From Failed Startup to AI Breakthrough: How Nick Donahue's Second Bet on Home Design is Winning Over Top Investors

Nick Donahue knows the housing industry inside and out. Growing up with a father who built homes for major developers and a mother selling to big-box builders across the East Coast gave him a front-row seat to an industry’s biggest problem: why does designing a custom home cost so much and take forever?

His first attempt at solving this problem, Atmos, showed real promise. The company went through Y Combinator, raised $20 million from investors including Khosla Ventures and Sam Altman, and employed designers who worked alongside software to streamline the custom home process. By the time interest rates climbed and the market froze, Atmos had designed $200 million worth of homes, built 50, grown to 40 people, and generated $7 million in revenue.

But Donahue saw the fundamental flaw: “It became this extremely operational business—kind of like a glamorized architecture firm.” The tech never truly replaced human labor, and when financing dried up, so did the business. Nine months ago, he shut it down.

The AI-First Pivot

Most founders would take a victory lap and write LinkedIn think-pieces. Donahue did something bolder. Drafted, his new venture, is the opposite of Atmos in almost every way. No in-house designers. No operational overhead. Just AI-driven software that generates residential floor plans and exterior designs in minutes based on user preferences.

The company is barely five months old but already securing serious backing. Jack Altman, Bill Clerico (who also invested in Atmos), Stripe’s Patrick Collison, Josh Buckley, and Warriors player Moses Moody just put $1.65 million into the company at a $35 million post-money valuation. Clerico was so convinced by Donahue’s vision over coffee that he allegedly repeated “Nick, please take our money” for two weeks straight until Donahue agreed.

The Economics Make Sense

Today, if you want a custom home, you face a binary choice: hire an expensive, slow architect or buy a cheap but rigid template. Drafted occupies the middle ground—customization at template prices. Complete plans cost $1,000 to $2,000.

The unit economics work because Drafted built its own specialized AI model trained on real house plans from actual built and permitted homes. While general-purpose AI costs 13 cents per floor plan, Drafted’s custom model costs just two-tenths of a penny. That’s not a rounding error; it’s a structural advantage.

Market Reality vs. Founder Ambition

Of America’s 1 million new homes built annually, only 300,000 are custom designed. Most people buy existing homes or accept whatever the big builders offer that year. Clerico’s bet is that this represents a chicken-and-egg problem—make custom design cheap and fast enough, and adoption explodes.

Donahue compares the opportunity to Uber’s transformation of transportation: “There’s really no reason in the future why everyone shouldn’t have a totally custom designed home.” Whether that vision matches market reality remains unclear. The housing industry has a reputation for resisting disruption.

Early traction is modest but steady. Since going public, Drafted has attracted about 1,000 daily users—not massive, but showing consistent growth for a product this young. The company is focusing on single-story homes for now, with multi-story and lot-specific customization coming later.

The Competitive Moat Question

The harder question: what stops an LLM player or another vertical competitor from acquiring similar datasets and replicating Drafted’s product? Donahue points to brand stickiness, citing his friend David Holz at Midjourney as proof. Despite an explosion of image-generation models, Midjourney’s user engagement barely wavers—customers keep returning to the same platform.

If Drafted moves fast, delights customers, and builds a community, it could claim the same defensibility. Speed, execution, and customer satisfaction might be the only moat that matters when the underlying technology becomes commoditized.

Nick Donahue’s journey from Atmos to Drafted tells a story about iteration in hard problems. The first swing taught him what doesn’t work. The second swing, armed with AI, founder credibility, and top-tier capital, might finally crack an industry that’s resisted change for decades.