Snapshot: Executive Summary
Palantir is two things at once, and an honest note has to say both. As a business, it may be the best in software: in Q2 2026 it grew revenue +93% YoY to $1.935B (+19% sequentially and still accelerating), at a ~155% Rule of 40, a 62% record adjusted operating margin, a 63% record adjusted FCF margin, 157% net dollar retention, and a first-class GAAP bottom line (Q2 GAAP operating income $912M, +239% YoY)1. US commercial, the AIP proof point, grew +149% to a record $764M. On the operating metrics, PLTR is in a class of one.
As a stock, it is one of the most expensive names in the market. After the Aug 3 blowout re-rated it ~+48% off the June low (still −3% YTD and −17% from the November 2025 high), PLTR trades at ~40× forward sales, ~98× forward earnings and ~105× trailing FCF, roughly 4× software norms. Our SaaS DCF, revenue × FCF margin, run on the raised guide, now lands a base intrinsic near ~$175/share, right at the ~$172 price. The re-rate closed the old DCF-below-price gap; that the DCF prints "fairly valued" is the expected result. So the case is not about the model; it is about what the price now requires:
- The reverse DCF is the real lens. At ~$172 the market implies only a ~5% perpetual terminal (modestly above the GDP norm), but layered on top of an already-aggressive ~50% base revenue CAGR and a mid-60s% terminal FCF margin. Charge the ~12%-of-revenue stock comp, or use a normal 3% terminal, and our own intrinsic drops back into the low-$130s to mid-$140s, below the price. The margin of safety is gone.
- The business grows into the price, but no further. Revenue compounds ~50% (FY26–30 CAGR), and per-share intrinsic, net of the ~12%-of-revenue stock comp, now rolls the ~$175 base to ~$175 in a year, in line with the price. That closes the discount that made this a lean-Reduce; it does not open one that would make it a Buy.
- Deceleration is mathematically normal. The raised guidance itself steps growth down from 93% to ~82% for FY26, and the law of large numbers takes it lower as the base scales past $10B. The bull needs the hyper-growth to persist longer than priced; the bear only needs it to normalize toward the 30s%.
We rate PLTR a Hold, 12-month PT $175 (bull $215 / bear $115), a genuinely balanced call now. The probability-weighted intrinsic today is ~$164 (≈4% below the price), the blended target is ~$167, and the base DCF ≈ the price: everything clusters at spot. We would Sell / trim for valuation-sensitive mandates and would not deploy new capital at fair value, but we do not sell the best franchise in software into strength. Downgrade-to-Sell trigger: US-commercial growth decelerating below the raised >134% guide, or a re-rating well above the ~$200 Street average without a commensurate fundamental step-up. Upgrade-to-Buy trigger: a pullback that restores a double-digit margin of safety. The risk is the price, not the business.
Tactical: PLTR at ~$172 trades roughly in line with our $175 12-month target, a Hold at fair value. The business is elite (Rule of 40 ~155%, 63% FCF margin), but ~40× forward sales leaves no margin of safety, and the probability-weighted intrinsic (~$164) sits just below the price. Swing factors: US-commercial durability and the multiple.
Investment Thesis
Bull Case
- US commercial holds >100% growth longer than priced; AIP becomes the default AI-application layer; the dual engine compounds for years
- Margins keep expanding (Rule of 40 stays well above 100%); FCF margin holds the mid-60s%
- The multiple holds as the durability de-risks; the government backlog (Army EA, NGC2) converts
- DCF at a ~9.5% WACC / 5.0% terminal on the bull cash flows (~$46B FY30 revenue) supports a ~$214 intrinsic; the $215 target adds a year of growth
Base Case
- Revenue ~$8.15B FY26 (+82%) decelerating to ~$42.2B by FY30 (~50% FY26–30 CAGR); FCF margin rises into the mid-60s%
- DCF intrinsic ~$175 today at ~10% WACC / 5% terminal; the per-share roll (net of dilution) keeps it ~$175, in line with spot
- The rich multiple holds roughly flat as growth stays hot; the base earns its price, but opens no margin of safety
- Net dollar retention stays ~150%+; the backlog (RDV $11.8B) underwrites the near-term line
Bear Case
- US-commercial growth normalizes toward the 30s% as the base scales; the "above 134%" guide slips
- The ~40× multiple compresses toward software norms (~15× would still be premium); SBC dilution is recognized as real
- Government revenue (~42% of the mix) hits a budget / continuing-resolution / political air-pocket
- DCF at ~10.5% WACC / 4% terminal on a slower path (~$36B FY30 revenue) re-tests ~$107 intrinsic; the $115 target assumes a less-severe glide
Rating: Hold, and now a genuinely balanced one. The probability-weighted blend of the 12-month targets is ~$167 and the probability-weighted intrinsic value today is ~$164, both essentially at the ~$172 spot, with the base DCF ~$175 right alongside. The blowout Q2 lifted intrinsic to meet the price, so the old downward skew has closed: this is no longer a lean-Reduce, but neither is it a Buy, because the re-rate left no margin of safety and a Buy would need heroic terminal assumptions. New capital can wait for a pullback. Downgrade-to-Sell trigger: US-commercial growth below the >134% guide, or a re-rating well above the ~$200 Street average without a fundamental step-up.
