Palantir Booked the Growth the Market Says Won't Happen
The most striking number in Palantir's second-quarter report was not the headline. It was the fact that a company already recognizing $764 million a quarter in U.S. commercial revenue managed to grow that line 28% sequentially — accelerating, not fading, against a year-ago comparison that was itself up nearly triple digits . Businesses at scale are supposed to obey the arithmetic of large numbers. Palantir's commercial engine is, for now, defying it, and that single fact is what the argument over this stock ultimately turns on.
Palantir builds the software layer that sits between an organization's raw data and the decisions it makes — historically Gotham for defense and intelligence agencies and Foundry for enterprises, and over the last two years the Artificial Intelligence Platform (AIP) that bolts large-language-model reasoning onto that same data "ontology." Revenue splits almost evenly between two reporting segments: government, which supplied 52% of the first half of 2026, and commercial, at 48% . Geographically the business is overwhelmingly domestic — the United States generated $3.32 billion of FY25's $4.48 billion, with the United Kingdom and the rest of the world the modest remainder . That concentration is the whole story in miniature: Palantir is a bet on American government and American enterprise adopting a single vendor's operating layer for AI, and the second quarter suggested both are doing so faster than the Street's models allow.
The house view here is a qualified long. The specific claim is narrow and falsifiable: the consensus revenue path for 2027 embeds a deceleration steeper than the company's own signed backlog can justify, and as that gap closes on estimate revisions the stock carries higher than a static valuation lens implies. This is not a call that the multiple is cheap — it plainly is not — but that the earnings power feeding the multiple is being under-modeled. Three things have to be true for that to hold, and the evidence base for each is worth laying out.
The Quarter That Reset the Base
Start with what is demonstrated rather than projected. Consolidated revenue grew 93% year-over-year and 19% sequentially to $1.935 billion in the quarter ended June 30, 2026 . U.S. commercial rose 149% to $764 million and U.S. government 90% to $809 million — the crucial detail being that government, long treated as the mature ballast that would drag the blended rate down, re-accelerated to 18% sequential growth of its own . Both domestic engines fired at once. Net dollar retention reached 157% , meaning the average existing account spent roughly 57% more than it did a year earlier; this is expansion-led growth, the highest-quality form software compounding takes, because it does not depend on a perpetual sprint to add logos.
The profitability is not a mirage layered on top of the growth. GAAP income from operations was $912 million, a 47% margin, and on management's adjusted framework — which adds back stock-based compensation and related payroll tax — operating income was $1.194 billion at a 62% margin . The combination produces a Rule of 40 score of 155% , a figure that has very few, potentially zero, real precedents among pure-play software companies at this revenue scale. Cash generation backs the accounting: operating cash flow of $1.216 billion actually exceeded GAAP net income of $1.062 billion , and adjusted free cash flow of $1.220 billion ran at a 63% margin . The balance sheet holds $9.2 billion of cash, equivalents, and short-term Treasuries against no debt , with capital expenditure immaterial — this is an asset-light business where free cash flow tracks operating income closely.
The forward signal is what elevates the print from "great quarter" to "reset base." Closed total contract value reached $3.373 billion, of which a record $2.132 billion was U.S. commercial, up 153% year-over-year . U.S. commercial remaining deal value — contracted but not yet recognized — stood at $6.238 billion, up 124% and 27% sequentially . When bookings grow faster than the revenue they will eventually become, the near-term recognized number is understating the trajectory, not flattering it. Management responded by raising full-year 2026 revenue guidance to between $8.150 and $8.158 billion and lifting the U.S. commercial guide to above $3.424 billion, a growth rate of at least 134% .
Why the 2027 Consensus Is Too Low
The mispricing this thesis rests on is subtle, because it is not a mispricing of the past — it is a mispricing of the slope. The sell-side consensus carries roughly $12.15 billion of revenue for FY27, an implied fade to about 49% growth from 2026's ~82%. The house estimate is $12.4 billion, or +52% . The gap is small in dollars but directional in meaning: the Street is modeling a deceleration curve that the signed backlog does not support. With U.S. commercial RDV up 124% and total remaining performance obligation climbing from $4.45 billion in the first quarter to $4.90 billion in the second , a meaningful portion of 2027 is already pre-contracted rather than pipeline-dependent. A business fading to sub-50% growth the year after re-accelerating sequentially, with expansion running at 157% retention, requires a demand break the data is not showing.
