Under the new arrangement, Anduril will deliver 1,000 missiles to the Pentagon, an order that forces the company to scale output from nearly zero.
"The amount that we're producing very, very quickly is, I think, enabled by the fact that we've thought about the scalability of supply chain so much," he said.
That focus matters because the usual defense route is slow. Established defense projects can require many years, sometimes several decades, before a concept becomes a functioning system.
Anduril belongs to a generation of younger companies trying to unseat established contractors by applying a Silicon Valley-style emphasis on speed, innovation, and artificial intelligence. Anduril is also trying to shake up its manufacturing processes. Schimpf argues that defense firms should bring more design work in-house, draw on commercial suppliers, and produce at the speed of a tech company, rather than depending on the sluggish, specialized supply networks that have long defined the sector.
The order is a concrete test of that approach. Anduril has built its name on the idea that a newer, software-oriented company can deliver military hardware on a shorter timeline, and the Pentagon's 1,000-missile purchase gives it a chance to prove that model at scale.
Anduril's pitch rests on a simple criticism of the existing defense industry: programs take too long. Schimpf argues that the old approach, with its sluggish, specialized supply networks, is exactly what Anduril is trying to avoid. His preferred model is to keep design work internal, work with commercial suppliers, and move at the pace of a software company. The 1,000-missile order is a chance to show whether that works in real production.
Get the market news that matters in a five-minute read with Market Briefs, our free daily newsletter
On The Tech Download podcast, CNBC's Arjun Kharpal interviewed Anduril CEO Brian Schimpf and brought up Apple as a comparison. Apple, the Cupertino-based tech giant, handles much of its own product development, from hardware design to the chips inside its devices and its software. Anduril is pursuing a similar strategy.
Schimpf was blunt about what Anduril does not want to build.
"Nobody wants the insane version of this. Like, nobody in the military is talking about it. Nobody thinks it's a good idea," he said.
He was describing a future where AI makes big decisions on its own, with no person in the loop.
He also rejected the alternative of keeping artificial intelligence out of weaponry. He called it "just a deeply unethical view, where you're saying, oh, actually, we should have dumber weapons."
His reply to that argument: "That seems like a strictly worse world to live in."
So where does the company draw the line? It comes down to human control of decisions.
"The fundamental thing is we have principles around how we think this should work ... which is very much around human supervision of decision making," Schimpf said.
Schimpf also pointed out that autonomous weapons are already a reality. As examples, he mentioned torpedoes and anti-radiation missiles, which can lock onto and strike particular target signatures without a person directing each small step.
Schimpf stressed that people must stay involved, and that the purpose of AI is to support better human judgment.
Join Market Briefs, our free daily newsletter, for a quick daily rundown of the markets