The Franchise: Ontology + AIP
Palantir sells two products over one foundation. Gotham serves government/defense; Foundry (with AIP, the Artificial Intelligence Platform) serves commercial enterprises. The foundation is the ontology: a living digital twin of an organization that maps its messy data to real-world objects (a plane, a patient, a shipment) and the actions that can be taken on them. AIP is the governance + orchestration layer that lets large language models operate on that ontology as auditable, permissioned agents. The pitch, in the CTO's words: LLMs are the brains; the ontology is the body, "agents go nowhere without it."
Revenue by segment (Q2 2026)
A dual engine: US Government 42% + US Commercial 40% are ~82% of revenue; the two international lines are the rest. Total US revenue grew +115% YoY in Q2 to $1.573B, US commercial nearly caught government.
Annual revenue ($B)
From $1.1B (FY20) to $4.48B (FY25) and a raised ~$8.15B FY26E (+82%): the AI super-cycle inflected the curve. The question the DCF answers is what that curve is worth.
Why the moat holds (and where it is debated)
- The ontology is sticky. Once an enterprise models its operations in Foundry and wires AIP agents into live workflows, switching means rebuilding the digital twin, high switching costs, reflected in 150% net retention.
- AIP is a governance layer, not a model. As inference gets cheaper (Jevons' paradox), more AI gets deployed and the need for permissioning, lineage and audit grows, which is what PLTR sells. It is a bet on AI operationalization, not on any one LLM.
- The debate: services vs software. Bears argue the boot-camp / forward-deployed-engineer motion is consulting in disguise and won't scale at software margins. The counter is the ~85% gross margin and the 63% record FCF margin, which look like software, not services.
AIP & US Commercial: The Engine
If there is a single number that drives the bull case, it is US commercial revenue: a record $764M in Q2 2026, +149% YoY, and +28% sequentially. US-commercial customers reached 615+ (+42% YoY as of Q1), and management raised the FY2026 US-commercial guide again, to >$3.424B (~+134%). The motion is the AIP boot camp: a hands-on workshop that takes a customer from pilot to a production use-case in days to weeks rather than months, Heineken (3 months vs a typical 3 years), Walgreens (~4,000 stores in 8 months).
US commercial revenue ($B)
From ~$0.7B (FY24) to a raised >$3.4B FY26E: the AIP engine more than doubling YoY. This is the line that has to keep compounding to justify the multiple.
Backlog / deal value ($B)
Total RDV $11.8B (+98% YoY); US-commercial RDV $4.92B (+112%); total RPO $4.45B; Q1 closed TCV $2.41B (+61%). The backlog nearly doubling is the bull's "durability" evidence.
The customer wins that anchor it
Recent expansions made the AIP story concrete: GE Aerospace (agentic AI into production + military-aviation supply chain), AIG (multi-agent underwriting/claims, targeting a 10%→20% growth re-acceleration), and an Airbus Skywise multi-year extension. The bull reads the boot-camp-to-production conversion as a repeatable, widening funnel; the bear watches for the day the >134% US-commercial growth rate decelerates, because that single line carries the valuation.
The funnel kept widening at AIPCon 10 (June 4, 2026): new or expanded customers including McCarthy Building, Stellantis, Bain, and GNP Seguros (the first public Mexico commercial customer), plus a Google Cloud partnership (Foundry on the Google Cloud Marketplace, two-way BigQuery↔Foundry federation, deeper Gemini↔AIP connectivity) that puts Foundry on all four major clouds. A Zeta Global strategic partnership followed (June 23, Zeta re-architecting its data cloud on Foundry; note the headline >$100M figure accrues to Zeta, not Palantir). None carry a disclosed contract value, so they are GTM/distribution signals, not modeled revenue. The most thesis-relevant launch is the general availability (March 2026) of the AI Forward-Deployed Engineer: an AIP agent that builds and operates Foundry pipelines and ontologies in natural language. If it scales, it directly answers the "services-in-disguise" bear by automating the deployment motion toward software margins.
