PLTR — Q4 FY25 Earnings Review

StockNovice (Joe) | February 2026 First look — no prior Joe analysis on file


The Headline

Okay. Sometimes a quarter is so loud you can't ignore it. Palantir printed $1.407B in Q4, up 70% year-over-year at a nearly $6B annual revenue run rate. U.S. Commercial grew 137%. TCV jumped 138%. Adj operating margin hit 57%. Rule of 40: 127%.

By almost any growth metric, this is the best quarter a large-cap software company has put up in years.

Atlas pegged it at Conviction 3 — "extraordinary business at an extraordinary price." That's about right. My job here is to figure out what Joe would actually do with it.


"Is" vs "Could Be"

This one is clearly "is." The acceleration is the story:

Quarter Revenue YoY Growth Adj Op Margin
Q1 FY25 $884M 39.4% 19.9%
Q2 FY25 $1,004M 48.0% 26.8%
Q3 FY25 $1,181M 62.8% 33.3%
Q4 FY25 $1,407M 70.0% 57.0%

Revenue growth is accelerating at $1.4B quarterly scale. That almost never happens. Usually companies decelerate as they get bigger — law of large numbers. Palantir is doing the opposite.

The AIP bootcamp model is clearly working. They're converting marketing spend into signed deals at a pace that's showing up in TCV (4.26B, +1384.38B, +145% YoY). That RDV represents 3+ quarters of U.S. Commercial forward revenue visibility. Leading indicators are running 2x ahead of reported revenue.

So: the business is delivering. This isn't a "could be" story anymore.


What I Love

U.S. Commercial approaching Government parity. $507M Commercial vs $570M Government in Q4. Commercial was a distant second 18 months ago. This matters because commercial contracts are theoretically more scalable and less politically exposed than government. If AIP keeps compounding, commercial could overtake government in FY26. That's a structural shift.

180 deals ≥1M.61deals10M. The AIP bootcamp flywheel converts tire-kickers into signed contracts. High-velocity deal closing at enterprise scale — that combination is rare.

Margin inflection is real. Going from 33% to 57% adj operating margin in a single quarter is extraordinary. Even if there's some quarterly lumpiness, the direction is unmistakable. FCF at $791M (56% margin) confirms it's not just accounting.

$7.2B cash, zero debt. The balance sheet is a fortress. They're not burning cash, they're printing it.


What Makes Me Nervous

The valuation. I have to be honest. 70x EV/TTM revenue. ~200x GAAP P/E. Even on FY26 guidance ($7.18B revenue), you're paying 44x forward EV/revenue and 78x forward EV/FCF.

For context: elite, fastest-growing software companies typically trade 15-20x EV/revenue. Palantir is trading at 3-4x that premium. You're not paying for growth — you're paying for the story around growth being permanent and enormous.

The NRR black box. Palantir doesn't disclose NRR or customer count with meaningful granularity. For a company where expansion within existing customers is presumably a key driver, that's a material blind spot. I can infer from RDV growth that expansion is happening, but I'd like to see it directly.

SBC at 14% of revenue. $640M in FY25 SBC on $4.47B revenue. That's real dilution. GAAP EPS and GAAP margins look better than they should because of how SBC gets treated in adj figures. The 57% adj margin is impressive; the GAAP picture (43% net margin but inflated by non-operating items) is more complex.

Government concentration risk. 41% of revenue (and growing QoQ in absolute dollars). DOGE-related budget pressure, shifting political winds — this isn't academic. If government contracts get restructured or delayed, that's a meaningful hit to near-term numbers.

Karp's language. "n of 1 company." "Commodity cognition." He's making bold claims at the moment of peak confidence. That language pattern is one I always watch. It's not a red flag yet, but it's a yellow flag to monitor.


The FY26 Setup

Guidance: 7.18 − 7.20Brevenue(+613.144B (+115%), adj FCF $3.9-4.1B.

Q1 FY26 guide: $1.532-1.536B. That's +9% QoQ off a monster Q4. The beat-and-raise pattern has been consistent — every quarter they've guided low and delivered high. If that holds in FY26, we could see $7.5-8B for the year.

