New York · AEC Technology
Acelab | Material Hub
The AI brain for architecture's $2.4 trillion materials problem - turning a scavenger hunt through PDFs into a single search box that actually understands what architects mean.
The AI brain for architecture's $2.4 trillion materials problem - turning a scavenger hunt through PDFs into a single search box that actually understands what architects mean.
It is a Tuesday. A project architect needs an exterior cladding that survives a coastal climate, hits a fire rating, carries the right certifications, and still looks like the renderings promised the client. The old way: a dozen manufacturer sites, three rep emails, a PDF graveyard, and a spreadsheet nobody trusts. The Acelab way: she types what she wants in plain English and the shortlist appears - ranked by aesthetics, performance, and sustainability, all at once.
That is the entire pitch, and it is deceptively boring until you realize how much of a building gets decided in exactly these moments. Acelab is the company quietly rewiring them. Founded by architects who got tired of doing this by hand, it now runs material research for more than 20,000 firms - including over half of the world's top 100 practices.
Every building is, at bottom, a stack of material choices - what goes on the roof, what wraps the walls, what the floor is made of and how it was tested. There are hundreds of thousands of products, thousands of brands, dozens of building codes, and a small mountain of certifications. None of it lived in one place. The knowledge that did exist tended to walk out the door whenever a senior architect retired or changed firms.
The construction industry is worth roughly $2.4 trillion in materials alone, and for decades the people specifying it worked from memory, manufacturer reps, and PDFs that may or may not be current. It was, depending on your temperament, either a tragedy or a market.
In 2019, a group of architecture graduates out of Harvard's GSD and MIT made a slightly unusual career move: instead of designing more buildings, they decided to fix the thing that made designing buildings tedious. CEO and co-founder Vardhan Mehta had worked on U.S. embassy and institutional projects - the kind of work where getting a material wrong is expensive in every sense. He teamed up with co-founders including Dries Carmeliet, Andrew Setiawan, and Carly McQueen.
The bet was that the people best positioned to digitize material expertise were the ones who had suffered without it. It helped that they brought in AEC-software pedigree: Dave Lemont, the former CEO of Revit, signed on as Executive Chairman - a useful person to know when your product needs to live inside Revit.
"Designing one of the most tedious, time-consuming parts of the profession out of existence."
- THE FOUNDING PREMISE"Becoming the basis of design - for architects, free; for manufacturers, the goal."
- THE BUSINESS MODEL, IN ONE LINEMaterial Hub does three things that, taken together, are harder than they sound. First, it understands architect language - you ask for a material the way you would describe it to a colleague, and the AI translates that into the right query across a database of more than 250,000 products and 10,000 brands. Second, it ranks results across aesthetics, performance, and sustainability simultaneously, tracking 200+ performance metrics, 2,000+ certifications, and 30+ building and energy codes.
Third - and this is the part architects didn't know they needed - it keeps that one selection synced across every document. Pick a product once and it flows through the specs, the submittals, the RFIs, and into the Revit model itself. The company's Firm Library quietly captures each decision so a practice's accumulated taste stops being tribal knowledge and becomes searchable.
Yes, and the list is the convincing part. More than 20,000 architecture, design, and owner firms run on Material Hub, including Gensler, AECOM, Stantec, and CannonDesign - over half of the top 100 firms on the planet. For a category as conservative as architectural specification, that is not a soft adoption curve.
Bars compare disclosed figures: 100,000+ materials at the March 2025 launch grew past 250,000; the $3.5M seed grew into a $13.5M Series A. The 50%+ figure is Acelab's own, and the kind of number a conservative industry doesn't hand out lightly.
"It speaks architect. That alone separates it from the usual product portals."
- THE CASE, SUMMARIZED"Free for firms, paid by manufacturers who want to be specified - the classic marketplace trade."
- WHY IT SCALESAcelab's stated ambition is bigger than search. It wants to be the intelligence layer for the entire building-materials industry - the place where the science of a material and the art of using it finally speak the same language. The backers seem to agree the prize is structural rather than incremental: alongside Navitas Capital, the Series A drew JLL Spark, DivcoWest, and a roster of architecture-world names including Kai-Uwe Bergmann of BIG and Christopher Sharples of SHoP.
The architect who is not losing her afternoon finishes the cladding decision before lunch. The selection is already in her specs, her submittals log, and her Revit model. The certification check happened on its own. The next person who opens the project sees exactly why she chose what she chose.
Multiply that Tuesday across 20,000 firms and the picture changes: the part of building design that used to leak time and lose knowledge starts compounding it instead. That is the unglamorous thing Acelab is really selling - not a search box, but a building industry that remembers what it learns. The materials were always there. Someone finally made them findable.