Government & Defense
The other engine kept accelerating: US government revenue $809M in Q2 2026, +90% YoY and +18% sequentially. Government is ~42% of total revenue (its share slipping only because US commercial is growing even faster), which makes PLTR a high-beta proxy for the federal-AI modernization wave, and a concentrated bet on appropriations. The backlog here is anchored by ceilings that are large but not booked revenue:
Key government contract ceilings ($B)
Ceilings, not bookings. Actual revenue depends on task orders drawn against each vehicle, the single most common way government-software backlog is overstated.
- Army Enterprise Agreement: up to a $10B ceiling over 10 years (Aug 2025), consolidating 75 contracts into one; the largest deal in PLTR history.3
- Maven Smart System: ceiling ~$1.3B through 2029; 20,000+ active users, more than doubled since Jan 2025. Now a formal DoD program of record (Mar 2026 SecDef memo), transitioning onto the Army EA vehicle by end-FY26, converting a special initiative into durable, funded multi-year revenue.
- TITAN: $178M for 10 prototype ground systems; an interim transition milestone falls in FY2026, but the full-rate production decision + initial fielding (100–150 units) slips to 2027–28 (PLTR is a team member with Anduril, Northrop, L3Harris).
- NGC2: the Army set the Next-Gen Command & Control data baseline on June 22, 2026, naming Anduril the lead (tactical/C2 layer) with Palantir a subordinate partner providing Foundry as the cloud data layer, under Anduril's ~$20B-ceiling NGC2 agreement (no separate PLTR dollar value disclosed; ceilings ≠ bookings); Project Convergence-Capstone 6 (~July 2026) is the next force-on-force test before fielding.
- Golden Dome (~$185B missile-defense program), Palantir is a named software co-developer (with Anduril) in the command-and-control consortium (Scale AI and others alongside); prototype testing is targeted for summer 2026, and PLTR is an approved vendor on the MDA SHIELD IDIQ (~$151B shared ceiling). No specific PLTR contract value is disclosed yet.
The exposure is two-sided. The DOGE-era federal-modernization push and contract consolidation are a tailwind; but ~42% government concentration brings appropriations lumpiness, continuing-resolution / shutdown risk, and reputational/headline risk (immigration-enforcement ties drew ACLU scrutiny and shareholder proposals; a UK Met Police contract was blocked in June 2026). It is both the durability argument and a key risk.
Financial Health & Trends
The financial profile is what makes "just a Hold" a genuinely hard call. PLTR posts a Rule of 40 of ~155% (93% revenue growth + a 62% record adjusted operating margin, both on the Q2 quarterly basis; on the FY26E annual estimate the charts plot ~60%), roughly 3.9× the 40 threshold, and the highest among the largest-cap companies. Critically, growth and margin are expanding together, which refutes the old "growth bought with spend" bear claim.
Rule of 40 (%)
From ~58% (FY23) to ~155% (Q2'26): accelerating growth and expanding margin. Almost nothing in software prints this.
Margin expansion (%)
Gross margin ~85%; adjusted operating margin to ~60%; adjusted FCF margin ~63% (record). The operating leverage is the bull's structural argument.
FCF and the GAAP bottom line
PLTR is genuinely GAAP-profitable, not just adjusted: Q2 2026 GAAP operating income was $912M (+239% YoY), and adjusted EPS was $0.41 (vs ~$0.34 consensus), the kind of bottom line that underpins its S&P 500 membership. Adjusted FCF ran at a record 63% margin on $1.935B of revenue.
Adjusted free cash flow ($B)
Adjusted FCF from ~$0.2B (FY22) to a guided ~$4.6B FY26E: the cash the DCF capitalizes. Remember the asterisk: this is before the ~12%-of-revenue stock comp.
FY20–FY30E: revenue & adjusted operating margin
Revenue compounds on the AI build-out while the adjusted operating margin holds in the low 60s. FY26 is guidance; FY27E–FY30E are author estimates used in the DCF.
The "adjusted" asterisk
Here is the catch behind the elite margins: the headline 63% FCF and 62% operating margins are adjusted: they exclude stock-based compensation of ~12% of revenue. SBC is a real economic cost that dilutes owners (see the next section). On a GAAP, fully-diluted basis the per-share economics are lower than the adjusted headline implies, which is exactly why the DCF, which has to reckon with the cash and the dilution, lands where it does.
