The Navy asks venture capital to fund the early research it used to pay for itself

The funnel has a stage filter on the intake. "We mostly buy Series D through F type companies," Fanelli said ^[8]. For a founder that reorders the sequence: the product reaches maturity on private money, and the Navy arrives as a customer after it works. Fanelli said the common form of co-investment is waiting for a company to mature a product on its own before buying it, and that taking an equity stake remains rare ^[7].

He put the obligation for that on his own side. "The cost of that, or the responsibility, is for us, if we're not going to do it ourselves, to cast a cleaner signal," he said ^[10]. Investors told him the first priorities publication changed how they thought about the Navy's buying plans, and he said that is part of why he is doing it again ^[4].

He also gave a size for the whole of it. "We spend in the like $150 billion range every year," Fanelli said, separating that figure, total Navy purchasing, from anything as narrow as direct equity investment ^[5]. Most of it still flows through traditional channels ^[6]. This month's MQ-25 award comes to roughly 0.4 percent of the annual figure he cited ^[17].

One of the smaller purchases explains adoption better than the large one. Gecko Robotics took over inspection work that used to be done manually and dangerously, and Fanelli said the move drew little pushback because almost nobody wanted that job in the first place ^[13]. Armada's servers packed into shipping containers and Domino Data Lab's work on the Navy's machine learning pipeline sit on the same recent list ^[12][14].

Fanelli is both the person building the funnel and the source for how it is working. Asked whether any of the commercial technology is in use around the Strait of Hormuz, he said, "I often don't know what's classified and unclassified because I'm normally talking to people with clearances," and said he would have to check ^[16]. He did not say how many pilots have been promoted into enterprise services ^[18].

Two tests sort a defense pipeline faster than a priorities list does. The first is whether someone inside wants the current job to go away, as with the inspection work ^[13]. The second is whether the company can reach the stage the Navy says it buys at without Navy money ^[8][9]. Both yes, and the pilot is worth its cost of sale. If a company clears the first test but not the second, the funding gap belongs to the fund, because the Navy is trying to stop paying for it ^[9]. Fail the first and the pitch lands in a program office content with the system it already has. The camera swap worked on a program that was already years late ^[15].