Andy Alder and Arjun Mather were roommates at Stanford University, where they both studied computer science and then spent 10 years at Palantir. Mr. Alder led product design for a team of 800 people building software for the U.S. defense and intelligence community, and Mr. Mather served as the company’s senior vice president of commercial affairs. When our friends decided to launch a startup early last year, there was no question we would do it together.
Despite having no direct background in architecture, Mr. Alder and Mr. Mather were troubled by the growing problem that, due to rising development costs, fewer buildings are being built in the United States today than they were decades ago.
“We’ve gone from dial-up internet to super-intelligent AI, but our ability to build the physical world around us has actually declined,” Mather told TechCrunch.
Believing that AI will improve productivity across the architecture, engineering, and construction (AEC) industry, they set out to found Dream, a startup that develops a global model to automate building design and engineering. Dream emerged from stealth with $17 million in seed funding led by Andreessen Horowitz, with backers including SV Angel, XYZ, and Zigg Capital. The startup aims to replace clunky and outdated AEC tools like Autodesk’s AutoCAD and Revit.
If you sit next to an architect or structural engineer, you’ll find that they’re still partially “drawing” every part of the design in legacy software, Mather said. “AI doesn’t really play a role in that equation.”
Standard LLMs like ChatGPT can produce images of buildings, but the results are not structurally sound. In contrast, Dream is based on real physics and building codes. Just as LLM learns from text, Dream’s world model is trained on over 100 million square feet of architectural data collected from permitting archives across the United States.
“Our model actually understands what’s inside the walls. It understands what materials are used, what the beams are made of, and how much weight they can support,” Alder said.
The startup calls its approach “structural-aware” because its models generate structurally accurate designs that take into account internal architectural elements such as plumbing and HVAC.
According to the founders, Dream has already started selling its product to AEC professionals and developers who use the model to conceptualize architectural designs and estimate overall costs. The platform has already been deployed in 10 active projects in Texas and California.
Dream isn’t the only company looking to update its AEC software with AI. Other startups working on this include Motif, Snaptrude, and Arcol. Alder and Mather claim to be the only company to have developed a world model that instantly designs entire buildings.
Kimberly Tan, a partner at Andreessen Horowitz, said her firm first invested in Dream over a year ago, when the startup model was just an idea, and has been very impressed with its progress. “I’ve never seen anyone take a technical approach,” she says. “They are all working on their own model.”
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