Balance Sheet, SBC & Dilution
PLTR runs a fortress balance sheet: ~$8B of cash and US Treasuries, no debt, which funds R&D and modest buybacks and supports a lower discount rate. But the capital story has a catch the bull case under-weights: dilution. Stock-based comp still runs ~12% of revenue, the fully-diluted share count (~2.571B) sits ~7% above the basic count (~2.40B) from in-the-money options/RSUs, and "adjusted" FCF/EPS exclude that SBC. The cash is real; so is the slow leak in per-share value.
A recurring overhang is insider selling: roughly $6B cumulatively since 2024 (CEO Karp and co-founder Thiel the largest), all under pre-arranged 10b5-1 plans. The pace has moderated: Thiel's last large discretionary sale was ~$290M in March 2026, with no new open-market sales since. Institutional ownership is ~55%, with 13F aggregates showing modest net trimming through Q2 2026. For a name whose entire thesis is durability, the people closest to it monetizing into strength is, at minimum, worth noting.
Valuation & Comps
There is no gentle way to say it: PLTR is among the most expensive software stocks ever. At ~$172 it trades at ~40× forward sales, ~98× forward earnings and ~105× trailing FCF2. Even after round-tripping the H1 de-rating from a peak of ~55× sales, it sits at roughly 4× the multiples of high-quality peers: CrowdStrike ~19× EV/sales, Datadog ~11×, ServiceNow ~14×, Snowflake ~15×. No historical software multiple anchors PLTR; the bulls' best single counter is a ~1.3 PEG on the hyper-growth4.
PLTR forward P/S history
From ~9× (2022) to a ~55× blow-off (late 2025), down to ~32× at the June low, and back to ~40× after the Aug 3 blowout and the subsequent run. The rebound is real, and ~40× is still ~4× the ~10× software median.
Forward EV/Sales: software peers
At ~40× PLTR is in a tier of its own, well above CRWD/SNOW/NOW/DDOG, all of which are themselves premium names. The growth is faster, but not 4× faster.
The Street, the shorts, and the de-rating regime
The set-up: the sell-side stays net-bullish: consensus is a "Moderate Buy" (roughly two-thirds Buy, a third Hold, a handful Sell) with an average target of ~$200, above both the ~$172 price and our $175 target. So even after the blowout, PLTR trades below the Street's average, though it has now closed most of the gap to consensus. The Street range is wide: a ~$255 high to a ~$70 low. The most prominent named bear is Michael Burry, who holds PLTR puts and pegs intrinsic "well under $50." Short interest is low (~3% of float), a valuation debate, not a crowded short.
Crucially, the H1 de-rating was not a PLTR-specific stumble, it was a sector regime. A 2026 "SaaSpocalypse" re-rated software broadly on fears that AI agents erode the per-seat subscription model: the median public-SaaS EV/Revenue compressed ~25% (toward ~5×). PLTR, as the most expensive name, had the most multiple to lose, and the most to regain: the Aug 3 blowout snapped it straight back. That reframes both the drawdown and the rebound, but it does not rescue the valuation: even post-recovery, the reverse DCF still requires near-flawless execution.
Why a multiple alone is the wrong frame
"40× for 90%+ growth" is the bull soundbite, and "40× is insane" is the bear's, but a sales multiple can't price the path: how fast growth decelerates, how durable the margin is, how much the SBC dilutes. The rigorous test is a cash-flow model that discounts the whole glide path. That is next, and for PLTR the most important number isn't the intrinsic itself; it's what the reverse DCF says the price already assumes.
DCF + Reverse DCF
A software DCF is simpler than an industrial one: we project revenue × FCF margin to get unlevered free cash flow, discount the explicit years, add a Gordon-growth terminal and net cash, and divide by shares. Toggle Base / Bull / Bear / Reverse and drag the WACC and terminal-growth sliders, the per-share value recomputes live. The base now prints roughly at the price: after the Aug 3 re-rate the intrinsic and the price converged. The reverse tab is the one that matters.
Base: revenue ~$8.15B → ~$42.2B by FY30 (~50% FY26–30 CAGR), FCF margin ~58–66%. WACC 10%, terminal growth 5.0% → a DCF intrinsic ~$175, right at the price, and note ~85% of that value sits in the terminal. It uses the mid-60s% adjusted FCF margin; charge the ~12%-of-revenue SBC (margins into the ~50s% → ~$144) or use a GDP-anchored ~3% terminal (→ ~$130) and it drops back below the price. Our $175 12-month PT ≈ the intrinsic ≈ spot; this is a fair-value Hold, not a claim the stock is cheap.
