operate, the lifecycle stage where asset-data and reality-capture plays had been building from the bottom up. One incumbent bought its way down the delivery chain. The other bought its way into the operational life of the asset after handover. Both were paying for position in a stage they did not previously own.

The venture capital, meanwhile, kept flowing to the AI-native early stage, and it concentrated in two wedges. Preconstruction drew the clearest conviction: LightTable closed a 22 million dollar Series A for AI error detection, on the same thesis a cluster of peers on both sides of the Atlantic is selling. Design generation drew the largest single cheque: Endra raised a 50 million dollar Series A from a16z for AI that generates code-compliant systems straight from a model, marking the shift from AI that assists a designer to AI that generates the design. Around them, a steady run of smaller rounds, an AI construction operating system out of Italy at 12 million euros, a procurement platform moving into AI-assisted estimating, a set of European design-tooling seeds, confirmed the pattern the funding data already showed: preconstruction and field copilots are the two cheques investors are happiest writing.

The competitive board also shifted underneath all of this. Hexagon confirmed the spin-off of Octave, creating a focused Nordic AEC and asset-software pure-play with around 1,4 billion euros of revenue, and sharpening a competitive set that now reads Octave against Nemetschek against Autodesk against Trimble. The suites are not only buying. They are reorganising into cleaner, more focused competitors.

Why it changed: the incumbents moved into the wedge

The most important thing the incumbents did this quarter was not spend money. It was choose where to spend it. They moved directly into the AI-native wedges the startups had treated as open ground.

Bluebeam, inside Nemetschek, acquired mbue and launched Bluebeam Max in June, carrying its AI from PDF markup into preconstruction document workflows and bundling automation and AI-powered review into the subscription tier. That is the preconstruction wedge, the same one LightTable and its peers were funded to own, now contested by an incumbent with millions of installed seats. Procore, at its June summit, launched agentic APIs and a managed marketplace, repositioning from the system of record to the platform of AI agents and making its managed interface the moat layer. For every company in or adjacent to the field-ops stack, that reframes the build-versus-integrate question overnight: ride Procore as a distribution rail and cede the customer relationship, or harden the direct motion before an agentic interface erodes the brand.

This is the deal rationale worth reading carefully, because the same logic runs from the top of the market down to the mid-market, across construction and property software, and across geographies. It is visible in the largest deals of the quarter, and it repeats in the European mid-market, where incumbents in both ConTech and PropTech are absorbing cloud-native architecture to leapfrog their own modernisation rather than build it, from construction financials to property management platforms buying their way to a modern operational-data layer. The logic is consistent wherever it appears. An incumbent with distribution buys the operational-data layer or the AI-native workflow it lacks, and in doing so raises what a customer is willing to pay for the combined suite while lowering what it must spend to modernise. The startup that sold into that wedge as open ground now competes against its own acquirer's distribution.

The startup that sold into that wedge as open ground now competes against its own acquirer's distribution.

What it means for you

Here is the read that matters if you are building software for the built environment, across ConTech, AEC SaaS, and PropTech, or backing someone who is.

The strategic buyers now set the top of this market, and they are paying for one thing above all others: a defensible operational-data layer that compounds inside their own platform thesis. Scalable software with recurring revenue, deep workflow integration, and a proprietary data footprint holds a platform multiple. Everything else reprices toward tuck-in economics, no matter how strong the growth curve on the slide. A feature-strong product growing at 60 percent still gets the meeting. It rarely gets the strategic premium unless its data anchors the buyer's platform.

The trap is subtle, because it feels like progress. Feature velocity is visible, fundable, and satisfying to build. It is also the thing incumbents can now buy or replicate fastest, and the thing the market has decided to price as a tuck-in. Platform gravity, the compounding operational data and the switching cost it creates, is slower to build and harder to show on a chart. It is also the only thing that holds its price when a strategic buyer runs the numbers.

Feature velocity gets meetings. The data layer gets the premium.

Source: Stratly framework

So price it in now. If you are raising in the next two quarters, do not rehearse the round a peer closed in 2024. Rehearse the one a strategic buyer would underwrite today. Show the workflow you sit inside and the data you compound, not the feature list you can extend. A 2027 exit built on feature velocity alone will read as a tuck-in to the strategic buyer even when you still see a platform. The founders who close that gap early, who build the data layer before they need it to defend a valuation, are the ones the reset rewards.