FEATURED

Palantir Is Down 20%. Its Business Never Ran Faster.

Revenue up 85% year over year. U.S. commercial revenue up 133%. Full-year guidance raised by 10 full percentage points in a single earnings call. And the stock is still down roughly 20% year-to-date.

Something doesn’t add up. Or maybe it does, and it’s the valuation that’s the real issue, not the business.

This is the Palantir question that keeps coming up. And right now, with the stock trading well below its late-2025 highs while the underlying business accelerates at a pace few software companies have ever sustained, it’s worth working through what you’re actually owning here.

What the Q1 Numbers Actually Say

Palantir’s Q1 2026 results were not subtle. Revenue came in at $1.633 billion, beating consensus of $1.54 billion, with adjusted EPS of $0.33 against a $0.28 estimate. The company raised full-year 2026 revenue guidance to approximately $7.650 to $7.662 billion, representing roughly 71% year-over-year growth. That’s not rounding error. That’s a business operating at a different speed than its peers.

U.S. total revenue grew 104% year over year. The company’s Rule of 40 score hit 145%, an extraordinary reading that reflects both growth rate and profitability margin combined. Adjusted free cash flow for the quarter came in at $925 million, a 57% margin. These are not the metrics of a company that is losing the AI race. These are the metrics of a company that may be winning it in ways that don’t yet show up in investor sentiment.

What’s interesting is how the commercial side of the business has evolved. U.S. commercial revenue grew 133% year over year to $595 million and is now guided to grow at least 120% for the full year. That’s the piece that used to be the knock on Palantir: too government-dependent, too hard to replicate commercially. That criticism is increasingly hard to sustain.

The Defense Side Is a Different Story

And then there’s the government business, which is accelerating in its own right. Palantir holds a contract with the U.S. Army worth up to $10 billion over the next decade, consolidating 75 separate military contracts into one enterprise deal covering the Army’s future software and data needs.

U.S. government revenue hit $687 million in Q1 2026, up 84% year over year. The Pentagon is also seeking $2.3 billion over the next five years to expand Palantir’s Maven Smart System, an AI-powered platform that identifies and analyzes battlefield data. That request is part of a proposed $1.5 trillion defense budget for fiscal 2027, a significant jump from current spending levels, and Palantir’s software is embedded at every layer of the architecture that budget would fund. It’s not a weapons contractor. It’s the operating system that makes the weapons contractors more effective. That’s a fundamentally different and more durable business model.

One more thing on the defense angle: classified government environments don’t run on commercial AI APIs. Anthropic can’t compete here. Microsoft’s consumer Azure tools can’t compete here. Palantir’s Gotham platform operates in environments where none of those alternatives are allowed. That’s a structural moat, not a temporary advantage.

The Valuation Tension Is Real

Here’s where I’m not going to pretend the bear case doesn’t exist. It does. Palantir trades at a forward price-to-earnings multiple of roughly 75 to 80 times, depending on which earnings estimate you use, and a trailing P/E that’s north of 120. Those are numbers that leave very little room for error. Even with 85% revenue growth, any deceleration in commercial momentum, any government contract cancellation, or broader multiple compression in software, and the stock reacts fast. It already has: roughly 20% from its January 2026 highs is not a minor correction.

The insider selling is also worth acknowledging. Directors Karp, Cohen, and Sankar collectively disposed of a significant volume of shares in May, though many of those sales coincided with RSU-to-Class-B conversions representing compensation-related liquidity. Some directors also acquired shares on the open market in early June. That’s not a clear signal either way, but it’s worth watching.

What’s different here from most high-multiple software names is the durability of the contracts. The company reported total remaining deal value of $11.8 billion as of March 31, 2026, up 98% year over year, providing forward revenue visibility that most software companies simply can’t point to. Remaining performance obligations hit $4.45 billion, up 134% from a year earlier. And the AIP boot camp model, where enterprises trial the platform in short, intensive sessions and convert to multi-year contracts, is a distribution approach that scales in ways traditional enterprise sales doesn’t.

The Bigger Picture

Palantir hosted AIPCon 10 on June 4, showcasing real-world deployments across companies including Hertz, Accenture, McCarthy Building, the U.S. Department of Agriculture, and law firm Kirkland and Ellis. That last one is worth a pause. A private equity fundraising platform built jointly with one of the world’s largest law firms suggests the commercial expansion is reaching corners of the market that weren’t on the radar a year ago.

The gap between where the stock is and what the business is doing is unusually wide right now. Whether that gap closes up or down depends almost entirely on whether U.S. commercial growth stays above 100% through Q3. If it does, the multiple will be easier for the market to digest. If it slips, the stock will adjust quickly.

The business, though, the actual operating machine underneath the valuation debate, is running as well as it ever has. Net dollar retention of 150%. A customer count of over 1,000 commercial clients. Q2 guidance of $1.797 to $1.801 billion. In a market where most AI software names are still trying to prove revenue traction, that’s not nothing.