Kozmo delivered DVDs and snacks in one hour — and burned $280M doing it

Kozmo.com, founded in 1998 by Joseph Park and Yong Kang in New York City, promised free one-hour delivery of DVDs, video games, snacks, and magazines. The service was free — no delivery charge, no minimum order. Kozmo delivered by bicycle, on foot, and by subway, targeting dense urban neighborhoods.

The model was popular but unprofitable. Delivering a $4 DVD or a $2 bag of chips in one hour, for free, in Manhattan, cost more than the product was worth. Kozmo expanded to 11 cities anyway, raising $280 million from investors including Amazon. The company burned cash on marketing, warehouse space, and delivery staff.

The moment it turned is on the public record. Kozmo had filed to go public in March 2000 — the IPO was what the expansion was being funded against — and on 18 August 2000 it filed Form RW with the SEC, formally withdrawing the registration. The dot-com window had shut. Amazon, which had put in $60 million, declined to put in more. Kozmo closed on 11 April 2001 and laid off 1,100 people, eight months after the withdrawal.

Kozmo scaled free delivery to 11 cities, raising $280M, but every order lost money; when Amazon declined more funding, it shut down.

Free one-hour delivery of low-margin products is structurally unprofitable — the delivery cost exceeds the product margin on every order.

Expanding to 11 cities before proving unit economics in one multiplied the losses without improving the model.

Amazon's $60M investment created a false sense of security; when Amazon declined to fund further losses, Kozmo had no fallback.

The dot-com funding environment rewarded growth metrics (deliveries, cities) over profitability, encouraging Kozmo to scale losses.

Free delivery is a cost, not a feature. Kozmo delivered $4 DVDs for free and lost money on every order. Scaling to 11 cities multiplied the losses.

Kozmo shut down in April 2001. Joseph Park later co-founded Askville (acquired by Amazon) and other startups. The one-hour delivery concept was eventually proven viable by companies like Gopuff and Getir — but with higher minimum orders, delivery fees, and denser warehouse networks. Kozmo was right about the demand and wrong about the economics.