Bull: durability wins, revenue ~$46B by FY30 (slower deceleration), FCF margin holds the mid-60s%, and the discount rate eases as the cycle de-risks; WACC 9.5%, terminal growth 5.0% → ~$214 intrinsic, supporting the $215 12-month PT.
Bear: growth normalizes toward the 30s% (revenue ~$36B by FY30), the FCF margin settles ~56–59% as SBC is recognized, and the discount rate lifts; WACC 10.5%, terminal growth 4.0% → ~$107 intrinsic, below the $115 12-month PT, which assumes a less-severe glide.
Reverse DCF: holds base-case cash flows and solves for the terminal growth the current $172.01 price implies at the slider WACC. The output tells you how much durability is already in the price, for PLTR, the single most important number in this note.
DCF Inputs
▶ DCF intrinsic ≈ market at these inputs, fairly valued only if this hyper-growth + margin profile persists.
5-Year FCF Forecast ($B)
| ($B) | FY26E | FY27E | FY28E | FY29E | FY30E |
|---|---|---|---|---|---|
| Revenue | $8.2 | $13.0 | $19.5 | $28.0 | $42.2 |
| FCF margin | 58% | 60% | 62% | 64% | 66% |
| Unlevered FCF | $4.7 | $7.8 | $12.1 | $17.9 | $27.9 |
FCF = revenue × FCF margin. The base uses the mid-60s% adjusted margin (terminal); charging the ~12%-of-revenue SBC would use the ~50s%. All forecast values FY27E+ are author estimates.
Sensitivity: DCF value / share ($) vs WACC × terminal growth
Base-case cash flows across the grid (WACC × terminal growth).
Reverse DCF: what is the market pricing in?
The reverse DCF is the honest counterweight to the "best business in software" story, but read it as a package, not a single number. The terminal the price implies (~5%) is right on our 5% base, so the "perfection" is not in that one input; it is that the whole aggressive stack has to hold at once: the ~50% FY26–30 explicit-growth CAGR, a sustained mid-60s% adjusted FCF margin, and a ~5% perpetual terminal (2× the GDP norm) into which ~85% of the value is loaded. Charge the SBC or normalize the terminal and independent bases fall into the low-$130s to mid-$140s, below the price. You are not buying a mispriced asset; you are buying a great one at a full price and underwriting near-flawless execution for years.
PT calculator (forward-sales cross-check)
FY26E revenue/share ~$3.40 × ~51.5× forward sales lands at our $175 target, roughly in line with spot: at ~$172 PLTR already trades ~51× current-year sales, so the target embeds essentially no multiple compression. Note how sensitive it is: at a "normal" premium ~20× it is ~$68; the entire valuation rests on the market continuing to pay a top-decile sales multiple.
Risk / Reward calculator
Defaults use the scenario endpoints (bull $215 / bear $115), not a tight setup. The ~0.8:1 endpoint ratio is slightly downward-tilted after the run: the probability-weighted intrinsic (~$164) sits just below spot, so this is the quantitative case for a Hold at fair value, not a Buy.
Note: The assistant reasons from the dashboard's data snapshot and thesis sections, it does not browse the web or access real-time fundamentals beyond what's in data.js. Treat its responses as scenario-modeling support, not primary research. Author judgments on rating, PT, and probabilities remain with the analyst.
