Last Price
$153.37
Jeff the Financier · Institutional Equity Research · Initiation
The best-executing software company in the market today, attached to one of the most demanding valuations ever assigned to a $5B-revenue business. We rate the business an A+; we rate the entry price a HOLD.
Section 01
Palantir is, on the operating data, the single most impressive software franchise in the public market: Q1 2026 revenue of $1.633B grew +85% YoY — the fastest growth the company has ever printed and an almost unheard-of acceleration at a $6B+ run-rate — at a 60% adjusted operating margin and a company-reported Rule-of-40 score of 145. The bet, however, is not on the business; it is on the price of the business. At $153.37 the stock carries ~50x forward sales and a ~254x trailing EV/EBITDA, embedding a reverse-DCF-implied 38% five-year forward revenue CAGR against a sell-side consensus that tops out near 30%. Our probability-weighted intrinsic value is $109 (−29%). We initiate at HOLD, conviction 3/5: the franchise deserves a place on any AI-exposed book, but the asymmetry at today's entry is skewed to the downside (bear −74% / base −35% / bull +37%). The patient buy is a pullback toward the base-case $100–110 zone.
Per the situational-awareness thesis on a trillion-dollar compute build-out and a national-security mobilization around AGI, Palantir is one of the few public vehicles levered directly to the application and deployment layer of that build-out rather than the silicon. If the AGI-race framing is even partially correct, government AI spend compounds structurally and the bull case gains credibility — but that same frame warns that "AI enthusiasm" is reflexive, and a sentiment air-pocket compresses 50x-sales multiples first and fastest.
Section 02
Palantir sells two things that look like one: a data-integration/operational backbone (the Ontology, productized historically as Gotham for government and Foundry for commercial) and, since 2023, the Artificial Intelligence Platform (AIP) — the orchestration and governance layer that lets enterprises wire LLM agents into that ontology to take real actions. The reported KPIs in Q1'26 are, frankly, anomalous for a company of this scale.
Revenue compounded from $1.91B (FY22) → $2.23B (FY23, +17%) → $2.87B (FY24, +29%) → $4.48B (FY25, +56%) — i.e., growth accelerated through scale, which almost never happens. Management then guided FY2026 to $7.65B (+71%), with the US-commercial line guided "in excess of $3.224B" (≥+120%). The chart below shows the trajectory plus the base-case extension to FY31.
FY22–FY25 actuals (Yahoo Finance statements); FY26 = management guidance; FY31 = our base-case DCF (27% FY26→31 CAGR). Bars scaled to FY31 base case.
| Segment / cut | Q1'26 Rev | YoY | % of total | Read-through |
|---|---|---|---|---|
| US Commercial | $595M | +133% | 36% | The inflection engine |
| US Government | $687M | +84% | 42% | Moated annuity |
| US total | $1,282M | +104% | 79% | Crossed 100% growth |
| Intl Commercial | $179M | +26% | 11% | The soft spot |
| Rest of World | $220M | — | 13%→falling | De-mixing toward US |
| Total revenue | $1,633M | +85% | 100% | Beat $1.54B cons. |
Gross margin runs ~82%. The headline tension is GAAP vs. adjusted: Q1'26 adjusted operating margin was ~60% and adjusted FCF margin ~57% ($925M), but GAAP operating margin on a trailing basis is ~31.6% and GAAP FCF margin ~47% — the wedge is stock-based compensation. This is the single most important number-quality caveat in the whole report: we model on the cash/adjusted basis because this is a DCF, but every multiple looks ~10–15% worse on pure GAAP.
| Metric | GAAP (TTM) | Adjusted (Q1'26) | Comment |
|---|---|---|---|
| Gross margin | 82.4% | ~83% | Best-in-class software |
| Operating margin | 31.6% | ~60% | ~28pp SBC/amort. wedge |
| FCF margin | 46.9% | ~57% | Both genuinely high |
| Net margin | 36.3% | 53% (Q1 GAAP) | Boosted by interest income on $8B cash |
| ROE / ROA | 22.0% / 18.3% | — | Asset-light, net-cash |
Palantir has beaten its own top-line guide every quarter for two-plus years and raised the full-year guide alongside; Q1'26 revenue of $1.633B beat the ~$1.54B consensus by ~6%, adjusted EPS $0.33 beat ~$0.28 by ~18–22%, and the Q2'26 guide ($1.797–1.801B) was set ~7% above the Street. The beat-and-raise cadence is itself a core part of the bull reflexivity — and a core part of the risk, because the bar the company must clear ratchets higher each quarter.
