Board Case Study · PropTech (Real Assets)
The firm bought thousands of homes on an AI valuation model. In a turning market the model was systematically wrong, and the portfolio is underwater nine figures. Investors were told it de-risked the strategy.
The situation
A PE-backed real-estate firm scaled a single-family-home buying strategy on an AI valuation model that priced and acquired thousands of properties. When the market turned, the model proved systematically over-optimistic, and the portfolio is now underwater by a nine-figure sum.
The model extrapolated from a rising market and underweighted regime change; no one stress-tested it against a downturn. Investors were told the algorithm reduced risk and improved discipline, the opposite of what happened. The lender's covenants are now in play.
The board must address the valuation write-down and LP disclosure, the governance of model-driven capital allocation, and accountability for a strategy sold on the premise that AI made it safer.
“We bought thousands of homes on the model's word and called it discipline. It was confidence without a downturn in it. Who owned that assumption?”
Chair, Investment Committee
The decision on the table
The full case, scenario architecture, board materials, and facilitation notes, is shared with boards and partners on request.