Upcoming Catalysts
| Catalyst | Window | Why it matters |
|---|---|---|
| Q3 2026 earnings | ~Nov 2, 2026 | The next hard catalyst. After the Aug 3 blowout (revenue $1.935B, +93%; FY26 raised to ~$8.15B), the market needs to see whether the ~90% growth and the raised >134% US-commercial guide hold. A priced-for-perfection name needs another beat-and-raise to justify the re-rate. |
| Q2 2026 earnings (reported) | Aug 3, 2026 · past | A blowout beat-and-raise: revenue $1.935B (+93% YoY), US commercial +149% to a record $764M, record 62% adjusted operating margin and 63% adjusted FCF margin, adj EPS $0.41 vs ~$0.34 est, FY26 revenue raised to ~$8.15B. The print drove the ~+48% recovery off the June low. |
| NGC2 fielding / Project Convergence follow-through | H2 2026 | Project Convergence-Capstone 6 (~July) tested the Army's Next-Gen C2, whose data baseline went live June 22 with Anduril as lead and Palantir's Foundry the subordinate cloud data layer. Clean readouts de-risk fielding; PLTR's dollar share is undisclosed. |
| TITAN milestones | FY26 interim; full-rate 2027–28 | An interim transition milestone falls in FY26; the full-rate production decision + fielding (100–150 units) is a 2027–28 event, a longer-dated pipeline than the prototype implies. |
| Golden Dome C2 prototype | Summer 2026 | PLTR sits in the Anduril-led command-and-control consortium for the ~$185B missile-defense program; a successful prototype is pure optionality, no PLTR award disclosed yet. |
| Federal budget / DOGE headlines | Ongoing | Two-sided and high-beta given ~42% government revenue: a federal-modernization tailwind vs continuing-resolution / shutdown / budget-cut risk (DOGE cuts were cited in the Jan 2026 selloff). |
| Multiple normalization | Ongoing | The swing factor that dwarfs the fundamentals near-term: at ~40× sales, a move toward peer multiples (~15–20×) outweighs a quarter of upside. The re-rate leaves the stock fully valued. |
Risk Factors
- Valuation / no margin of safety (the pre-eminent risk). At ~40× forward sales, ~98× earnings and ~105× FCF, PLTR has essentially no cushion. Our base DCF (~$175) now only equals the price; on a normal terminal or SBC-charged margin it drops into the low-$130s to mid-$140s. A disappointment has a long way to fall.
- Growth deceleration (mathematically normal). The raised guidance already steps growth from 93% to ~82% for FY26; as the base scales past $10B, the law of large numbers pulls it lower. The whole valuation rests on US-commercial holding >134%, a single line decelerating toward the 30s% collapses the reverse-DCF math.
- Stock-based comp & dilution. SBC at ~12% of revenue; the diluted share count sits ~7% above basic. "Adjusted" FCF and EPS exclude it, so the real per-share economics are lower than the headline, and the dilution compounds.
- Government concentration & political risk. ~42% of revenue rides federal budgets, exposing PLTR to appropriations lumpiness, continuing-resolution / shutdown risk, and reputational/headline risk (immigration-enforcement scrutiny; a blocked UK contract). Top-3 customers are still ~16% of revenue.
- European digital-sovereignty losses (a 2026 development). France's DGSI moved to replace Palantir's Gotham with a domestic alternative (ChapsVision/ArgonOS, June 2026), part of a ~€655M French sovereign-AI push. That move, plus UK NHS-contract scrutiny, fed a broader European-sovereignty pressure that contributed to a ~7% single-day PLTR drop on June 22. International government (~10% of revenue) is most exposed, and the "Palantir = US-dependence risk" theme could broaden across allied governments.
- Insider selling, heavy cumulatively, but the pace has moderated. ~$6B cumulatively since 2024 (Karp, Thiel the largest), under pre-arranged 10b5-1 plans. The pace has eased: Thiel's last large discretionary sale was ~$290M in March 2026, with no new open-market sales since (through late June). Not a legal signal, but a steady supply of stock near the highs.
Scenario Stress Tests
| Scenario | Mechanism | DCF intrinsic | vs spot ~$172 |
|---|---|---|---|
| Base | ~50% FY26–30 rev CAGR; ~58–66% FCF margin; WACC 10%, tg 5% | ~$175 | +2% |
| Charge SBC as real dilution | FCF margins into the ~50s% (the GAAP-honest cash margin); base WACC/tg | ~$144 | (16%) |
| GDP-anchored terminal | Terminal growth to 3% (the long-run norm); base cash flows | ~$130 | (24%) |
| Risk re-rating | WACC to 12% (multiple/rate shock); base cash flows | ~$121 | (29%) |
| Growth halves | Rev growth halves to ~25%/yr as AIP adoption slows; margins held | ~$87 | (49%) |
| Full bear | Deceleration + 10.5% WACC + 4% terminal + ~57% margin | ~$107 | (38%) |
| Bull: durability holds | ~$46B FY30 rev, ~66% margin, WACC 9.5%, tg 5.0% | ~$214 | +24% |
All values are DCF intrinsic values today (not the 12-month PTs, which roll forward ~a year of growth), each shock changes only the inputs noted, holding the 2.40B basic-share count and $8B net cash constant. Deltas are vs the ~$172 spot. Note that the base only just reaches spot, and every normalization, SBC-charged margin, a GDP terminal, a higher WACC, drops the intrinsic into the $87–144 range below the price: even after the re-rate, there is no margin of safety.
