PLTR IC Memo | IWANNAVY LAB

Executive Summary & Action Plan

Verdict

SELL (Reduce) — The best operating quarter in enterprise software (+85% revenue, 46% GAAP operating margin, NRR 150%) attached to a price that pre-pays two flawless years and then some: at 140x EV/EBITDA even our bull-case value ($111) sits below the $127 tape, and the last two blowout prints were both faded. The ontology moat is real; the moat-duration premium is not fundable.

Loading chart

Price $126.79 | Market cap ~$304B | Target $105 | Upside -17.2% IWANNAVY Fair Value · price as of 2026-07-11 close · Street consensus $183.12 (reference; n=27, 101% dispersion) — we are 43% below Street because we expense SBC, strip interest income from margins, and refuse to hold a 45x+ forward multiple through an arithmetic-certain deceleration

Loading chart

Abstract

Palantir's Q1 was close to operationally flawless: revenue grew 85% (US commercial +130% on AIP deal velocity — 47 deals over $10M in one quarter; US government +84% on Maven and the $10B-ceiling Army vehicle), gross margin reached 86.8%, GAAP operating margin 46.2%, and SBC fell to 12.4% of revenue — real leverage, fully expensed. Two flags temper the print: the 53.3% net margin carries ~8 points of non-operating support (interest on $7.8B net cash plus a ~$68M one-off equity gain), and TCV bookings (+61%) now grow slower than revenue (+85%) while FY26 guidance embeds deceleration every quarter. Meanwhile the forward-deployed model is being copied at hyperscaler scale — Microsoft's $2.5B/6,000-engineer Frontier Co. (7/2), Amazon's $1B commitment, OpenAI/Anthropic deployment ventures — and Databricks runs larger and faster in commercial data-AI at less than half the sales multiple. Our $105 fair value is the log-median of an SBC-expensed DCF ($61), 45x forward EBITDA ($105), and 45x forward FCF ($109); the multiples legs already sit at NVDA-peak territory. Rating: SELL/reduce ahead of the 8/10 print; re-enter at $105/$82.

Action Plan

History says even blowouts get faded at this multiple (2/2: +11% premarket → −13% same session; 5/4: +85% quarter → −7% next day). Reduce now and into any pre-print strength; re-enter $105 (FV) and $82 (compression zone).

Loading chart

New-money R/R: negative — spot exceeds bull FV ($111) | Prob-weighted 12M return −26.8% (Bull 30% × −12.5% + Base 45% × −17.2% + Bear 25% × −61.4%) | Confidence: Medium

Last Four Quarters

Every margin line expanded every quarter — the cleanest operating trend in our 18-name universe. Gross margin 80.8→86.8% as AIP productized; operating margin 26.8→46.2% with SBC falling from 17.6% to 12.4% of revenue (leverage is real, not add-back cosmetics). The net-margin line overstates it: Q1'26's 53.3% includes $66M interest income and a ~$68M one-off gain on private equity holdings — the durable line is the 46.2% GAAP OM.

Metric	Q2'25	Q3'25	Q4'25	Q1'26
Revenue ($B)	1.00	1.18	1.41	1.63
Gross margin	80.8%	82.5%	84.7%	86.8%
Op margin	26.8%	33.3%	40.9%	46.2%
FCF margin	53.0%	42.4%	54.3%	54.6%

Loading chart

Revenue & Profit Mix

Two US engines mask a flat third: US commercial (+130%) and US government (+84%) are 78.5% of revenue; international commercial grew just +26%. Government overall is still ~53% of the book. Q1'26 per the 5/4 8-K.

Loading chart

Business Lines

AIP / US commercial (36% of revenue, +130%): Land via bootcamps, expand to $5–10M+ production deals; US commercial RDV $4.92B (+112%). Must know — this segment carries the entire multiple, and it is exactly where MSFT Frontier Co., Amazon's FDE push, and Databricks Agent Bricks all aim.
Gotham / US government (42% of revenue, +84%): Maven, $10B-ceiling Army vehicle, IL5/IL6 accreditations. Must know — re-accelerating but lumpy and politically exposed; June's NGC2 award made Anduril prime with Palantir as data-architecture teammate — Palantir does not automatically prime every next-gen program.
Foundry / ontology (cross-segment): Must know — the ontology is the switching-cost engine and the honest rebuttal to the "consultancy" critique: 86.8% gross margin and 150% NRR are not services economics.
International (~21%, commercial +26%): Must know — the persistent hole: Europe's adoption gap means US demand carries everything; NATO eAirC2 (three-way bake-off vs Anduril and Athea) is the swing factor.

IWANNAVY Fair Value

Log-median lands at $105. The multiples legs ($105–109) already grant NVDA-peak forward multiples (45x) — holding them through an arithmetic-certain deceleration to ~45% FY27 growth is the generosity ceiling. The SBC-expensed DCF ($61) is what the cash flows support if the multiple normalizes. Our stale 4/30 mark ($157) is retired: the Q1 blowout raised the numbers, but the de-rate math raised faster.

Loading chart

Thesis breaker (for the SELL): an 8/10 print with revenue >$1.85B, US commercial holding ≥125–130%, a third FY26 raise toward $7.9–8.0B AND the market finally not fading it — sustained multiple support above 45x forward would force our multiples legs up; a NATO eAirC2 win and NGC2 dollarization add the defense leg.

Catalysts & Risks

Catalyst	Timing	Impact	·	Risk	Prob	Impact
Q2'26 print (cons. rev $1.81B, EPS $0.28; guide $1.797–1.801B; bar = 3rd FY raise toward $7.9–8B)	8/10 AMC (confirmed)	High (two-sided)	·	Deceleration math vs 140x: FY26 guide embeds Q4 exit ~55–60%; FY27 cons ~45%	~55%	35–40x FY27 FCF = $82–95 (−25–35%)
NGC2 dollarization (William Blair: ~$30M initial → $150M+/yr path)	Q3–Q4'26	Med (+)	·	Government concentration (~53% of revenue) into FY27 reallocation + midterms	~35%	−20–40% episodes (Jan'26 precedent)
NATO eAirC2 bake-off (vs Anduril, Athea) + AIPCon 11	H2'26	Positive drip	·	FDE model copied at scale (MSFT $2.5B/6,000 eng; AMZN $1B; OpenAI/Anthropic units) + Databricks larger & faster at <½ the multiple	~40% (24M)	US-commercial decel cuts the thesis segment

References

Palantir Q1'26 results — revenue +85%, US commercial +130%, FY26 guide raised (8-K, 2026-05-04) · Q1'26 10-Q (SEC) — SBC, interest income, one-off equity gain detail
CNBC — Microsoft's $2.5B, 6,000-engineer Frontier Co. (2026-07-02) · CNBC — Databricks $6.9B run-rate +80% (2026-06-16)
Defence Blog — NATO eAirC2 three-way selection (2026-07-07) · Foreign Policy Journal — NGC2: Anduril primes, Palantir data backbone (2026-06-27) · Benzinga — "the bounce is the trap" (2026-07-10)