Palantir Trades at 200 Times Earnings. Here Is Whether That Can Possibly Make Sense.
PLTR has pulled back 35% from its peak. The valuation is still stratospheric. The growth is genuinely extraordinary. Here is how to think about it honestly.
Start With What Is Actually True
The temptation with Palantir is to lead with the valuation, 200x trailing P/E, roughly 80x forward, and stop there. That number is so extreme that it becomes the whole story for most analysts. But it was also extreme at $20, at $50, at $100, and at the $207 peak in late 2025. The investors who dismissed it at each of those levels were technically correct about the multiple and spectacularly wrong about the stock. So the multiple alone is not sufficient analysis. You have to actually look at what is happening to the business.
What is happening to the business is remarkable. Q4 2025 revenue hit $1.41 billion, up 70% year-over-year. U.S. commercial revenue, the segment that matters most for the long-term thesis, grew 137% in the same quarter, and it grew over 121% the quarter before that. This is not a company where the growth is decelerating. The Rule of 40 score, which combines revenue growth rate and profit margin, came in at 127. For context, most software companies are thrilled to hit 40. A score of 127 at this revenue scale is historically rare.
The 2026 guidance of $7.2 billion in revenue implies 61% growth. At the current $130 share price, you are paying approximately $310 billion for a software business that did not exist in its current form four years ago.
The AIP Boot Camp Strategy Is Smarter Than It Sounds
Palantir's primary commercial growth engine is the AIP Boot Camp, an intensive, on-site workshop where enterprise clients spend two to five days building AI workflows on Palantir's Artificial Intelligence Platform. The average time from boot camp to signed contract has reportedly compressed to under a month. For enterprise software, where sales cycles typically run six to eighteen months, that is a structural disruption of the buying process.
The reason it works is specific: Palantir does not sell a product, it sells a solved problem. They walk in, use your data, build something that works in your environment, and show you the output before you have signed anything. It is a demonstration-led sales model that eliminates the "we need to see it work first" objection, the most common reason enterprise software deals die. The expansion from boot camp to broader deployment is where the real revenue comes from.
The Government Business Is Not a Liability, It Is a Moat
A common bear argument is that Palantir is too dependent on government contracts, which are lumpy, political, and potentially cancelable. This argument was more compelling two years ago. Today, government revenue grew 55% year-over-year in the most recent quarter. The USDA signed a $300 million contract for AI-powered agricultural analytics. The U.S. Army expanded the Maven Smart System. The TITAN program cemented Palantir's position in battlefield AI.
The subtler point: government contracts are not just revenue. They are references. A defense contractor that can say "the U.S. Army trusts us with battlefield data" carries a credibility in commercial enterprise sales that no marketing budget can buy. The government business validates the commercial business in a way that investors systematically undervalue.
The Honest Bear Case
Michael Burry, who made his name being right about the 2008 housing crisis, disclosed a short position against Palantir in late 2025. His argument, as best as can be reconstructed, is that AIP boot camp contracts are pulling forward demand that would otherwise take years to develop, and that normalized growth rates will be dramatically lower than current trends suggest. There is also the hyperscaler threat: Google, Microsoft, and Amazon are all building AI analytics platforms that will eventually compete for the same enterprise budget.
The valuation math is punishing to any growth slowdown. If Palantir's revenue growth decelerates from 60% to 30%, which would still be exceptional, the stock would likely need to reprice at a much lower multiple. At 30x revenue (generous for decelerating software growth) and $10 billion in 2027 revenue, that implies a market cap around $300 billion. Roughly where the stock is today. The upside requires the extraordinary growth to persist; the downside is repricing to fair value.
The Framework That Actually Matters
The right way to think about Palantir is not as a traditional software investment, it is as a call option on becoming the default AI operating system for both government and enterprise. If that vision materializes, 200x earnings will look cheap in retrospect. If it does not, and the company re-rates to 40x forward earnings like a normal enterprise software business, the stock falls 70%+. That is not a thesis to size at 15% of your portfolio. It is a thesis to size at 3-5% if you believe it, and zero if you do not. There is no moderate position that makes sense here.