§ 01 Business Overview
Palantir builds software that helps large, complex organizations understand their data and do something with it. Gotham serves government and defense clients, helping intelligence agencies and militaries connect data and make decisions in high-stakes environments. Foundry is the commercial product, designed to break down internal data silos so a company can see how all its moving parts connect. AIP is the newest layer and the fastest-growing. It lets companies run AI agents inside their own secure environments without shipping sensitive data to outside servers.
How they sell is worth understanding. They now run AIP Bootcamps: five-day sessions where a potential client builds a real working AI tool using their own data. Sales cycles have collapsed from months to days, and customer count surged 69% in the last fiscal year.
§ 02 Competitive Moat · Strong (Gov) / Moderate (Commercial)
The government moat is close to unbreakable. Security clearances take years to obtain. Classified integrations run so deep into defense infrastructure that replacing Palantir is practically impossible. No startup is replicating this in five years.
The commercial moat is real but more contested. The key mechanism is what they call the ontology, a proprietary framework that maps an organization's entire operational reality into a structured data model. Once a company builds its business processes on top of Palantir's ontology, switching means rebuilding that entire digital foundation from scratch. The longer a client uses the platform, the more accurate and indispensable it becomes.
The risk is that Databricks and Snowflake are moving fast into adjacent territory. The hyperscalers (AWS, Azure, Google Cloud) are not far behind, and they can undercut on price. In commercial markets, Palantir's moat is strong but not yet airtight.
The government flywheel is the real insight here: government deployments prove the product at the highest stakes, which builds commercial credibility, which drives enterprise adoption, which generates more data, which improves the AI, which wins more government contracts. The government business is not just revenue. It is the ultimate reference customer.
§ 03 Financial Snapshot
Revenue growth has been accelerating, not decelerating, which is rare at this scale.
| Period | Revenue | YoY Growth |
|---|---|---|
| FY 2022 | $1.91B | 24% |
| FY 2023 | $2.23B | 17% |
| FY 2024 | $2.87B | 29% |
| Q3 2025 | $1.18B (quarter) | 63% |
| Q4 2025 | ~$1.28B (quarter) | ~70% |
Gross margins have reached 82%, with operating margins hitting 31.6% in Q4 2025. U.S. commercial revenue, the highest-growth segment, grew 121% year-over-year in Q3 2025. ROIC stands at an extraordinary 271%. The company carries almost no debt. Debt-to-equity is 0.03 and they hold over $5 billion in cash.
The one number that stops you cold: the stock trades at a P/E of roughly 238x at a price of about $152, against a 52-week range of $66 to $207. The software sector average P/E sits around 35 to 40x. Palantir trades at roughly 6x that.
§ 04 Risk Rating
Great business, but the risk is one thing: valuation. A 238x P/E means the market is pricing in perfection for years. If revenue growth decelerates for any reason, the stock re-rates hard. The business could be executing perfectly and the stock could still fall 40% if sentiment shifts.
Beyond valuation, there are real operating risks. Government budget cuts are a variable nobody controls. AI commoditization is moving faster than most people expected two years ago. And stock-based compensation has been diluting shareholders at a pace that does not show up clearly in the headline numbers.
§ 05 Bull vs. Bear
Bull case: Palantir is becoming the operating system for AI-era enterprises and governments. The ontology moat compounds with every new client. AIP Bootcamps have cracked the commercial sales problem and customer count is accelerating. Revenue is speeding up as the company gets bigger, which almost never happens at this scale. If they sustain 40 to 50% growth for three more years and margins keep expanding, today's multiple starts to look less extreme in hindsight. The government AI arms race, especially given the current geopolitical environment, could drive a decade of defense contract expansion that analysts are not fully pricing in.
Bear case: The stock already reflects every good thing that could happen. At 238x earnings, there is zero margin for error. The commercial moat is genuinely challenged. Databricks and Snowflake are improving fast, and hyperscalers can undercut on price while bundling AI tools into existing cloud contracts. Government revenue, while sticky, is concentrated and subject to political winds. If growth slows to 25 to 30%, the stock likely re-rates toward 80 to 100x, which implies a price around $55 to $70. That is a 55 to 65% drawdown from today.
Palantir is a world-class business built on a genuinely durable moat. The operational execution over the past two years has been remarkable. But the stock at $152 is not a business purchase, it is a bet on continued hypergrowth at a scale where hypergrowth gets harder. My price to get seriously interested: $90 to $100. That would still represent a roughly 120x forward earnings multiple, not cheap, but at least it leaves room for the thesis to play out.
§ 06 What to Watch
The number I am watching is U.S. commercial growth. If it decelerates below 50% for two consecutive quarters, something has changed. A major government contract cancellation or direct hit from defense budget cuts would be more serious. And if a hyperscaler manages to bundle a good-enough ontology alternative into existing cloud contracts at real scale, that would be the most important signal of all. Any one of those would make me rethink the thesis.
Going into this, I thought Palantir was just a government surveillance company, the kind of name you hear in conspiracy theories, not earnings calls. I was wrong about what they actually build, and that was the first lesson: never analyze a company based on its reputation before you understand its product.
The number that genuinely stopped me was the P/E ratio. 238x. I had to look up what that even meant in context because I had never seen a multiple that high on a profitable company. A P/E ratio by itself tells you almost nothing. You have to ask why the market is paying that premium and whether the growth rate actually justifies it. In Palantir's case, when revenue is accelerating at 60%+ and margins are expanding at the same time, the conversation changes. It does not make the valuation comfortable, but it at least makes it explainable.
The ontology concept took me the longest to understand. I had to read three different explanations before it clicked. The simple version: Palantir builds a digital map of how your entire organization works. Once that map exists, ripping it out and starting over is essentially impossible. That is the switching cost. That is the moat. It made me realize moats are not always obvious from the outside. Sometimes they are buried in the product architecture.
The government-to-commercial flywheel was the most interesting strategic insight. I always thought of government contracts as slow, boring revenue. What I did not see is that winning classified defense contracts is actually the best marketing a B2B software company can have. If the Pentagon trusts your platform with national security decisions, a Fortune 500 CFO has a much easier time saying yes.
The honest thing I still do not fully understand is how to value a company like this with confidence. The standard tools exist but Palantir sits in a category where they do not cleanly apply. A P/E of 238x is technically a number but it does not tell you much on its own. That is something I am going to keep working on.