Section 03
Palantir reports two segments — Government and Commercial — but they carry very different growth, durability and "should-be" multiples. A clean SOTP forces the question the blended 50x-sales headline obscures: what is each engine actually worth?
| Segment | FY26E Rev | EV/Sales | Implied EV | Rationale |
|---|---|---|---|---|
| Government | $2.7B | 14× | $38B | Annuity-like but budget-cyclical / lumpy |
| Commercial | $4.9B | 28× | $137B | Hyper-growth, AIP optionality |
| SOTP enterprise value | $7.65B | — | ≈$175B | + $5B net cash → ≈$180B equity |
| SOTP per share | ≈$73 | ÷ 2.45B shares | ||
| Current market price | $153.37 | Market premium to SOTP ≈ +110% | ||
Even on generous standalone multiples — 14x sales for government and 28x for the hyper-growth commercial engine (both rich versus the peer set) — the SOTP lands near $73/share. The market is paying a ~110% premium to that. That premium is not a "conglomerate" discount/premium in the classic sense; it is a platform-monopoly option premium — the market is pricing a meaningful probability that Ontology+AIP becomes the de-facto standard AI substrate, a winner-take-most outcome no segment-by-segment multiple can capture. Cross-segment synergy is real and underrated: government deployments (Maven, NATO) are an unusually credible lead-generation and proof-of-trust engine for regulated commercial buyers (banks, healthcare, utilities). The cannibalization risk is minimal; the dependency risk is that both engines lean on the same ontology thesis, so a single technological disruption (e.g., hyperscalers commoditizing the orchestration layer) hits both at once.
Section 04
The most useful valuation question for a stock like this is not "what is it worth?" but "what must be true for today's price to be fair?" We solve the DCF backwards.
Fixing a 35× exit EV/FCF, a 50% terminal FCF margin, 10% WACC, FY26 anchored at the guided $7.65B and 2.77B terminal shares, today's $376B equity value implies a 38.0% FY26→FY31 revenue CAGR → ~$38B of FY31 revenue (≈8.5× FY25). That is materially above the ~25–30% sell-side consensus ceiling.
This is a far more honest number than a naïve trailing-multiple reverse-solve (which spits out an absurd ~51% CAGR) precisely because the FY26 +71% step-up is already locked into guidance. The real debate is therefore narrow and specific: can Palantir compound at ~38% for five years off a $7.65B base? Q1'26 at +85% and 60% margins proves the current run-rate is above that line; the entire bear case is that this rate is a spike, not a plateau. Historically, only a tiny number of software companies (early Snowflake, peak Salesforce in the high-20s%) have sustained anything near this from a multi-billion base — and none did so while defending against hyperscaler bundling of the very orchestration layer at issue.
The prevailing narrative is an AI re-rating layered on a government-AI secular thesis: that Palantir is not an enterprise-SaaS comp at all but the "Oracle/ServiceNow of AI-native decision infrastructure," with a national-security tailwind that the situational-awareness frame would argue is only beginning. The skeptic narrative is pure multiple-compression mathematics: 50x forward sales has historically been a number you grow into over a half-decade, during which the stock can go sideways-to-down even as the business triples.
| Ticker | Price | Mkt Cap | EV/Sales | EV/EBITDA | P/E ttm | Fwd P/E | FCF yld | Rev gr. | Gross % | FCF mgn | Rule 40 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| PLTR | $153 | $376B | 81.7× | 254× | 243× | ~167× | 0.7% | +56% | 82.4% | 46.9% | 102.9 |
| NOW | $127 | $133B | 9.9× | 43.6× | 76× | ~54× | 3.4% | +20.9% | 77.5% | 34.1% | 55.0 |
| SNOW | $259 | $89B | 19.0× | n/m | n/m | ~85× | 1.3% | +29.2% | 67.2% | 23.9% | 53.0 |
| DDOG | $269 | $95B | 27.9× | n/m¹ | n/m | ~67× | 1.0% | +27.7% | 80.0% | 26.7% | 54.4 |
| CRWD | $760 | $193B | 39.1× | n/m¹ | n/m | ~93× | 0.6% | +21.7% | 74.7% | 25.8% | 47.5 |
| MDB | $397 | $32B | 12.6× | n/m | n/m | ~57× | 1.6% | +22.8% | 71.7% | 20.3% | 43.1 |
| Peer median | — | — | 19.0× | 43.6× | — | ~76× | — | ~23% | — | ~26% | ~54 |
¹ DDOG/CRWD GAAP EBITDA is near-zero (heavy SBC) so EV/EBITDA is not meaningful; on a non-GAAP EBITDA basis both sit ~55–70×. SNOW/MDB are GAAP-loss; n/m. Only PLTR and NOW are unambiguously GAAP-profitable.
