Kuro has raised €10 million in a round led by UVC Partners, with Foundamental, Realyze Ventures and helloworld participating. UVC announced the round on 30 September.
UVC’s announcement does not attach a letter to the round. NEON, the law firm that advised UVC on it, files it as a €10 million Series Seed, and that is the only formal label on the record we found.
The Berlin company was founded in November 2024 by Lea Scherer, its chief executive, and Jonas Minkler, its chief product officer. UVC says both previously held leadership roles at the construction-technology companies alcemy and Capmo. Kuro employs sixteen people.
The Number That Defines the Strategy
Buried in UVC’s bullet points is the figure that makes this round unusual: six of Kuro’s twelve existing customers are among the twenty largest general contractors in the German-speaking region. UVC’s “About Kuro” section names six — Goldbeck, Hochtief Infrastructure, Köster, Implenia, Geiger and Riedel — and says the largest contractors in Germany, Austria and Switzerland “have rolled out Kuro to hundreds of users each”.
Twelve customers is a tiny number. Half of them being tier-one contractors is not.
That ratio is the whole go-to-market thesis, and it runs against the grain of the usual construction AI seed round. The conventional path into this industry is bottom-up: find the firms that are underserved by enterprise software, sell cheaply, land thousands of them, move upmarket later. Conmeet raised €6 million in August to do precisely that in the same DACH market, explicitly targeting firms with ten to 500 employees — the segment, as its pitch has it, too big for spreadsheets and too small for SAP.
Kuro appears to have gone the other way and started at the top, with the organisations that have procurement departments, security reviews and incumbent software. The upside of landing a Goldbeck or an Implenia first is that the reference carries weight with every other tier-one contractor in the region, and that the contract values are large enough to fund a sixteen-person company. The cost is concentration: at twelve accounts, each renewal is a material event, and the sales cycle that won them is not a cycle you can run at volume.
What the Product Actually Does
The stated problem is not that construction is behind on software. It is narrower and more credible than that: the documents contradict each other, and the contradiction is the contractor’s liability.
UVC’s description is worth quoting at length because it is specific in a way these announcements usually are not. A project’s data “quickly adds up to hundreds of thousands of pages” and teams in “sales, estimating, procurement and site management spend around half their time sifting through that information and making sense of it.” Then: “the project’s documents regularly contradict each other. If a contradiction goes unnoticed, the general contractor bears the cost, which can run into millions depending on the project.”
That is the commercial mechanism. In a lump-sum or design-build arrangement, a discrepancy between a drawing and a bill of quantities that nobody catches before the price is submitted becomes the contractor’s problem after it is. The value of finding it is not time saved; it is a loss not taken.
Kuro’s platform, per UVC, ingests those documents and links them into a single data model, with the contractor’s own rulebook layered on top — “its building components, review criteria and work packages.” It then “finds contradictions between drawings and building specifications, analyses what needs to be built and where the risks lie, enriches components in 3D models, processes public tenders and evaluates subcontractor bids.”
The Rulebook Is the Interesting Part
One sentence in UVC’s post is doing more work than the rest of the product description: “Corrections made by estimators stay in the rulebook and apply to the next project.”
That is a per-customer learning loop, and it is a different asset from a model. A general-purpose document-reading system improves for everyone when the vendor improves it. A rulebook that accumulates one contractor’s own corrections improves only for that contractor — and becomes progressively more expensive for that contractor to abandon. Kuro is, in effect, selling a place to put institutional judgement that currently lives in the heads of senior estimators.
Minkler frames the expansion from there as following the same data: “We started in estimation because that’s where a project is born and decided. But the problems carry through into everything that follows: too much contradictory data and knowledge that isn’t passed on from project to project.” UVC’s Dr Ingo Potthof, a managing partner at the firm, describes the roadmap as extending “to other back-office processes such as procurement, accounting, and compliance.”
The demographic case for all of this is the one Scherer makes, and it is the strongest line in the announcement. UVC cites around 19,800 workers in Germany’s main construction trade retiring in 2024 against roughly €1.6 trillion of construction investment by EU member states in 2025 — both figures UVC’s, not ours. Scherer’s conclusion: “An industry that forgets faster than it learns cannot deliver on these contracts.”
The Numbers, As the Company Gives Them
Kuro’s reported results come from the company and its lead investor, and are presented here as such. Customers report savings of around 90 hours per bid, UVC says, and 70 per cent of users say Kuro found information they would otherwise have missed. Kuro’s own site says the platform scalably processes more than one million pages a month and puts a Köster case study at six hours saved per week per employee.
The 70 per cent figure is the more informative of the two. Hours saved is a productivity claim, and productivity claims in estimating software are easy to make and hard to audit. Finding something the team would otherwise have missed is a claim about errors avoided, which is the thing the contractor is actually buying. Nils Köster, chief executive of Köster GmbH, says exactly that in the announcement: “We price more bids with the same team, but that’s a side effect. What matters is that we spot risks in the tender quickly, before we submit our price.”
Where This Sits
Reading drawings and specifications for what they really commit you to is now a well-populated category rather than an opening. The founder who sold Pype to Autodesk raised $5 million in September to attack the same problem on the American side, and Document Crunch’s agentic contract review works the contract layer above it. European construction AI has had an active autumn in adjacent territory too — syte’s €9 million for German land-development analytics, announced 17 September, and Rayon’s €10 million for browser CAD, announced 23 September, both landed in the second half of the month.
What distinguishes Kuro in that company is not the technology claim, which is roughly the category standard. It is the customer list. Six tier-one DACH contractors, which UVC says have rolled the product out to hundreds of users each, is a harder thing to assemble in twenty-two months than a model that reads a drawing — and it is the part a competitor cannot replicate by shipping.
The open question is the obvious one. A company that wins the top of a market first has to find out whether the rest of the market is reachable at all, or whether the product it built for Hochtief Infrastructure is simply the wrong product for everyone smaller. Kuro has sixteen people and €10 million to answer it.