That U.S. Commercial guide of >$3.144B implies they exit FY26 at a $1B+/quarter commercial run rate. Government + International presumably $4B combined. This company is on a trajectory to be a $10B+ revenue business by FY27-28 if it holds pace.


Prior Beliefs → Updated Beliefs

Prior: No prior Joe view on file. Palantir was always "too expensive" in my mental model — a government contractor learning to be a software company, great product but priced for perfection.

Updated: The AIP commercial inflection is real and data-confirmed. This is no longer primarily a government contractor — it's becoming a commercial AI platform at scale, with the government business as a stable, high-margin base. That changes the narrative meaningfully.

The product moat (ontology layer + AIP) appears real based on customer retention signals and the bootcamp model's velocity. Gartner ratings are high. Competitive alternatives don't exist at this layer.

What I was wrong about: The commercial acceleration arriving at this speed and this scale. I'd have expected some deceleration by now.

What I still worry about: Whether 70% growth is durable at $7B+ annual revenue, and whether the current valuation already prices in 5-7 years of compounding.


Position Sizing Verdict

Status: Bench / Watchlist. Not a current starter or tryout.

Here's how I'm thinking about it: the business quality is elite — maybe top decile of anything I'd consider. The results are clearly "is" not "could be." But paying 70x revenue for even an extraordinary business is asking a lot. The market has priced in several years of continued acceleration at a premium multiple.

For it to earn a tryout spot, I'd want to see either:

  1. Multiple compression to somewhere in the 30-40x TTM revenue range (roughly 40-50% stock decline from current), OR
  2. Revenue acceleration sustaining at 70%+ for 2 more quarters while the multiple stays flat, allowing growth to "earn into" the valuation, OR
  3. NRR disclosure confirming the expansion engine is as strong as TCV/RDV implies

The stock is already down ~27% from its January highs. That's movement in the right direction. But 70x EV/revenue is still 70x — it would need to come down substantially more before the risk/reward works for me.

I don't need to own everything that's working. Passing on an expensive great business is a legitimate choice.


Key Metrics Table

Metric Q4 FY25 Q3 FY25 YoY
Revenue $1,407M $1,181M +70%
U.S. Commercial $507M $381M +137%
U.S. Government $570M $498M +66%
Adj Op Margin 57% 33% +37pp
FCF Margin 56%
TCV $4,262M +138%
RDV (U.S. Comm.) $4,380M $3,617M +145%
Rule of 40 127%
Cash $7.2B
SBC (FY25) $640M ~14% rev

FY26 Guidance

Metric Guide Implied Growth
Revenue $7.182-7.198B +61% YoY
U.S. Commercial >$3.144B >+115% YoY
Adj Op Income $4.126-4.142B ~57% margin
Adj FCF $3.925-4.125B ~55% margin

Valuation Context

Metric Current Peer Median
EV/TTM Rev ~70x 15-20x
EV/FY26E Rev ~44x 10-15x
EV/TTM FCF ~141x 40-50x
EV/FY26E FCF ~78x 30-40x
GAAP P/E ~200x 50-80x

Multiple compression from here doesn't require bad news — it just requires growth to slow from 70% to, say, 50%. That's still elite, but the stock could drop 30-40% in that scenario.


Conclusion

Palantir is the real deal. The business is delivering at a level that would earn a starter allocation in virtually any other portfolio I'd consider. The AIP commercial flywheel is working. The margins are inflecting. The leading indicators (TCV, RDV) are running hard.

But I've learned the hard way that paying extreme multiples for even extraordinary businesses creates fragility. One quarter of deceleration — from 70% growth to, say, 45% growth — and a 70x multiple compresses fast. You can be right about the business and still lose money badly.

For now: bench. Watch. Admire. If the valuation comes in meaningfully, this is a company I want to own.

As usual, thanks for reading.


Analysis Date: 2026-02-22 Quarter: Q4 FY25 (ended December 31, 2025) Source: Scout brief at briefs/PLTR_earnings-review_2026-02-21/, Atlas analysis Transcript: Unavailable (Motley Fool/SA access issue — 95% complete data otherwise)