This is the reason the mispricing frame is credible despite an obviously rich valuation. The re-rate does not need the multiple to expand — it needs consensus estimates to catch up to the bookings. Each quarter that U.S. commercial holds a sequential pace above roughly 20% forces the Street to lift its too-steep 2027 number, and the stock can carry higher on revisions alone even as the headline multiple compresses. That is a different and more durable engine than "the market will keep paying 50 times sales."
Which brings the discussion to what "the market is getting wrong" actually means here. It is not that investors have failed to notice the growth — the stock's best week since 2024 followed the print, and price targets moved up across the board. It is that the crowd still handicaps Palantir as a name that must obey the standard SaaS scaling law and fade hard in 2027. The bookings say it fades later and more gently. That nameable, testable disagreement is what lets a rational holder own the stock against a full multiple rather than genuflect to it.
Sovereign AI and the Government Half
The government segment is where the thesis draws its second leg of support and, simultaneously, its sharpest risk. The near-term tailwind is real: U.S. government revenue re-accelerating to 90% year-over-year growth reflects both defense programs ramping and a broader "sovereign AI" demand pull, as agencies move to own their AI operating layer rather than rent generic model access. Third-quarter government seasonality — the federal fiscal year ends September 30 — has historically been a source of strength as agencies spend into their appropriations, and management's third-quarter guide of $2.160 to $2.164 billion in total revenue builds that in.
But the same calendar cuts both ways, and this is the risk the bull case cannot wave away. Government contracts carry termination-for-convenience provisions — Palantir's own filing states plainly that many of its contracts, including government contracts, contain them, and that a terminating customer may be entitled to a pro-rata refund of its deposit . That is a structural feature of selling to the state, not a Palantir-specific flaw, but it means the government book is not the locked-in annuity a commercial RPO implies. The third-quarter print, expected around November 2, 2026, is the appropriations-sensitive quarter: a continuing resolution that delays task orders would land on precisely the segment the market is least braced for a stumble in. A soft government line into a maximalist "both engines firing" narrative would read larger than the dollars involved, and Palantir's holder base — heavily retail, largely passive institutions, no marquee activist defending the valuation — tends to reprice violently in both directions.
There is also a quieter political-optics thread worth naming rather than dismissing. Palantir's public-sector prominence draws scrutiny that a pure enterprise vendor never faces; the reporting that the company paid roughly £2 million of UK corporation tax in 2024 despite substantial public-sector contracts is the kind of story that recurs precisely because the government relationship is the point. It is not a fundamental threat, but it is a persistent headwind to the reputational latitude the company needs to keep winning sensitive deployments.
The Quality-of-Earnings Frictions Are Real but Disclosed
The bear case against this stock has three legs beyond valuation, and honesty requires taking each seriously. First, stock-based compensation is rising, not falling — total SBC was $265 million in the quarter, up 66% year-over-year, driven by new restricted-stock, performance-RSU, and stock-appreciation-rights grants . That is the entire wedge between the 47% GAAP operating margin and the 62% adjusted figure, and it is why the diluted share count drifts up roughly 1% a year (2.569 billion diluted in the quarter ) rather than shrinking. Growth here is partly bought with equity, and any honest valuation must charge that cost — which the house model does, by carrying the rising share count rather than pretending the adjusted margin is free.
Second, the receivables build is conspicuous. Accounts receivable climbed to $1.5 billion at June 30 from $1.0 billion at year-end, and a single customer — "Customer I" — represented 27% of that balance, up from 25% . That is genuine working-capital concentration; a delay or dispute with one counterparty would swing reported cash conversion. It is worth weighing against the fact that no single customer exceeded 10% of revenue in the period , so the concentration is a collection-timing exposure rather than a revenue-dependency one.
Third, in March 2026 Palantir amended a third-party cloud agreement committing it to spend at least $5.6 billion over ten years, with annual minimums ranging from $268 million to $979 million . That is a sizable fixed obligation for an asset-light company, and it is the direct cause of the adjusted gross margin stepping down toward ~86% as government cloud-hosting costs are absorbed . The charitable read — that this is a deliberate near-term margin give to deepen agency relationships and remove a delivery bottleneck — is plausible, but it is a commitment that must be fed with revenue growth to earn its keep. None of these three is a thesis-breaker; all are disclosed, widely discussed, and already in the price. The frictions matter for how one weights the adjusted numbers, not for whether the operating story is intact.
