Dark stores and the last-mile overbuild.

Gopuff, Gorillas, JOKR, Getir, Fridge No More. At the peak of the rapid-delivery wave, more than a dozen venture-backed companies were competing to deliver groceries in fifteen minutes or less. To do it, they needed density: small fulfillment points positioned within one to two miles of customers. The fastest way to get that density was to lease retail space.

The logic was sound in isolation. A 3,000-square-foot former dry cleaner or sub shop could be converted into a dark store for $80,000–$150,000 in fit-out — a fraction of the cost of building dedicated last-mile logistics. The resulting micro-warehouse could serve a trade area of roughly 1.5 miles, handling 200–400 deliveries per day at maturity. If the economics worked, the model was a brilliant arbitrage: retail rents for logistics output.

The economics did not work. By mid-2022, Fridge No More had shut down. JOKR exited the United States. Gorillas retreated from most markets. Getir began closing locations. The rapid-delivery wave left behind a trail of vacated retail units, most of which had been stripped of customer-facing fixtures — no display windows, no signage, no floor finish. The cost to re-tenant these spaces for traditional retail use runs $40–$60 per square foot, creating a hidden impairment for landlords who leased to the sector.

The unit economics problem

Rapid delivery failed for a specific, quantifiable reason: average order values were too low to cover delivery costs. The typical rapid-delivery order ran $25–$35, against which the operator bore $8–$12 in picking, packing, and delivery labor, plus $6–$9 in allocated rent, overhead, and spoilage. At those margins, profitability required either dramatically higher basket sizes (which customers resisted) or delivery fees that eliminated the convenience premium (which customers also resisted).

The venture capital subsidies that masked these economics were extraordinary — Gopuff alone raised $3.4 billion — but the subsidies could not permanently alter the cost structure. When funding tightened in late 2021, the math surfaced immediately.

The landlord lesson

For community-retail landlords, the dark-store episode offers a clear lesson about tenant underwriting. A tenant whose business model requires ongoing venture-capital subsidy to cover occupancy costs is not a tenant — it is a short-term leaseholder whose departure timing is governed by funding cycles, not real-estate fundamentals.

The rent check is not the question. The question is whether the rent check comes from revenue or from a balance sheet that is burning down.

We passed on every rapid-delivery inquiry we received between 2020 and 2022, not because the rents were unattractive (they were above-market in several cases) but because the tenant credit analysis did not survive contact with the income statement. In community retail, the cost of a vacancy created by a tenant failure — the re-tenanting timeline, the TI dollars, the lost co-tenancy traffic — almost always exceeds the premium rent that an unsustainable tenant offered on the way in.

— End of entry —