Bull vs Bear Debate
| Issue | Bull view | Bear view |
|---|---|---|
| Is the valuation a problem? | A fast compounder (~50% FY26–30 revenue CAGR) grows into a rich multiple over time; a ~1.3 PEG is reasonable for the best growth-plus-margin profile in software. | ~40× sales / ~105× FCF leaves no margin of safety; the DCF only just equals the price on generous inputs, and the SBC makes the "FCF" overstated. |
| Growth durability | US-commercial +149%, 157% NDR, FY26 guide raised again; AIP is becoming the AI-application standard, durability is the whole point. | 93% → ~82% guided deceleration is just the start; past $10B the base mathematically slows toward the 30s%, and one soft US-commercial print breaks the thesis. |
| Software or services? | ~85% gross margin and 63% FCF margin are software economics; boot camps are a sales motion, not the product. | Forward-deployed engineers and bespoke deployments look like high-end consulting; the question is whether it scales without re-acceleration in headcount/SBC. |
| The government engine | A structural federal-AI tailwind, contract consolidation (Army EA), and a widening defense backlog (NGC2, Maven, TITAN). | ~42% concentration in appropriations brings lumpiness, CR/shutdown risk, and headline/reputational exposure; ceilings ≠ booked revenue. |
| What about the insiders? | 10b5-1 plans are pre-arranged, mechanical, and common for founder-led companies; it is not a fundamental signal. | ~$6B of selling since 2024, near the highs, by the people closest to the durability thesis, at minimum, not a vote of scarcity. |
Technical Analysis
PLTR trailing-12-month closes
RSI (multi-timeframe)
Stretched-to-overbought after the August earnings gap; the H1 momentum washout has fully reversed.
MACD vs Signal
Crossed back above the signal line on the August gap; momentum flipped positive after the H1 downtrend.
Relative strength (2026 YTD)
Still a slight laggard but recovered: −3% YTD vs the Nasdaq-100 +16% and software (IGV) ~+10%, after round-tripping the H1 drawdown on the Q2 print.
EMA stack (current)
Trader's view
- Breakout: price (~$172) has reclaimed the 50-, 100- and 200-day averages on the August earnings gap, after the H1 slide to the $106 June low.
- Key support: the reclaimed moving-average band (~$132–146), then the $106 June low far below.
- Key resistance: the $180+ shelf and ultimately the $207.52 high; extending needs another beat, not just a multiple that holds.
- Momentum (RSI/MACD) has turned decisively up on the August print; the near-term risk is an overbought pullback, not a lower low.
Glossary & Methodology Notes
- Ontology
- Palantir's core abstraction: a living model that maps an organization's data to real-world objects (assets, people, events) and the actions/decisions taken on them. It is the "digital twin" that AIP agents operate over, and the source of the switching costs.
- AIP / boot camp
- The Artificial Intelligence Platform, the governance + orchestration layer that turns LLMs into permissioned, auditable agents on the ontology. A "boot camp" is the hands-on workshop that converts a prospect from pilot to a production use-case in days/weeks, the core commercial sales motion.
- Rule of 40
- A software health metric: revenue growth % + profit margin % (Palantir uses adjusted operating margin). Above 40 is good; PLTR's ~155% is exceptional and rare.
- SBC & dilution
- Stock-based compensation, paying employees in equity. It is a real economic cost that dilutes shareholders (the share count grows). "Adjusted" FCF/EPS add SBC back, flattering the per-share picture; a rigorous DCF must reckon with the dilution.
- RDV / RPO / TCV
- Remaining Deal Value, Remaining Performance Obligations, and Total Contract Value, measures of contracted backlog. Government contract ceilings (e.g. the $10B Army EA) are maximums, not bookings, task orders determine actual revenue.
- SaaS DCF & Reverse DCF
- For a software name, free cash flow ≈ revenue × FCF margin. The DCF discounts the projected FCFs plus a terminal value to an intrinsic value. A reverse DCF inverts it: hold a discount rate and solve for the growth the current price implies, telling you how much of the future is already priced.
- Why a great business can be a Hold
- Quality and price are separate questions. A best-in-class company bought at a price that already discounts a near-flawless decade can still offer no margin of safety and a balanced risk/reward, the definition of a Hold.
Methodology
- Snapshot anchor: August 7, 2026 (price = August 7 close). Live price patches via the Cloudflare-Worker quote proxy on page load.
