In late 2021, Zillow's 'Offers' arm — which used machine-learning 'Zestimates' to iBuy homes sight-unseen at algorithmically-set prices — had accumulated thousands of houses it had overpaid for as the model failed to keep up with a whipsawing pandemic real-estate market. On November 2, 2021, CEO Rich Barton shuttered the entire iBuying program, announced a $881 million inventory writedown, and laid off 25% of Zillow's workforce. Internal docs later showed Zillow was losing an average of $25,000 per flipped home. Barton's memo admitted 'we've been unable to accurately forecast future home prices.' It is still the single largest publicly-reported dollar loss from a deployed AI/ML system in an industrial setting.