What Fifty Times Forward Cash Flow Actually Buys
At $176.01 the equity is capitalized near $404 billion [Tenzing market data], which on the raised FY26 revenue guide of ~$8.15 billion is roughly 50 times current-year sales — a multiple that survives only if the growth ahead is genuinely exceptional. Netting the $9.4 billion of cash and Treasuries against no debt (\$2.03 billion cash plus $7.38 billion marketable securities ) leaves an enterprise value near $395 billion. Against the FY26 adjusted free-cash-flow guide of $4.5–4.7 billion that is roughly 85 times this year's cash flow, falling to about 56 times the ~$7.1 billion the house model carries for FY27 . The point of regrounding onto free cash flow rather than the EV/EBITDA the screens quote — 130 times trailing [Tenzing market data] — is that SBC is the whole distortion in the EBITDA family here; free cash flow is the clean lens, and even the clean lens is expensive.
The relevant peer check sharpens rather than rescues the number, and it does so precisely because no other software name trades where Palantir does. CrowdStrike, at ~$225 and a ~$229 billion market capitalization [Tenzing market data] against ~$5.94 billion of consensus FY27 revenue [Financial Modeling Prep analyst-estimates feed], sits near 38 times forward sales while growing around 20% — the richest of the elite-growth cohort, and still less than four-fifths of Palantir's ~50 times current-year sales. ServiceNow, a larger and more mature platform, trades closer to 7–8 times sales (roughly $112 billion of market value on ~$14.7 billion of trailing revenue). Palantir is therefore paying up not merely versus the median software name but versus the single most expensive comparable, for roughly four times its growth rate at higher margins. Whether that premium is deserved is the entire debate, and it does not resolve on the multiple in isolation — it resolves on whether the growth durability is real. A disciplined discounted-cash-flow anchor, discounting the house revenue track at an 11.5% cost of equity, lands closer to $110–115 per share on a forward basis, because it underwrites a real multiple fade as growth decelerates toward the mid-20s by 2030. The gap between that intrinsic anchor and the tape is not an error in either — it is the market pricing a scarcity premium the DCF refuses to extend, and the honest reading is that intrinsic value sits below spot unless the premium multiple proves durable.
Where This Resolves
The tension everyone already knows — elite franchise, nosebleed multiple — is the starting point, and it does not resolve into a comfortable answer. What tips the balance is the specific, checkable disagreement over the 2027 slope. The signed backlog says the consensus deceleration curve is too steep, and while that gap persists the stock can grind higher on upward revisions rather than on further multiple expansion. That is the constructive case, and it is a real one: bookings running ahead of revenue, expansion at 157% retention, and both domestic segments accelerating at once are not the fingerprints of a business about to obey the fade the Street has penciled in.
But the honesty the valuation demands is that the downside is not shallow. A single quarter of U.S. commercial sequential growth below roughly 15% would validate the maturity narrative, pull 2027 back toward the consensus $12.15 billion, and compress the multiple toward the peer median in the same motion — a path that clears well below spot. There is no version of the future where the growth breaks and the multiple survives; the two are welded together at this valuation. The appropriations-sensitive third-quarter print is the near-term test of exactly that, and the government book's termination-for-convenience exposure means the risk is not evenly distributed across the calendar.
Weighing it: the operating evidence is strong enough, and specific enough, that the market is more likely under-modeling the 2027 earnings power than over-paying for a franchise about to disappoint. The stock is not mispriced cheap — on any static lens it discounts near-flawless execution — but the dynamic case, that estimates chase the bookings higher, is the sturdier read of a business whose signed contracts are still outrunning its recognized revenue. The risk-reward favors the long side into the coming print, with the clear understanding that the entire thesis lives or dies on a single sequential growth line that the market will not wait to re-price.
Disclaimer: Tenzing Analytics, Inc. is not a registered investment adviser, broker-dealer, or financial institution, and all content provided by Tenzing Memo is AI-generated for informational purposes only. It does not constitute investment advice, a recommendation, or an offer or solicitation to buy, sell, or hold any security or financial instrument. Content is general in nature, not tailored to any individual's financial situation or objectives, and may contain errors, omissions, or inaccuracies; it should not be relied upon as a primary basis for any investment decision. Any forward-looking statements are subject to risks and uncertainties and may not reflect actual future results. Users should conduct their own research and consult qualified financial, legal, or tax advisors before making investment decisions, and by accessing this content, you acknowledge and accept these limitations.