- All figures in USD. FY2025 + Q2 2026 are reported; FY2026 is management's raised guidance; FY2027E–FY2030E and all DCF inputs are author estimates, the largest swing factor.
- Market cap uses basic shares (~2.40B); per-share intrinsic uses the same. The ~2.571B diluted count is the dilution overhang discussed in Balance Sheet & Dilution.
- The 12-month PT rolls the DCF intrinsic forward ~one year of growth. Conclusions are the author's view. Illustrative, not investment advice.
Sources & Citations
Inline citations
Superscripted numbers in the body link here. Click any N in the report to jump back to the source.
- Palantir Technologies, Second Quarter 2026 Results (reported Aug 3, 2026): revenue $1.935B, +93% YoY (+19% QoQ); US commercial $764M (+149%); US government $809M (+90%); total US $1.573B (+115%); GAAP operating income $912M (+239% YoY); adjusted operating margin 62% (record); adjusted FCF margin 63% (record); adjusted EPS $0.41 (vs ~$0.34 est); Rule of 40 ~155%; NDR 157%. FY2026 guide RAISED: revenue ~$8.15B (~+82%), US commercial >$3.424B (~+134%), adjusted FCF $4.5–4.7B. Palantir Investor Relations. ↩
- Market data (stockanalysis.com, as of August 7, 2026 close): price $172.01 (+10.3% on the day); market cap ~$413B on ~2.40B basic shares (diluted ~2.571B); 52-week range $106.37–$207.52; Dec-31-2025 close $177.75 (YTD −3.2%); forward P/S ~40× (NTM); forward P/E ~98×. ↩
- Government programs per Department of Defense / Army disclosures and company statements: Army Enterprise Agreement (up to $10B ceiling, 10-year ordering period, Aug 2025); Maven Smart System (~$1.3B cumulative ceiling, 20,000+ users; designated a DoD program of record, Mar 2026); TITAN ($178M, 10 prototypes; full-rate production 2027–28); Navy ShipOS ($448M); NGC2 common data layer (baseline established June 22, 2026 with Anduril as lead integrator and Palantir a subordinate partner under Anduril's $20B-ceiling agreement); Golden Dome (Anduril-led C2 consortium; no PLTR award disclosed). Ceilings are maximums, not booked revenue. ↩
- Valuation: independent third-party DCF base cases (Alpha Spread and others) sit below the market on ~30%+ multi-year growth and mid-40s% net margins; charged for SBC or a GDP-anchored terminal, the author's in-house SaaS DCF (revenue × FCF margin) lands in the low-$130s to mid-$140s, documented in the DCF section. Peer multiples (CRWD, DDOG, NOW, SNOW) per company disclosures / consensus, August 2026. All FY27E+ figures are author estimates. ↩
Background reading
- Palantir Q2 2026 Form 10-Q + the quarterly business update / shareholder letter, segment revenue, margins, RDV/RPO, customer counts, SBC.
- Palantir FY2025 Annual Report (Form 10-K), full-year financials, segment/customer disclosures, risk factors, share-count detail.
- US Department of Defense / Army contract announcements, the Army Enterprise Agreement, Maven, TITAN, NGC2 / Project Convergence.
- Peer disclosures (CRWD, DDOG, NOW, SNOW), the EV/Sales and Rule-of-40 comparison and the software-multiple context.
- Palantir AIPCon 10 (June 4, 2026) + the Google Cloud and Zeta Global partnership announcements, and the AI Forward-Deployed Engineer GA (March 2026), the commercial/distribution catalysts.
- Army.mil / Breaking Defense / DefenseScoop (June 22, 2026), the NGC2 data baseline and the Anduril-lead / Palantir-subordinate structure; the Maven program-of-record memo (Mar 2026); the Golden Dome consortium.
- French DGSI / Gotham-replacement reporting (June 2026) and the European digital-sovereignty theme; MarketBeat / TipRanks (analyst distribution, ~$200 Moderate Buy consensus); the Michael Burry short disclosures.
- Market data: stockanalysis.com / ytdreturn.com (price, shares, market cap, 52-week range, −3.2% YTD, forward multiples) as of August 7, 2026.
Disclaimer. This report is the author's institutional equity-research view, prepared for portfolio and educational purposes. It is not a recommendation to buy, sell, or hold any security. Forward-looking statements are subject to risk and uncertainty; past performance is not indicative of future results. Consult a licensed financial advisor before making investment decisions. All third-party trademarks are the property of their respective owners.