Revenue +96% and a Bigger Guide, but Bookings Grew 6% and the Margin Slipped Again: Downgrading to Hold

Results vs. Consensus

On every income-statement line, this was another record quarter. Revenue of $2,313M grew 96% year over year and 19.5% sequentially, both the highest rates the company has reported as a public company, on a quarterly base that crossed $2B for the first time. The revenue beat over the Street was $133M (+6.1%) and the beat over Palantir's own August guide was $151M (+7.0%). That sits inside the 6–8% range the company has now delivered for three consecutive quarters, after a beat of just under 6% in Q4 2025, and it is the cadence we named in August as the one the multiple embeds. Growth accelerated again, from +85% in Q1 to +93% in Q2 to +96% in Q3. The deceleration the guide keeps implying has still not arrived in the reported numbers.

What changed is the relationship between what was recognized and what was booked. Revenue grew 96% while total contract value grew 6%. In Q2 the leading indicator ran ahead of revenue; in Q3 it ran far behind it. The rest of this note is largely about which of those two numbers deserves more weight for the next four quarters.

Q3 2026 Scorecard

Year-over-Year Comparison

Sequential Comparison

Quality of the Beat

Revenue. No acquisitions, no FX distortion, no accounting change. Of the $378M sequential add, $364M came from the United States: U.S. commercial added $220M and U.S. government $144M, while the international book added $13M. The 10-Q adds a useful split. Of the $619M year-over-year increase in commercial revenue, $407M came from customers who existed at the end of 2025 and $212M from customers added since. In Q2 the same new-customer figure was $88M. New logos are now doing far more of the commercial work than they were a quarter ago, which fits management's account of pilots converting into large first contracts inside a single quarter. Strategic commercial contracts (the legacy investment-linked book) contributed about $0.4M and will end after Q4, so there is no low-quality revenue in the number.

Margins. Adjusted operating margin of 61.7% matched Q2 and beat the 59.9% guide by 180 basis points, but the composition is the same as Q2's, only more so. Cost of revenue grew 29.1% sequentially against revenue growth of 19.5%, and adjusted gross margin fell 120 basis points to 85.1%. The operating margin held only because operating expenses grew 16.3% sequentially. On a year-over-year view the picture is still favorable: adjusted gross margin is up 130 basis points from 83.8% and contribution margin rose to 73% from 66%. The direction of travel inside the year, though, is the one we flagged in August: gross margin down, operating leverage carrying the whole load.

EPS. GAAP and adjusted diluted EPS both printed $0.44, and this quarter the below-the-line items worked against the result rather than for it. Other income swung to a net expense of ($25.8M) from a $91.8M gain in Q2. The 10-Q attributes the year-over-year change to upward adjustments on privately-held securities in the prior-year period plus higher losses on marketable securities, including $25M of net unrealized losses on publicly-traded equity holdings. That is roughly a one-cent drag in Q3 against the three-cent tailwind Palantir disclosed in Q2. The effective tax rate was 1.2%, so pre-tax income is still almost untaxed, but the adjusted EPS figure applies a normalized 23% tax charge and still printed $0.44. Against a $0.41 bar, the beat is clean and operating.

Free cash flow. Adjusted free cash flow of $1,216M was flat on Q2 even as revenue grew $378M, and the margin fell to 52.6% from 63.1%. The swing sits in working capital. Receivables rose $398M in the quarter to $1,883M, up 26.8% against revenue growth of 19.5%, and days sales outstanding moved to roughly 75 days from 70 in Q2 and 68 at year-end. Contract liabilities grew 13.2% sequentially, slower than revenue, so the deferred-revenue cushion did not keep pace either. None of this signals a collection problem, and the full-year free cash flow guide went up $300M. But it is the second line, after bookings, where Q3 looked weaker than Q2.

Segment Performance

Segment dollars are as reported by management, rounded to the nearest million. The Total row is the income-statement figure of $2,313.4M.

