You're seeing the preview. Pro unlocks the full Atmos teardown, the rebuild plan, every technical spec in the database, and 5 fresh report requests each month.
This report was generated by our Deep Research agent and may contain mistakes.
Did we get something wrong? DM @oscrhong and we'll fix it ASAP!
Atmos was a San Francisco-based homebuilding marketplace that operated from approximately 2018 until March 2025.[1][2] The company positioned itself as an end-to-end platform for custom home design and construction, connecting buyers in Sun Belt and Southeast markets with vetted builders, handling design approvals through 3D technology, and coordinating financing — all under one digital roof.[3] It graduated from Y Combinator's Summer 2020 batch and raised approximately $20 million across three rounds, including a $12.5 million Series A led by Khosla Ventures.[4]
Atmos failed for two compounding reasons. Its technology never automated away the human designers and project coordinators at the core of its service, leaving it structurally exposed as an operationally heavy business. When the Federal Reserve's 2022–2023 rate hiking cycle made custom home financing unaffordable for its target customers, the pipeline of months-long design engagements collapsed without converting to revenue.
The company shut down around March 2025 after completing a lifetime total of roughly 50 homes and designing over $200 million in projects that never broke ground.[5] CEO Nicholas Donahue subsequently launched Drafted, an AI-first floor plan generator with no designers on staff — the product he described as "the thing we always wanted to build."[6]
Nicholas Donahue grew up inside the homebuilding industry. His father built homes for major developers; his mother sold homes to big-box builders across the East Coast.[7] That background gave Donahue an unusually granular understanding of where the process broke down for buyers — the opacity of builder selection, the fragmentation of design and construction, the absence of any single coordinating layer that a homebuyer could trust. He dropped out of NC State and moved to the Bay Area before founding the company.[7]
Donahue co-founded Atmos with Trent Hedge, Matthew Rastovac, and Austin Kahn, with Crunchbase placing the founding year at 2018.[1] The team's initial concept was VR-focused — an immersive visualization tool for homebuilding — before pivoting to the managed marketplace model that would define the company's public identity.[8] The details of the original VR product and the precise reasons for abandoning it are not documented in any public source, but the pivot suggests the team recognized early that visualization alone was not a defensible business — the value had to sit in the transaction itself.
By the time Atmos entered Y Combinator's Summer 2020 batch — the first fully remote cohort, shaped by the COVID-19 pandemic — the core thesis had crystallized: homebuilding was a $400 billion industry still coordinated largely by phone calls, paper contracts, and fragmented local relationships.[9] Atmos would sit above that chaos as a managed marketplace, providing buyers with a single interface and builders with a qualified, pre-vetted client pipeline.
The founding team's domain credibility was a genuine asset in early fundraising. Sam Altman and Adam Nash participated in the March 2020 seed round, and Khosla Ventures followed within months.[10] The investor base was consistently design-tech and AI-forward — Figma CEO Dylan Field joined the Series A cap table — suggesting the original vision was always software-first, even as execution forced a services-heavy detour.
Rastovac later became Head of Engineering at Atmos before departing to become CEO of Respell, which was subsequently acquired by Salesforce.[11] Hedge went on to found Pylon.[12] The exact timing of these departures relative to Atmos's wind-down is not publicly documented.
Read the complete post-mortem, the rebuild playbook, and the exact reasons Atmos is still worth studying now.