Tendos, a Munich company founded in 2024, announced a €12 million seed round on 27 October led by Insight Partners. The amount is unremarkable for an enterprise-software seed. The customer is not. Tendos does not sell to contractors, architects or owners — the three buyers almost every company this publication covers is chasing. It sells to the manufacturers who supply them.
What Was Announced
The round is seed-stage and led by Insight Partners, which the German coverage identifies as a US investor and which describes itself as supplying “flexible capital across every stage of growth”. Alongside it, per the announcement, are the founders of four unicorns — Personio, Forto, Flix and SumUp — and Stephan Tahy, the former chief executive of the sanitaryware manufacturer Duravit. Trending Topics reports three founders: Christian Sedlmeier as co-founder and chief executive, Matthias Hilscher as chief technology officer and Daniel Kappler as chief product officer. The same report puts Sedlmeier’s background at Celonis and at a company called ivy, where it says Hilscher was also CTO, and Kappler’s at TradeLink and IBM. Those biographies rest on that single outlet; this publication has not independently confirmed them.
Three things the round does not come with, and it is worth being precise about which document is silent. Tendos disclosed no valuation, no individual investment amounts and no prior financing — a point Startbase makes explicitly rather than one inferred from an absence. Tendos has announced no earlier round, and the release names none. Several database pages for a similarly named American healthcare company, “Tendo”, report totals in the tens and hundreds of millions; none of those figures belongs to the Munich company, and the name collision is close enough that it will be mis-cited.
“The building materials industry is one of the largest and least digitized markets in Europe,” Sedlmeier said in the announcement, which is in German. “Its sales teams are highly qualified but are drowning in paperwork and repetitive tasks.” (Translated from the original: “Die Baustoffindustrie ist einer der größten und am wenigsten digitalisierten Märkte Europas.”)
Dealroom converts the round to roughly $13.4 million and says the round “ranks in the top 1% of all-time seed deals for enterprise software startups in Germany” — a ranking it bases on 1,034 comparable rounds. That is Dealroom’s calculation against Dealroom’s own dataset rather than an independently checkable figure, and the kind of ranking that depends entirely on how the comparison set is drawn.
What the Product Actually Does
The company’s site describes Tendos as “eine KI-Plattform, die speziell für Hersteller in der Bauindustrie gebaut ist” — an AI platform built specifically for manufacturers in the construction industry — organised into six modules: tender search, inbox, quotes, orders, integrations and insights. The summary line is blunter: “Findet Ausschreibungen, verarbeitet Anfragen, Angebote und Bestellungen – automatisiert, wo es geht.” Finds tenders, processes inquiries, quotes and orders, automated where it can be.
Mechanically, the agents find relevant tenders, read the incoming documents, match the stated requirements against the manufacturer’s product catalogue, draft the quote and push the resulting order into whatever CRM and ERP the manufacturer already runs. The company is explicit that sales teams keep the decisions that need domain judgement. That division — machine reads and drafts, human commits — is the same one Bobyard drew around submittals, where an estimator reviews each detection before it is locked into the estimate. PataBid’s sub-trade estimating puts a comparable gate in the opposite order: the estimator marks one item on a drawing and the software finds the rest. It is becoming the default architecture for AI in construction commerce, and for a legible reason: a wrong quote on a tender is a priced, signed, enforceable mistake.
Tendos names six customers in the announcement: Wilo, Grundfos, Hansgrohe, Spelsberg, KSB and Kermi — pumps, taps, electrical enclosures and heating. Its website displays a wider wall of manufacturer logos. The company says quotes with a combined value in the double-digit billions of euros have been processed on the platform. That is Tendos’s own figure, reported here as the company states it; the announcement does not break it into customers, periods or win rates, and a volume processed is not revenue, a pipeline, or business won.
Why the Supply Side Is the Interesting Part
Construction-AI capital has gone overwhelmingly to the people who assemble buildings. Scroll this publication’s funding coverage and the pattern is hard to miss: estimating, takeoff, field reporting, scheduling, progress capture, safety — all of it pointed at the contractor and the design team. The materials half of the industry has drawn far less, and in this publication’s coverage what it has drawn has mostly gone to buying rather than selling. BRKZ raised $31 million for Saudi materials procurement and HomeRun raised $12 million for materials quick-commerce in Bengaluru — both marketplaces, both standing between the manufacturer and the site, both optimising the buyer’s experience.
Tendos is on the other side of that transaction, inside the manufacturer, working the seller’s problem. The analytical case for it — and this is this publication’s reading rather than a claim from the company — is that the document flow is unusually well shaped for the technology. A tender is a structured demand for specified products. A catalogue is a structured list of products with specifications. Matching one to the other is a retrieval-and-constraint problem with a correct answer that exists before the machine starts, which is precisely the condition under which language models are least likely to invent something. Compare that to generating a design or sequencing a schedule, where the output is a judgement and the verification is expensive.
Insight Partners’ stated reasoning points the same way. Trending Topics reports its managing director Teddie Wardi as saying that building materials and manufacturing had largely been skipped by the first wave of AI companies. Read as a thesis rather than a courtesy, that is an argument about which part of the desk is unattended. The manufacturer’s sales function is not unserved: ConstructConnect sells building-product manufacturers project leads and specification tracking — Project Intelligence and Specpoint, with a Salesforce integration — and Dodge Construction Network runs a comparable business. What those incumbents sell is the top of the funnel, finding the opportunity. What Tendos says it sells is the bottom of it, producing the quote. On this publication’s reading, that is the gap the round is aimed at.
What to Watch
The concentration question is the one the announcement leaves open. Six named customers across pumps, taps, heating and electrical enclosures is a credible start and a narrow one, and billion-euro manufacturers are slow, procurement-heavy buyers whose contracts are large and few. International expansion beyond Germany, Austria and Switzerland is the stated use of funds, which means the next test is whether tender formats, product-classification conventions and procurement law travel — or whether each market is a fresh integration problem wearing the same product name.
The second is scope creep in a useful direction. The company says it intends to add use cases across the manufacturer’s sales process and further along the building-materials value chain. That is the right ambition from inside the ERP, and it is also how a focused product becomes a broad one that competes with the ERP vendors themselves.