PLTR trades at a +331% premium to the 19× peer-median EV/Sales and +8.3× the closest GAAP-profitable comp, ServiceNow (9.9×). The defense for the premium is real — PLTR's Rule-of-40 of 103 (TTM) / 145 (Q1 reported) is roughly double the peer cluster's ~54, and its GAAP profitability separates it from the SBC-masked loss-makers. But a ~2× Rule-of-40 advantage does not arithmetically justify a ~4–8× multiple. The gap is the option premium on the platform-monopoly outcome.
NOW is the most honest GAAP-to-GAAP yardstick. Both are genuinely profitable (NOW +13.7% / PLTR +31.6% GAAP op margin), both sell mission-critical, workflow-embedded platforms into large enterprise + government with high switching costs and structural net expansion. The difference is positioning and price: NOW compounds ~21% on a $13B base at 9.9× sales and 54× forward earnings — priced for steady compounding; PLTR is priced as if dominating AI-decision infrastructure is already underwritten, at 8× the sales multiple and ~3× the Rule-of-40. SNOW is the more common "data-platform" comp, but its GAAP losses and negative operating leverage make it a dirtier fundamental match than NOW.
Section 05
We build three six-year explicit forecasts (FY2026–FY2031) anchored on the guided FY26 revenue of $7.65B, ramp FCF margin linearly from ~56% toward each scenario's terminal margin, discount the FCF strip plus a PV-of-terminal value (exit EV/FCF multiple as primary method, Gordon-growth as a conservative cross-check), add ~$5B net cash, and divide by ~2.77B FY31 diluted shares (2.5%/yr SBC dilution). WACC is set at 10% base (β≈1.4 → CAPM cost of equity ~11.4%, adjusted down for the $8B cash fortress and >$800M/quarter FCF), 11% bear, 9.5% bull.
| Valuation bridge ($B unless noted) | Bear | Base | Bull |
|---|---|---|---|
| Sum PV of explicit FCFs (FY26–31) | $23.0 | $31.6 | $41.3 |
| Terminal EV (FCF × exit mult.) | $155.0 | $424.6 | $925.4 |
| PV of terminal EV | $82.9 | $239.7 | $536.8 |
| + Net cash | $5.0 | $5.0 | $5.0 |
| Equity value | $110.9 | $276.3 | $583.1 |
| Value / share | $39.99 | $99.68 | $210.35 |
| Upside / downside vs $153.37 | −73.9% | −35.0% | +37.2% |
(0.30 × $39.99) + (0.45 × $99.68) + (0.25 × $210.35) = $109.44 → −28.6% vs $153.37. Note the weighted value sits below all three mid-scenario points because the bear tail is severe: at 50x sales, deceleration doesn't dent the stock — it halves it.
The picture is stark: the current $153 print sits to the right of our probability-weighted fair value and just left of the Street's $193 — i.e., the market and consensus are both anchored near our bull case, not our base. Only the bull scenario, requiring a sustained 36% CAGR to $35.6B of FY31 revenue, returns positive value from here. That is the definition of a price that has run ahead of even a generous fundamental case.
Section 06
Probability/impact on a Low/Medium/High scale. The dominant, correlated risk pair is multiple compression × growth deceleration — they are not independent, and they fire together in a bear case.
Section 07
HOLD · conviction 3/5. 12-month tactical target band $95 / $150 / $215 (bear/base/bull), with a base-case 12-month target of ~$150 (roughly flat) that intentionally sits above our $109 DCF intrinsic value: momentum names with beat-and-raise reflexivity persist at elevated multiples far longer than a DCF allows, and shorting quality on valuation alone is a career-ending trade. We are not bearish on the company; we are disciplined on the entry.
For a diversified institutional equity sleeve: 0.5–1.0% starter as an AI-thematic call-option, scaling to a core 2.5–3.5% only on a pullback into the $90–110 base-case zone. Above $190, trim back toward the starter weight.
Section 08
Section 09