Company: Spacial
Co-Founders: Maor Greenberg (CEO), Ami Avrahami (CPO)
Website: https://spacial.io/
About Spacial
Spacial builds an AI model that turns architectural plans into permit-ready structural and MEP engineering plans. Every set is reviewed and stamped by our in-house, licensed engineers, so the output is not a suggestion; it is a document a building department can accept.
In many cases, the problem arises when a document passes between different departments. For example, a designer could draw a layout. Then, two weeks later, an engineer flags an issue in the layout. The layout then changes, which breaks something else downstream, and another round of revisions begins. Each round is a week or three of correction work, not design work. Most companies are making one part of that loop faster. Spacial is eliminating the loop. Engineering knowledge sits inside the design surface, so a structure that will not work never gets drawn. Cut that cycle and the savings build more housing.
Spacial was founded in 2024 in Palo Alto by Maor Greenberg, who spent two decades in construction and development watching this loop drain time and money from architects, builders, and engineers. Ami Avrahami, credited on more than 30 patents in property technology and AI, leads product, and Jay Klein leads AI. Most of the team comes from structural engineering and architecture, so the problem is not theoretical to them.
The first challenge had to do with data. There isn’t a public dataset of engineered, permitted residential buildings, so the company opened a licensed practice, Spacial Services, delivering structural, architectural, MEP, and Title 24 work to real clients. Every project yields validated proprietary training data and a paying customer. Revenue funds the research, and the research sharpens the practice.
The second challenge was the code itself, scattered across separate databases owned by different jurisdictions, in different systems and formats, and revised constantly. Gathering and reconciling it, and keeping it current, is an ongoing effort.
The third was technical. How do you get a machine to understand physics rather than lines and shapes, to know what a wall is and detect on its own whether a door placed in it is physically sound? Existing tools draw geometry without knowing what it is made of or what it carries. Spacial built its own physics engine so the software reasons about a building as a structural system.
Spacial has delivered 47 projects with 228 in production, concentrated in California and extending into other states. The work spans single-family homes, townhome communities, and accessory dwelling units, for modular and prefabricated home manufacturers and custom builders. Engineers are hired and licensed in the United States, not offshore.
Growth means opening the software directly to architects and builders, adding code coverage in new states, and training the model on every project the practice delivers. Spacial is raising a Series A to fund it.
Want to find out what TechRound can do for your business?
