Startup Spotlight: Donkey sells container-sized orders to small importers under one contract

Why it matters

Donkey is taking on the sourcing, delivery and price-commitment burden that smaller importers usually navigate through intermediaries. Its one-contract model could simplify container buying, but the company has disclosed little about completed volume, repeat customers or margins, and its 38% savings example has not been independently verified.

Benjamin Martindale grew up in his family's importing business. Now he and co-founder Minghao Tan are building Donkey, a San Francisco startup that sells container-sized orders to smaller businesses that typically buy through distributors. Customers upload a list of products, quantities and specifications; Donkey says it finds factories, calculates duties and freight, and sells the delivered order under one contract. Y Combinator lists Donkey as a five-person company founded in 2025 and part of its Summer 2026 batch. The same model appears in Donkey's YC launch post.

The workflow begins with public customs records. Donkey says its AI agents use them to identify factories behind shipments and obtain quotes, distinguishing manufacturers from trading companies and freight intermediaries. The pricing calculation itself is deterministic: Donkey says it stacks duties line by line, adds freight and other landed costs, then gives the buyer one delivered price. Its homepage shows a 48-hour quote lock, says no payment is due immediately, and says the price holds once the customer signs. Donkey describes a 40-foot high-cube container with 68 cubic meters of volume and a 28,000-kilogram payload.

That contract is central to Donkey's pitch. The buyer purchases from Donkey, which says it takes responsibility for the transaction from factory floor to delivery dock, including inspection before the factory is paid. Donkey says the buyer's balance is due when the goods are accepted at the destination dock. Donkey's public materials do not spell out its importer-of-record role or how it finances purchases before customer payment, details that affect the risks and working capital behind this model.

Donkey frames the opening as a problem of scale: small businesses buy through several intermediaries because they cannot fill a container alone, while freight, currency exchange and tariffs can make quotes move. Donkey says those intermediaries can add 15 to 30 points. It plans to combine demand from smaller buyers in shared containers, with software packing orders and its team checking goods before shipment. Donkey says each container also adds factory pricing data for future quotes. Those are Donkey's descriptions of its model and expected benefits; its public materials do not report how many shared containers it has completed.

Donkey's clearest proof point is a company-reported Oakland home-builder order. The unnamed customer put 15 SKUs, including cabinets, flooring, bathroom fixtures and site materials, into one container. Donkey says the order saved 38% compared with piecemeal distributor purchases. Donkey's homepage displays an anonymized delivered quote of $51,318.92 and comparisons of $76,250 and $73,233 against big-box purchases of the same or equivalent specifications. The homepage presents the buyer as Acme Construction, but does not identify the customer behind the example; the savings and comparison have not been independently verified. One case study shows the shape of the offer; it cannot establish repeatable savings across other buyers or orders.

Martindale's background gives Donkey an operating connection to both sides of that transaction. YC's founder profile says he grew up in his family's Toronto importing business, advised Chinese industrial companies including Sany and Zoomlion on overseas expansion, and ran a Wuhan office in Mandarin. He also founded a Toronto marketing company that closed $4 million in construction deals over six months. The construction-sales work offers a plausible route into Donkey's initial focus on builders, while his importing background is directly relevant to factory sourcing and cross-border sales.

Tan brings the software background. YC says he builds Donkey's data systems to parse customs filings and distinguish factories from shells and freight forwarders. On its current homepage, Donkey says James built agent systems that ran nine figures of revenue; the site does not attribute that work to InstaLILY. Donkey operates between San Francisco and Wuhan, according to its YC profile.

The contracting role sets Donkey apart from tools that help buyers find suppliers or manage procurement. Alibaba.com describes a B2B marketplace connecting buyers and suppliers, with services for discovery, requests for quotes, payments and logistics. Didero sells software that automates procurement workflows across companies' existing systems. Donkey, by contrast, says the customer buys the delivered goods from Donkey itself under one contract. The available evidence does not establish that Alibaba.com or Didero is a direct substitute for that principal-trading arrangement. Didero's February 2026 announcement of a $30 million Series A offers a marker of investor interest in procurement automation, not evidence of direct competition with Donkey.

The public record is thin on commercial scale. Donkey's site and YC profile identify Y Combinator, but do not disclose a Donkey-specific investment amount, valuation or other financing. YC's published standard deal terms do not confirm that Donkey accepted those terms. The reviewed public sources also do not establish Donkey's customer count, revenue, ARR or completed-container total. The Oakland example is Donkey's most concrete disclosed customer result; the available figures do not yet show how often its pricing advantage translates into repeat orders or sustainable margins.