On the commercial-versus-government cut, total commercial revenue of $1,168M (+113% YoY, +24% QoQ) passed total government revenue of $1,146M (+81%, +16%) this quarter. In August we wrote that commercial would cross over in the second half on any reasonable extrapolation; it did so in the first quarter of the second half. For five years Palantir has argued it should be valued as a commercial software company rather than a government contractor, and as of Q3 the revenue mix supports the argument on its own.

U.S. Commercial: The Revenue Line Did Not Slow; the Bookings Line Did

The U.S. commercial growth sequence through the AIP era now reads +121% (Q3 2025), +137% (Q4 2025), +133% (Q1 2026), +149% (Q2 2026), +148% (Q3 2026). The segment added $220M sequentially, more than the $169M it added in Q2, and grew 28.8% quarter on quarter from a starting base 28% larger than the one behind Q2's step. Customer count reached 686, up 29% year over year and 5% sequentially, so quarterly revenue per U.S. commercial customer rose to about $1.43M from about $1.17M in Q2, a 23% increase in a single quarter. The full-year guide moved to at least $3,559M (+143%), which after nine months of $2,343M implies at least $1,216M in Q4, another 24% sequential step.

Against that, U.S. commercial TCV fell to $1,463M from $2,132M in Q2 and grew only 11% year over year. Context matters here, and it cuts both ways. Q2 was an outlier: $788M above the prior record. Q3's $1,463M still exceeds every quarter before Q2, including the $1,344M of Q4 2025. Trailing-twelve-month U.S. commercial TCV of $6,114M is up 62%. So this is not a collapse in absolute bookings. It is, however, a collapse in bookings growth relative to revenue growth: segment TCV was 2.8x segment revenue in Q2 and 1.5x in Q3.

"Our U.S. commercial business grew 148% year-over-year and accelerated to 29% growth sequentially, and our U.S. government business grew 96% year-over-year and 18% sequentially." — David Glazer, Chief Financial Officer

"An industrial materials company piloted a nearly $3 million engagement in Q2 2026 before converting to a $29 million ACV deal last quarter." — Ryan Taylor, Chief Revenue Officer and Chief Legal Officer

Assessment: We raised our FY26 U.S. commercial estimate to +140–150% in August. Nine-month revenue is running at +144% and the guide floor implies at least +143%, so we narrow the range to +143–148%. The revenue line needs no defense. The question for FY27 is whether the remaining deal value that funds it is still compounding, and in Q3 it grew 4% sequentially against 27% in Q2. We address that directly in the Key Topics below.

U.S. Government: Accelerating Every Quarter Since Q3 2025

U.S. government revenue of $953M grew 96% year over year and 18% sequentially, and the growth rate has now risen in each of the four quarters since Q3 2025. We raised our FY26 estimate for the segment twice this year and were too conservative both times. Nine-month U.S. government revenue of about $2.45B is running roughly 90% above the prior-year period. The quarter's disclosed awards were operational rather than headline: an Army prime agreement to produce 8 TITAN systems, Maven above 100,000 users and now available to all of the services, and the DORA edge device bringing Maven into disconnected environments.

"Maven continues to deliver for the joint force, exceeding 100,000 users. Maven is now available to all of the services beyond just the combatant commands and many combat support agencies." — Shyam Sankar, Chief Technology Officer

Assessment: The government segment is no longer a ballast line. It is growing at 96% and contributed $144M of sequential growth. The one caution is that RPO excludes most government work by construction (short initial terms and termination-for-convenience clauses), so this segment's forward visibility is lower than the revenue trend suggests. We raise our FY26 U.S. government estimate to +87–90% from +75–85%. That range implies Q4 U.S. government revenue of about $1.02–1.08B, or +79% to +89% year over year and +7% to +13% sequentially, a slowdown from Q3's +96% and +18%. We assume the slowdown for two reasons: the Q4 2025 base stepped up 17% sequentially to $570M, and RPO covers little of this segment. If Q4 instead repeats Q3's sequential pace, the segment grows more than 92% for the year, so the risk to our range is to the upside.

International: Decelerating Again, and Still Not Discussed

International commercial revenue of $184M grew 21% year over year (from +26% in Q2) and 1% sequentially; international government revenue of $192M grew 31% (from +42%) and 6% sequentially. The combined book added $13M in the quarter and its share of revenue fell to 16.3%. On the call, the chief executive said it was better to avoid talking about Europe. Both international lines decelerated year over year.

