Palantir's 'Otherworldly' Quarter Is Real. The Valuation Isn't Done Shrinking. (Don't Chase)

Palantir CEO Alex Karp called the quarter "otherworldly." For once, I'm not going to argue with the management team's chosen adjective.

Revenue surged 93% year-over-year to $1.94 billion, well above the roughly $1.80 billion consensus expected. Adjusted EPS of $0.41 cleared the 33–35 cent range. The U.S. commercial segment - the part of the business Wall Street has been watching for proof that Palantir isn't just a government contractor - grew 149% to $764 million. GAAP operating income hit $912 million, a 47% margin. Adjusted free cash flow was $1.22 billion, or 63% of revenue.

Then management raised full-year revenue guidance to $8.15–$8.16 billion, up from $7.65–$7.66 billion. That implies roughly 82% growth for the full year. U.S. commercial guidance was lifted to over $3.42 billion, representing at least 134% year-over-year growth. The remaining U.S. commercial deal pipeline - revenue contracted but not yet recognized - sits at $6.24 billion, up 124% from a year ago.

The numbers are not a mirage. They are the kind of quarter that makes you question whether the market's been asleep at the wheel.

But here's the problem I keep coming back to: Palantir stock is down 29% this year from its 52-week high of $207.52. As of Monday's close it was at $125.65. That drawdown has done some work on valuation, but it hasn't done nearly enough to make the risk/reward obvious.

The forward P/E - the ratio of today's price to next year's expected earnings - still sits at 278x. That sounds absurd, and it is, unless you have a very specific view about how fast earnings compound over the next two years. Even the PEG ratio (forward P/E divided by the earnings growth rate, a rough gauge of whether a multiple is justified by growth speed) comes in at 0.46, which is technically attractive but only because the growth denominator is so enormous. A sub-1.0 PEG implies the market is underpaying for growth, but it doesn't tell you what happens if that growth rate normalizes from 93% to something more ordinary. And for a company this size, at $301 billion in market cap, it almost certainly will.

I'm not saying Palantir's thesis has cracked. I'm saying the market's price reaction to this quarter tells us the valuation hasn't fully collapsed, and the downtrend hasn't been reversed either.

The moat is intact - that's the easy part

Palantir's competitive position has gotten stronger, not weaker, over the past year. The company's AIP (Artificial Intelligence Platform) has become the operating system for what it calls "sovereign AI" - meaning enterprises and governments that want to run AI workloads on their own infrastructure, with full data control, rather than sending sensitive information to cloud providers or model vendors. That positioning is a moat because it takes years to build the trust, integrations, and enterprise relationships that make switching costs meaningful.

The U.S. commercial TCV (total contract value, or the total value of deals signed in the quarter) hit $2.13 billion, up 153% year-over-year. Ninety-eight deals were at least $5 million each. Seventy-three were at least $10 million. This isn't a startup selling pilot projects. This is a company closing multi-million-dollar contracts with enterprises that have decided to build their AI operations on Palantir's platform.

Government revenue grew 90% to $809 million, so the legacy base is still expanding too. The 115% growth in total U.S. revenue - commercial plus government - is the number that matters most for durability. Both engines are firing.

The competitive moat survives the stress because the stress, such as it was, came from valuation pressure and macro skepticism about the AI trade, not from customers leaving or competitors eating market share.

What the price action says the market hasn't decided

This is where the contrarian instinct needs a reality check. The stock is trading below its 50-day moving average ($130.49) and its 200-day moving average ($152.62). The RSI - a momentum oscillator that measures whether a stock is overbought or oversold on a 0-to-100 scale - sits at 47.7, roughly neutral. The MACD, a trend-following indicator, is negative at -1.53. The stock is up 2.1% on the day but down 4.5% over five days, 5.2% over 20 days, and 12% over the past four months.

In plain English: the trend is still down. The after-hours pop on Q2 results hasn't broken the structure. A stock that's lost nearly a third of its peak value doesn't flip from "fallen knife" to "confirmed bottom" on one earnings night, even a very good one.

I don't think the thesis is broken. But I do think the market is telling us something: it doesn't believe the 278x forward multiple is justified yet, even for this growth rate. And I'm inclined to listen.

The GARP test

GARP - growth at a reasonable price - is the framework I use to separate genuinely attractive setups from expensive momentum traps. The test has three gates: fundamental quality, valuation relative to growth, and moat durability.

Palantir passes the quality and moat gates easily. Revenue growth at 93%, gross margins above 84%, operating margins above 42%, ROIC at 32.6%, ROE at 38.4%, $9.2 billion in cash and short-term Treasuries, and essentially no debt - this is a machine. The balance sheet gives management enormous optionality. A $1.22 billion adjusted free cash flow quarter means the business is generating cash faster than most companies half its size generate revenue.

The valuation gate is where it gets complicated. A forward P/E of 278x is not a number that gets more comfortable with time. It gets more comfortable only if earnings accelerate faster than the market expects. If Palantir hits the mid-point of its $8.15 billion guidance and maintains something close to these margins, full-year earnings would still leave the stock at a very high multiple. The PEG of 0.46 looks cheap, but the PEG ratio is a backward-looking lens applied to a forward-looking beast. It works best when you're confident the growth rate will persist for several years. At $301 billion in market cap, that confidence needs to be earned, not assumed.

The setup

Here's how I see it. Palantir is one of the most compelling businesses in the AI infrastructure stack. The commercial acceleration is real, not a narrative. The deal pipeline is deep. The moat is wide. Karp's "sovereign AI" positioning - enterprises owning their AI infrastructure rather than renting it - is the right long-term bet in an era of data regulation, national security concerns, and competitive anxiety.

But the stock hasn't fallen far enough to make the entry obvious. A 29% YTD decline from a peak that already reflected enormous optimism is a haircut, not a discount. The technical structure hasn't confirmed a bottom. The valuation, while compressed from its recent peak, still demands multi-year hypergrowth to look defensible.

I'm not selling here. I'm not buying here either - at least not in any size that would keep me up at night. The better risk/reward is likely on a deeper pullback toward the $106–$110 zone (near the 52-week low of $106.37) or after the stock closes above its 50-day moving average on volume, signaling that the downtrend has actually broken. If it does, the GARP case becomes significantly more compelling.

If you already own Palantir at better averages, holding is the call. The fundamentals are strong enough to justify patience. If you've been waiting on the sidelines, I'd say the same thing. The quarter was otherworldly, as Karp put it. But the setup for a lower-risk entry hasn't formed yet.

I would reassess this view if the stock cracks below $100 - that would create a genuine contrarian entry zone for a company with this cash flow and this growth. I would also reassess if guidance turns cautious or if the U.S. commercial growth rate falls below 100% for two consecutive quarters, because that's where the moat narrative starts to face real questions.

Until then, patience. The best growth setups reward the investor who can wait for the price to meet the thesis, not the other way around.