Trae Stephens on Anduril's Arsenal One factory, Thunder autonomous attack helicopter, and defense tech overhype

Anduril's Arsenal One, Thunder, and the Defense Tech Overhype Problem

Trae Stephens, co-founder of Anduril and partner at Founders Fund, makes the case that defense tech has moved from a scrappy contracting fight into a full-scale manufacturing moment — but warns the category is now drowning in hype that will be difficult to sort through.

Thunder autonomous attack helicopter

Anduril announced Thunder at the Farnborough Air Show, an autonomous attack helicopter designed to operate alongside crewed aircraft. The concept mirrors the Fury collaborative combat aircraft program: a human pilot commands a fleet of autonomous vehicles from their own cockpit rather than putting a crewed helicopter directly in harm's way. Stephens describes it as the pilot becoming "Ender in Ender's Game," directing autonomous assets to take fire, gather intelligence, or strike — while the crewed aircraft stays out of range. Anduril has completed test flights with a full-scale surrogate and is targeting Thunder's first flight in 2027.

“Thunder is an autonomous attack helicopter that we just announced at Farnborough in The United Kingdom last week... We just rolled our first unit off the line to the Air Force earlier this week. The proliferation of drones has turned the near surface fight into a robotic kill zone. And so crewed helicopters or crews, they're at risk.”

Arsenal One and the Fury rollout

Anduril rolled its first Fury collaborative combat aircraft off the Arsenal One production line in Columbus, Ohio this week, delivered to the US Air Force. The facility went from groundbreaking to operational in roughly 18 months — ribbon-cutting in January 2025, factory running by May 2026. The building is 800,000 square feet, originally a concrete shell, built out with support from JobsOhio, Ohio's economic development agency.

Stephens credits COO and co-founder Matt Grimm with the execution and is candid about what made the timeline possible: the team didn't wait for the factory to open before hiring. Everyone now working on the Fury line was already doing the same work at Anduril's Orange County headquarters months earlier. Supply chain constraints were real — rare earths and natural resources remain a bottleneck, with the Department of Defense acting as an active partner on offtake agreements. A second building shell is already standing next to Building 1.

Defense tech overhype

The sector has matured considerably since the early Palantir era, when Palantir and SpaceX had to sue the government under Title 10 USC 2377 just to get commercial contracts through. Pathways like the Defense Innovation Unit, AFWERX, Army Futures Command, and SBIR stratification funding now give startups routes to tens of millions in matching government capital. But Stephens draws a sharp line between R&D funding and production funding: the government's default posture is still to spread small amounts across many companies rather than commit to large-scale acquisition programs, and that muscle remains underdeveloped.

His bigger concern is the venture dynamic. Defense tech has become a category, and once something is a category, the window for generational returns has largely closed. He points to SpaceTech and crypto infrastructure as analogies: investors who missed the anchor company in each category mostly lost money. He argues there are now too many defense companies chasing venture capital at inflated prices, and the signal-to-noise ratio is deteriorating.

He sees a genuine opportunity in the opposite direction — smaller businesses building specific components like drone motors, structured with private equity rather than venture capital, targeting $50 million in revenue with reasonable EBITDA and no IPO ambitions. His observation is that every company that should be doing exactly that is instead raising tens of millions in venture dollars with an unclear endgame.

Founders Fund

Founders Fund recently added Ryan Beiermeister as a partner. She and Stephens worked together at Palantir in their early twenties; she subsequently went to Meta and then OpenAI. The firm continues to run a main venture fund alongside a growth fund and has no plans to move toward evergreen or public market structures.