Assessment: Nine-month international revenue of about $1.09B is running roughly 32% above the prior-year period, and a flat-to-modest Q4 puts the full year near +27–29%, inside the +25–35% range we set in August. The strategic point is unchanged: Palantir is conceding international commercial share while U.S. demand exceeds its delivery capacity. That is still a reasonable trade, and it is still a trade.

Key KPIs

Q1 2026 values marked with a tilde were derived in our prior notes from disclosed growth rates. Year-over-year changes are as disclosed by the company where available.

Three rows carry the analytical weight. First, the coverage ratio: U.S. commercial remaining deal value divided by quarterly segment revenue was 8.3x in Q1, 8.2x in Q2 and 6.6x in Q3. In August we used the stability of that ratio, while the denominator compounded, as the main reason to push the deceleration question out of FY27 and into FY28. It did not hold this quarter. Second, the deal-size distribution moved the other way from Q2: deals above $10M fell to 64 from 73 while deals above $5M rose to 107 from 98, consistent with large commitments that are smaller or shorter at signing. Third, net dollar retention rose 800 basis points to 165%. Existing customers are expanding faster than ever, which is the strongest single argument that the bookings figure understates demand rather than measures it.

Key Topics & Management Commentary

Overall Management Tone: Management was as assured as in August and less specific. Prepared remarks leaned on customer anecdotes and the Sovereign AI frame, the analyst session again ran to three questions, and none of them touched bookings, margins, cash or guidance. The one financial soft spot in the release, the bookings slowdown, was addressed in a single prepared paragraph. The chief executive's closing remarks moved from conviction toward grievance.

1. Revenue Accelerates Again, and U.S. Commercial Passes U.S. Government

Revenue growth of 96% is the highest rate Palantir has reported, on the largest base it has reported, and sequential growth of 19.5% is the highest too. Three $1B thresholds fell in the same quarter: commercial revenue, government revenue and GAAP operating income. The symbolic milestone is the U.S. crossover. U.S. commercial revenue of $984M exceeded U.S. government revenue of $953M for the first time, and total commercial revenue exceeded total government revenue as well. U.S. commercial is now the largest single segment.

"We delivered our highest ever reported year-over-year revenue growth rate of 96% and our highest ever reported sequential revenue growth rate of 20%." — David Glazer, Chief Financial Officer

"Again, it's so extreme that a business that was viewed as a nice-to-have part of our business as late as 3 years ago is now the dominant part of our business." — Alexander Karp, Chief Executive Officer

Assessment: The revenue line confirms every operating claim we have made about this business since February. Nothing in this note argues that the quarter was weak. The argument is narrower: the numbers that forecast FY27 were weaker than the numbers that report Q3, and the price already assumed the reported numbers would keep accelerating.

2. The Bookings Air Pocket

Total contract value of $2,918M grew 6% year over year and fell 13% sequentially. Commercial TCV of $1,652M grew 17%, and U.S. commercial TCV of $1,463M grew 11% against +153% in Q2. Bookings-to-revenue fell to 1.26x for the company and 1.49x for U.S. commercial, from 1.74x and 2.79x in Q2. Deal counts were mixed: 212 deals above $1M, 107 above $5M and 64 above $10M, against 220, 98 and 73 in Q2.

Two pieces of context keep this in proportion. Q2's $2,132M of U.S. commercial TCV was itself a one-quarter spike, $788M above the prior record, so some reversion was always likely. And Q3's figure still exceeds every quarter before Q2. Trailing-twelve-month U.S. commercial TCV of $6,114M grew 62%, which is a deceleration from +117% in Q2 but not a reversal.

"Over the past 12 months, we closed $6.114 billion of U.S. commercial TCV bookings, a 62% increase from the prior 12 months, highlighting the accelerating demand for AI that creates real operational value while also moving power away from the labs and back into your enterprise." — David Glazer, Chief Financial Officer

Assessment: In August we called bookings "the leading indicator that matters" because they had inflected harder than revenue. Consistency requires us to weigh them the same way when they inflect the other way. One quarter is not a trend, and a 62% trailing-twelve-month growth rate is still very strong. But a business priced at 32x forward revenue needs its leading indicator to keep confirming the trajectory, and this quarter it did not.

