Bloom raises $3.6M to broker the hardware supply-chain work it once planned to do itself

A company on Bloom posts a contract job. Another company bids for it, and Bloom handles the quote, the booking and the payment ^[9]. The work on offer covers contract manufacturing, assembly, design and engineering, freight, warehousing, repairs and hazardous-materials shipping ^[10].

The pitch around that flow is bigger. Kosmides, Bloom's CEO, has compared the company to an AI-driven version of what Alibaba did for contract manufacturers in China ^[16]. TechCrunch describes the product as, essentially, supply-chain AI agents that customers use to find suppliers, parts and services ^[4]. Under the agent language, what a member gets is matching plus the transaction ^[9].

Kosmides built Bloom around one failure he kept seeing at mobility companies: they could not or would not outsource the hard work ^[1]. The marketplace helps with finding suppliers. After a booking, the supplier does the work and the buyer keeps the suppliers in step ^[17]. A drone maker that books an assembler and a freight firm through Bloom has to manage the handoff between them itself.

TechCrunch links the change to tariffs. In its account, Trump's tariffs on dozens of countries sped up a US hardware scene that was already growing ^[11]. Robotics startups and drone makers joined mobility firms in looking for domestic suppliers ^[11]. Serving that wider group took a partial reinvention, the outlet reports ^[3], and the reinvention slowed Bloom's fundraising ^[12]. Kosmides positions the result against the incumbents. "Our friends at the parts marketplaces like Fictiv, Xometry, MacroFab and others are really good yesteryear marketplaces for finding one particular part," he said ^[14]. He set them against what a drone company or an electric-motorcycle company needs ^[15].

Spread across customers, Bloom's match count comes to about 14 matches each, taking both reported figures at their floors ^[18]. SNAK's revenue comparison puts this year's monthly pace at about 2.4 times last year's average ^[19]. The firm had passed on Bloom's pre-seed. "We liked the founder and the thesis and passed anyway: we wanted to see more traction," it wrote ^[13]. SNAK also said memberships grew fivefold "with low churn" ^[8], the nearest thing in the record to a retention figure. Neither Bloom nor SNAK gave absolute revenue, member counts, a split of customers by industry, or how many matches became booked jobs.

An operations lead deciding whether to route sourcing through Bloom can sort the choice on two facts about the team. One is how many supplier categories a single product needs. The other is whether someone on staff can manage suppliers once they are booked. One category with a manager in place points to the parts marketplaces Kosmides names ^[14]. Several categories with a manager is the buyer Bloom's marketplace now fits. Several categories with nobody to manage them is the buyer the 2023 plan was written for ^[1], and the marketplace no longer does that work ^[3]. One category with no manager is a single-supplier relationship, and a marketplace adds little to it. I think Bloom is a sensible trial for the second group, provided the team runs one real job in its own category and asks for the match-to-booking rate before moving production. The tradeoff is committing on traction figures that are ratios, most of them supplied by the lead investor ^[7][8].