# Anduril treats failed tests as part of product development

Markdown copy of https://realalpha.app/s/anduril-treats-failed-tests-as-part-of-product-development/ on Real Alpha, the best minutes from long interviews with the people building AI. Cite and link the page, not this copy.

- Speaker: [Palmer Luckey](https://realalpha.app/builders/palmer-luckey/) · Palmer Luckey is cofounder of Anduril. · Co-founded Anduril, which builds autonomous defense systems, and designed the Oculus Rift that revived VR.
- Host: [Molly O'Shea](https://realalpha.app/people/molly-oshea/) · Founder, Sourcery · Ex-Upfront Ventures investor; has interviewed Alex Karp and Palmer Luckey.
- Show: Sourcery
- Episode: [Palmer Luckey on UFOs, Anduril, Invisibility Tech, & Moon Wars](https://realalpha.app/v/WEGJh-4Iq30/), Dec 22, 2025
- Moment: 4:28 to 7:51 of the episode (3 min), playing on https://realalpha.app/s/anduril-treats-failed-tests-as-part-of-product-development/; the episode from there: https://realalpha.app/v/WEGJh-4Iq30/?t=268

## Quote

> “This is how you actually make functional products. You test them to fail you. **You push the limits**.”
> — [Palmer Luckey](https://realalpha.app/builders/palmer-luckey/) at [6:02](https://realalpha.app/v/WEGJh-4Iq30/?t=362) · Sourcery

## Takeaways

1. Testing to failure helps find product limits.
2. The trials were part of planned test campaigns.
3. Anduril says customers confirmed the tests met criteria.

## About this segment

Palmer Luckey explains why Anduril tests systems until they fail and says the much-criticized trials still met customer criteria. The approach lets the company move quickly on its own dime.

From “Palmer Luckey on UFOs, Anduril, Invisibility Tech, & Moon Wars”, Dec 22, 2025. [Watch the full episode](https://realalpha.app/v/WEGJh-4Iq30/)

## Transcript

Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there.

[4:33](https://realalpha.app/v/WEGJh-4Iq30/?t=273) Can you just go through the history of testing and why failure is part of it? Yeah. I mean, there was a handful of stories. Reuters and Wall Street Journal were both basically trying to write the same story. So you may not know this. I was a journalism major before I started Oculus, my first company, and of course before Anduril. And so I care about this on two levels. One is the kind of distraught tech founder who is upset when the press either doesn't understand or willfully misrepresents what his company is doing.

[5:02](https://realalpha.app/v/WEGJh-4Iq30/?t=302) And separately, you know, the single tear that I shed for the state of journalism in America. Remember, I wanted to be a journalist at one point. So people have said Palmer hates journalism. He doesn't want, you know, a free press. It's kind of nonsensical if you look at me and you look at my career where I used to be the online editor of one of the largest college papers in the country. The insult doesn't quite connect. In this case, they were making a huge deal out of the fact that we had had two aircraft that failed to release during a test. So they failed to detach from another aircraft. Also, they're making a huge deal out of a fire that we started during a test. And to be clear, this is a Very, very small fire.

[5:40](https://realalpha.app/v/WEGJh-4Iq30/?t=340) It was about 0.00002% of a weapons test range specifically designed to have fires on it. And the real, they said satellite imagery, reveals the extent of the damage. There was a bunch of grass that caught on fire, and then it got put out, and it was fine. Like, nothing burnt down. Nobody was injured. It's exactly what it was like. And the crazy thing is, Anduril has started hundreds of fires during our test procedures over the course of years.

[6:08](https://realalpha.app/v/WEGJh-4Iq30/?t=368) This is how you actually make functional products. You test them to fail you. You push the limits. You find out where things are going wrong. Also, in this case, it was a counter-drone system. What do you think is going to happen when a system designed to destroy drones full of flammable lithium-ion batteries is tested? Do you think that's never going to end in things on fire falling onto the ground?

[6:29](https://realalpha.app/v/WEGJh-4Iq30/?t=389) And they said Andrel runs into roadblocks and obstacles as tests fail. That's actually not the case. Those tests were extremely valuable. They're part of our usual test campaign. We've had hundreds, if not thousands of failed tests over the eight years that we've existed. We've started hundreds of fires. Usually by the time we're doing a release, it actually works. It's actually quite rare for that not to work.

[6:49](https://realalpha.app/v/WEGJh-4Iq30/?t=409) And of course, they don't mention that the Air Force gave them a statement saying our stuff was awesome and passed the certification. They don't run the statement for the Navy saying, by the way, the exercise you're literally talking about was a success for Andrel and they met all of the criteria that we talked about. Nobody mentions those things. They just want to talk about what they see as problems because they have no idea what product development looks like.

[7:10](https://realalpha.app/v/WEGJh-4Iq30/?t=430) If I ran a product development cycle that was two years late and a billion dollars over budget, they'd have almost nothing to say about it. But then if I have a successful failed test as part of moving quickly, then they're going to write a whole bunch of what they believe are exposés. The only difference between us and a lot of the other companies that are working in the defense space is that we are doing this on our own dime.

[7:34](https://realalpha.app/v/WEGJh-4Iq30/?t=454) We're spending our own money to move as quickly as we can. It's not like taxpayers are paying for that drone test that we were doing internally. Anyway, I've said this publicly, but we're not going to be changing our approach.