The more I follow Palantir, the less I think it's just an AI software company.

A few years ago I would've described Palantir as a defense contractor with great software. Now I'm not so sure. It feels like it's evolving into an operating system for organizations that need to make decisions from massive amounts of data.

The government side is still a huge part of the story, and that's probably not changing anytime soon. Defense, intelligence, logistics, battlefield software... those businesses continue to grow. But what caught my attention recently is how much the commercial side has accelerated with AIP. U.S. commercial growth has been extraordinary, customer counts keep climbing, and management keeps talking less about selling AI models and more about helping companies actually deploy AI into real operations. That seems like an important distinction.

I also think Palantir occupies an unusual position in the AI ecosystem. Companies like OpenAI, Anthropic and Google build foundation models. Palantir isn't really trying to compete with them. Instead, it's trying to become the layer that helps governments and enterprises integrate those models with their own data, workflows and decision-making. If that's the future of enterprise AI, it's a pretty interesting place to sit.

The obvious risk is that expectations are now incredibly high. The company has been putting up exceptional numbers, but the stock is priced like investors expect years of near-perfect execution. There's also the reality that a meaningful portion of the business still depends on government contracts, and the company continues to attract political and ethical scrutiny because of the customers it serves.

Maybe that's why I find PLTR such a fascinating company. I'm less interested in whether it's "expensive" today than in whether it ends up becoming an indispensable enterprise AI platform or simply remains an exceptional defense software company.

Curious how others see the business.