Palantir: Brilliant Business, Brutal Valuation (Downgrade)

I've been watching Palantir all year with a divided mind. The stock fell roughly 30% in 2026 as the broader software sector got caught in fears that AI agents would automate the very tasks enterprise SaaS was built to handle. Then on August 3, Palantir reported second-quarter results that made the selloff look inexplicable: revenue of $1.94 billion, up 93% year-over-year. U.S. commercial revenue surged 149%. Earnings per share came in at $0.41 against a consensus of $0.35. Management raised full-year guidance to $8.15 billion–$8.16 billion from $7.65 billion–$7.66 billion. And the stock responded by jumping roughly 29% in one session, erasing most of the year's decline.

If you're looking for the reason the business should work, the quarter delivered it in full. The question I've been wrestling with is whether the reason the stock has been punished — and the reason it just bounced — is the same one. And the answer is no. The selloff was about sector sentiment. The rally was about earnings excitement. Neither one was really about the number that should dominate every Palantir investment decision: the price you're paying for what is already extraordinary.

The Valuation That Nothing Can Fix

Palantir trades at a forward P/E of 382x. The trailing P/E is 137x. The price-to-sales ratio sits at 67x. For comparison, even CrowdStrike — another software darling with its own premium — trades at 43x sales. Datadog is at 21x. Airbnb, which is far less growth-heavy, is at 8x.

You can't dismiss these multiples by pointing to the growth rate alone. Palantir's revenue is growing at roughly 79% year-over-year, with free cash flow margins of 51% and operating margins near 43%. Those are the best kind of numbers: high growth with high profitability, which means the company isn't burning cash to achieve the trajectory. Free cash flow for the trailing twelve months came in at $3.36 billion, growing 97% year-over-year. Return on invested capital is 32.6%.

The problem isn't that Palantir isn't growing. The problem is that a forward P/E of 382x prices in a world where this growth never meaningfully slows and margins never meaningfully contract. Not for one quarter, not for one year, not for the five-year window the market implicitly demands before the multiple becomes defensible. At 382x, even a deceleration from 93% to 50% — still an extraordinary growth rate for any public company — would crush the stock. The market has priced in perfection, and perfection is a standard that compounding itself makes harder to meet.

Michael Burry — the investor who correctly called the 2008 housing crash and has been betting against Palantir since late 2025 — puts a number on this. In public posts, Burry has said he values Palantir's fundamentals at well under $50 per share. At the current price of roughly $172, that's an enormous gap. Burry has called the stock a "sand castle supported for now by the AI applications narrative." His $50 estimate is likely extreme — the business is clearly worth far more than a fraction of its current revenue. But the underlying diagnosis stands: the valuation has detached from any scenario where growth normalizes even partially.

The Moat Is Real. That Doesn't Make the Price Right.

One of my first checks before calling any stock overvalued is whether the competitive position can justify the premium. With Palantir, this is where the business actually shines.

The U.S. government side — which generated $809 million in Q2, up 90% — benefits from what amounts to a structural monopoly. The Department of Defense has a historical preference for a single primary supplier on mission-critical data infrastructure, and Palantir's Gotham platform has become that supplier. Displacing it would require rebuilding years of classified integrations, domain-specific workflows, and security clearances. That's not a product switch; it's a political and operational undertaking.

On the commercial side, the 149% growth is equally telling about the moat. Microsoft tried building its own enterprise AI analytics platform called Fabric and ultimately pulled back to pursue a partnership with Palantir instead of competing directly. Microsoft has every incentive to sell to every enterprise customer on earth. If Microsoft couldn't see a path to displacing Palantir, smaller competitors face an even steeper wall. The U.S. commercial remaining deal value more than doubled year-over-year to $6.24 billion, and bookings surged 153%.

Net revenue retention above 120% means existing customers are expanding their deployments, not just renewing. The AIP bootcamps are oversubscribed with waiting lists extending months into the future. These aren't vanity metrics — they're evidence of a land-and-expand cycle that's accelerating, not decelerating.

The moat is durable. I don't doubt it. That's what makes the valuation gap so frustrating. A business this good at this price isn't an investment opportunity; it's a bet that nothing goes wrong for years.

The Insider Signal and the Governance Noise

Two additional factors reinforce caution. First, insider activity paints a picture that's hard to reconcile with the narrative that the stock is headed significantly higher from here. Over the trailing five years, Palantir insiders have executed roughly $7.2 billion in net stock selling. Since the company's direct listing in September 2020, there has been only one recorded insider purchase, totaling approximately $1.16 million. Not every insider sale signals overvaluation — compensation-driven tax selling is real and mechanical. But a complete absence of buying over six years is a data point worth respecting.

Second, governance concerns have surfaced. A $17.2 million CEO private-jet reimbursement sparked investor backlash earlier this year, and the company faces a legal dispute with AI startup Percepta over alleged misuse of confidential information. Neither of these breaks the investment thesis. But together with the valuation and the insider selling, they add layers of friction for an already stretched stock.

Price Action: The Bounce Is the Problem

The technical picture complicates the entry further. Palantir's RSI-14 sits at 72.8, well into overbought territory. The stock closed near $172 after the earnings pop, up roughly 40% over the past five trading days. It sits above both the 50-day moving average ($132.60) and the 200-day moving average ($152.28), but the velocity of the move — 29% in one session followed by continued upside — is the kind of reaction that creates worse risk/reward, not better.

The stock's 52-week range runs from $106 to $208. At $172, it's roughly 68% of the way up from the low and still 17% below the high. The prior decline to the $125–$130 zone before Q2 earnings was closer to what I'd consider a defensible entry. The current level is the result of a post-earnings reflexive pop that has moved the goalposts on risk/reward.

What This Means for the Thesis

AInvest's aggregate signal labels Palantir a Buy, with a fundamental rating of 9.53 out of 10. The numbers justify the fundamental score. Revenue growth, margin expansion, cash flow conversion, and competitive positioning are all at elite levels. CEO Alex Karp told investors that the current growth trajectory is expected to continue for "at least another 18 months," and the Q2 print gives him the receipts to make that claim.

But the fundamental grade and the investment thesis are two different things. You can acknowledge that Palantir is a phenomenal business and still conclude that 382x forward earnings is a number that turns even modest disappointments into multi-month drawdowns. The forward P/E tells you how much future perfection the market requires. It's not a prediction; it's a demand. And the demand right now is unsustainable.

My view: Hold current positions. Don't add. Wait for a pullback to the $120–$135 zone before considering new entries. The business has earned the right to be in your portfolio. It hasn't earned the right to be bought at this price. The 30% decline in early 2026 was the market's first honest reaction to the valuation. The 29% earnings-day pop was the market's excitement overriding it. Neither move was the final word, and the final word on Palantir's investment case won't come until the price reflects a growth rate that's high without being impossible.

I'd reassess the stance if the stock pulls back into the $120–$135 range with the growth trajectory intact, or if the revenue base compounds to the point where the forward multiple compresses below 200x organically. Until then, the catch Burry identified is real — and it's the one that matters most.