The budget the demo comes out of

Three very different buyers moved on the same layer of this industry inside a fortnight, and the layer they bought is the one that decides whether anybody's software actually gets used.

On 20 October, Accenture acquired Avicado, a Miami consultancy of more than 190 people whose daily work is implementing and integrating Procore, Autodesk Construction Cloud, Trimble, Oracle Primavera and Kahua, and whose specialists also sit on the owner's side of the table as owner representatives. On the same day, CBRE agreed to buy Onnec of London for £460 million in cash plus an earnout of up to £300 million, taking the largest real estate services firm in the world further inside the data center stack. Two days later Newforma, a software company rather than a services one, absorbed Lumi AI of Toronto for an undisclosed sum, folding an analytics layer into its own project platform.

So inside one fortnight a systems integrator, a real estate services firm and a software vendor all bought into the same part of the chain. When three buyers with almost nothing else in common pay for the same layer, that layer is worth more than the companies selling software into it have assumed.

Everyone loves the demo, and the deals still do not close

If you sell software into design and construction, you know the shape of this. The demo goes well, the room is engaged, somebody says this is exactly what we have been missing, and then the deal sits. It sits through a quarter, it sits through a budget cycle, and eventually it is quietly reclassified as a bad fit.

The usual diagnosis is sales execution: the deck, the pitch, the rep, the follow-up. Hire somebody more senior, buy a methodology, run the pipeline review weekly. Realistically that diagnosis is wrong in the majority of cases, because the demo did exactly what a demo is supposed to do. It convinced the people who would use the product every day. What it did not do is reach the person who owns the line of money the purchase would have to come out of, and in this industry those are very rarely the same person.

The demo convinces the people who will use the software, but the money sits with someone who was not in the room.

Now the part that makes this urgent rather than permanent: the line of money is moving. At Release AEC in Paris on 20 October, where I was in the room, Gendo presented a survey of nearly 1.000 of its users in which 49% said they are using V-Ray, Enscape or Lumion less, and 19% said they are using external CGI studios less. It is a vendor's own survey of its own users and should be read as such, but the direction matches what capital is paying for. On 21 October Mach9 raised USD 34 million, co-led by First Round Capital and Bessemer Venture Partners, for software that turns laser scans into engineering-ready civil drawings, and the claim it sells on is not speed alone but work delivered at roughly a fifth of the cost of the conventional route.

In both cases the first thing the technology takes is not a salary. It is an invoice from outside the firm. That matters commercially, because it means the budget your software has to come out of is increasingly a services budget rather than a software budget, and services budgets sit with different people, get defended with different arguments, and are released against a different kind of evidence.

The large players are already repricing

Watch what the vendors with the most to lose are doing. In Paris, Motif disclosed its licensing model from the stage: viewing is free for the whole firm, floating seats are consumed by a day's activity, and tokens are passed through at cost. Simply put, that is a vendor refusing to charge for the people who only look.

Autodesk went further in public. Speaking to AEC Magazine in October, it confirmed that "Forma will replace Revit over time", and indicated that for the expensive AI features there will probably be a published rate table so a customer knows what a button costs before pressing it, with routine AI staying inside the subscription. The same reporting makes the uncomfortable observation out loud: if agents take over the bulk modeling and documentation, a firm needs fewer seats to do the same volume of work.

Overall, any company whose revenue still rests on counting the people who open the software is selling against a definition of value that its own market is in the middle of abandoning. And the route into the budget that replaces it is precisely what changed hands in October. Avicado's business is implementing other people's software for owners, which means it sits between a vendor and the decision to standardize on that vendor. When that position belongs to a firm with its own commercial interests, a channel that felt neutral last year is not neutral any more.

So if good demos keep ending in stalled deals, three questions are worth more than another sales hire. First, which budget line does this purchase come out of, and have you ever met the person who owns it? Second, is the price expressed in a unit that person recognizes, which is rarely a seat and is usually a project, a job, or an invoice they no longer have to pay? Third, who owns the route into that budget, and what happens to you if that route is acquired next year? Definitely the third is the hard one, because it is the only one of the three that cannot be fixed after the fact. It should be written down as a named strategic dependency with a stated fallback, rather than carried as a relationship everyone assumes will hold.

The scan

Axion PM publishes its price. A working general contractor has launched a construction project management platform with one flat price per company on the website, unlimited users, and no share of the customer's construction volume. Construction Dive. Worth noting because published pricing is itself a positioning claim in a market where the standard is a private quote tied to the value of the work you win.

Mach9 raises USD 34 million for AI drafting. Co-led by First Round Capital and Bessemer Venture Partners, for software that drafts civil engineering drawings from survey scans, with nearly 100 beta customers and more than 2.500 projects behind it. TechCrunch. The pitch leads on cost rather than on capability, which tells you which budget the money is expected to come from.

The European rental market splits in two. The ERA 2026 Market Report puts European equipment rental at a compound rate of 1,8% to 2027, with France at negative 0,7% while Spain, Portugal and Italy grow on infrastructure, logistics and data centers. Construction Briefing. Rental lags construction activity rather than leading it, so read this as a map of where budgets will be defended hardest next year.