He spent years drawing the skins of U.S. embassies. Then he noticed architects were losing whole afternoons to PDFs and spec sheets - and built a search bar for buildings instead.
Architects in the United States steer more than $100 billion in building product decisions every year. There has never been a single place to look them up. Vardhan Mehta found that absurd, and he found it personally. As an architect he had worked on building envelopes for U.S. embassies and projects orbiting MIT and Yale - the kind of work where one wrong material spec can cost a season. The design was thrilling. The sourcing was a scavenger hunt through brochures, spec sheets, and the voicemail of a sales rep who knew the lead times.
Today he runs Acelab, the company behind Material Hub, an AI platform that lets a designer query thousands of products against the constraints that actually matter on a job: cost, lead time, embodied carbon, code compliance. More than 20,000 architecture and design firms across North America use it, including half of the 100 largest firms in the country - Gensler, AECOM, Stantec, CannonDesign. In October 2025 the company raised a $13.5 million Series A led by Navitas Capital, with JLL Spark and DivcoWest along for the round, bringing total funding to roughly $35.8 million.
We're at an inflection point where AI can finally tackle the complexity of architectural decision-making. Our vision is to become the trusted AI infrastructure for every material decision in the built environment - from skyscrapers to schools to single-family homes.Vardhan Mehta, on the Series A, October 2025
Mehta grew up in a small town in central India, in the crossfire between an architect father and an engineer uncle. The built environment was the family dinner conversation. By the time he reached the Harvard Graduate School of Design for his Master of Architecture in Urban Design, the obsession had a target: the material research cycle, the gap between the idea on the screen and the product that could actually be ordered, certified, and installed.
That gap is where most founders see annoyance. Mehta saw a missing system of record. He teamed up with MIT alumnus Dries Carmeliet, and in 2019 the two started Acelab as a student venture. It won the Harvard Real Estate Venture competition in 2020, picked up grants from MIT DesignX and MIT Sandbox, and reached the finals of the Harvard Innovation Lab's President's Innovation Challenge. By August 2021 it had a $3.5 million seed round led by Pillar VC, with Alpaca VC, Draper Associates, and the MIT MET fund. Mehta collected his MAUD that same year.
The next big trend in AEC is everything where we don't have a system of record. That stuff in 2025 should not be a thing - it should be totally automated.Vardhan Mehta, on what comes next
Material Hub ingests manufacturer data, normalizes the specifications, and turns a folder-diving chore into a query. Ask for a cladding that hits a budget, a lead time, and a carbon target, and the system filters structured datasets instead of asking a human to open forty PDFs. But Mehta is careful about the word "automated." He pushes a human-in-the-loop model, where designers validate the machine's output against real evidence and their own past projects rather than trusting it blindly.
That distinction is the whole thesis. The point is not to remove the architect. The point is to give the architect back the hours that were never about design in the first place - and to raise the quality of the decision made at the moment a product gets written into a spec.
The goal is not just efficiency, but decision quality at spec time, where architecture truly meets impact.Vardhan Mehta
Before Acelab, Mehta did time at WEISS/MANFREDI and limonLAB, firms where building skins are treated like engineering problems with an aesthetic conscience. He earned his Bachelor of Architecture at Pratt Institute before Harvard. The throughline is consistent: he likes the place where design meets the unglamorous reality of getting something built. Now he leads a global team that has grown past 85 people, and in early 2025 Forbes added him to its 30 Under 30 list - recognition for a bet that the most boring part of architecture was also the most valuable to fix.
The market he is chasing is enormous and slow: a $2.4 trillion construction materials economy that still runs on tribal knowledge and two-year-old folders. Mehta's argument is that the firms who win will lean on a small number of platforms that are useful to everyone at the table - architect, manufacturer, contractor, client. He wants Acelab to be one of the two or three a firm cannot work without.
If you don't have all the right data, making decisions based on historic specs can still require hours of digging through folders to find something written two years ago.Vardhan Mehta
The aspiration is plain and large. Mehta wants Acelab to be the trusted AI infrastructure behind every material decision in the built environment, and he wants to automate the parts of the AEC workflow that, in his words, should not still be manual. With the Series A he is expanding the AI, entering new geographies, and building deeper hooks into the major design software platforms - the Revit-shaped places where architects already live. A former architect, in other words, quietly rewiring the supply room of an entire profession.
Bars scaled for illustration. Figures from public funding announcements and company statements.
He is a licensed-track architect who now ships software instead of buildings. The tools changed; the obsession with how things get built did not.
Acelab was born as a Harvard GSD and MIT student project before it was ever a venture-backed company. The pitch deck started as a personal complaint.
He insists on human-in-the-loop AI. The machine drafts; the designer signs off against real project history. No blind trust.
Half of the 100 biggest architecture firms in the country run on the product he co-founded. The other half is the roadmap.