Anduril Just Won a US Air Force Contract to Build Autonomous Fighter Jets

I initially thought this story was mostly about Anduril's growth trajectory, sort of a "look how far the startup has come" narrative. But after reading through the context more carefully, I think the more interesting question is what this signals about how the US military is thinking about AI autonomy in combat systems right now.

The timing isn't nothing. We're in a moment where the government is simultaneously trying to restrict AI model access to foreign nationals and handing contracts for autonomous weapons to a company that's less than a decade old. Those two impulses, tight control on one side and rapid deployment on the other, are going to create friction eventually.

This raises questions about oversight and accountability. Well, multiple things, actually. Who's responsible when an autonomous fighter makes a decision that causes civilian casualties? How do existing international laws of armed conflict apply to systems where human judgment is removed or reduced from the loop? These aren't hypotheticals anymore.

For what it's worth, Anduril has consistently argued that their systems keep humans in control of lethal decisions. Whether that holds as the systems get more capable and operational pressure mounts is something we'll have to watch closely. It's too early to say whether this contract represents a genuine shift in how autonomous weapons are being deployed or whether it's a more constrained system than the "autonomous fighter" framing suggests.

What I do think is clear: the pace of this is accelerating faster than the policy frameworks around it. And that gap is worth watching.

A causal adaptation model hits a Cohen's kappa of 0.88 against human raters, while a depth-vision fusion system outpaces recent baselines on two standard benchmarks. The gap between lab and corridor is narrowing.

James Chen · Yesterday · 5 min