3. The Duration Explanation, and the Number That Would Prove It

Management's explanation is that customers adopting Sovereign AI signed large contracts on shorter terms than usual, so TCV understates the annual value of what was signed. The mechanism is real: a $30M annual commitment signed for one year books $30M of TCV, while the same commitment on a three-year term books $90M. The chief executive described the same pattern from the field: faster decisions, larger starting commitments, shorter initial terms, then expansion.

"We signed multiple 8-figure deals in the quarter that were shorter in duration than usual, as customers rapidly adopted initial Sovereign AI capabilities in their enterprise. It is important to remember that the Sovereign AI revolution is only 1 quarter old. As we continue to deliver value with Sovereign AI, over time, we expect both the duration and size of these contracts to increase." — David Glazer, Chief Financial Officer

The explanation is testable, and the test was not supplied. On the Q2 call, management said total TCV grew 129% on a dollar-weighted duration basis and U.S. commercial TCV 271%, figures that strip out exactly this effect. The Q3 release still defines the dollar-weighted duration basis, but neither the release, the call nor the 10-Q reports a Q3 figure on it, and no aggregate ACV figure was given. Individual conversions were quoted in ACV terms ($29M, about $21M, $15M, $11.5M), which shows the company tracks the measure. We could not read the CEO letter or the investor presentation for this note, and the release defines the measure for use in the letter, so a Q3 figure may appear there. Every statement in this note that something was not disclosed refers to the release, the call and the 10-Q only.

Assessment: We think the duration explanation is probably a large part of the story. Net dollar retention of 165% and a Q4 revenue guide of +14.6% sequentially are hard to reconcile with a demand problem. But "probably" is not good enough for a stock priced on the forward curve, and the company tracks the number that would settle it. Until we can verify a Q3 duration-adjusted bookings figure, we treat the bookings deceleration as partly real.

4. Remaining Deal Value and RPO Stall; Coverage Falls to 6.6x

The more important consequence of shorter contracts is visible in the backlog measures. Total remaining deal value rose 2% sequentially to $13.4B (+55% year over year, against +83% in Q2), and U.S. commercial remaining deal value rose 4% to $6,517M (+79%, against +124%). Remaining performance obligations were flat at $4.9B, with about 44% expected to convert within 12 months. Total remaining deal value now covers 5.8x quarterly revenue, against 6.8x in Q2; for U.S. commercial the ratio fell to 6.6x from 8.2x.

"We ended the third quarter with $13.4 billion in total remaining deal value, an increase of 55% year-over-year and 2% sequentially, and $4.9 billion in remaining performance obligations, an increase of 88% year-over-year and flat sequentially." — David Glazer, Chief Financial Officer

Assessment: This is the change that moves our view. If contracts are shorter, then more of next year's revenue has to be won next year rather than sitting in backlog today. That raises the dependence of FY27 on continued conversion and renewal at the current pace. Shorter terms do not reduce demand, but they do reduce visibility, and visibility is what supported the multiple. We are moving the FY27/FY28 deceleration risk from CONTAINED to EMERGING on this evidence.

5. Sovereign AI: Pilots Converting at Scale, and New Customers Doing More of the Work

The Sovereign AI frame introduced in August was the organizing theme again, now with conversion evidence attached. Management cited an industrial materials company moving from a nearly $3M pilot to a $29M ACV deal, a global athletic company from a $1.6M pilot to about $21M ACV, a home appliance company from a $1M pilot to $15M ACV within two quarters, and an apparel company from a $2.5M pilot to $11.5M ACV. The 10-Q is consistent: $212M of the year-over-year commercial revenue increase came from customers added after December 31, against $88M in Q2.

"So instead of being, like, $250,000, it's like $25 million to start." — Alexander Karp, Chief Executive Officer

The technology argument was sharpened too. Management's position is that the post-trained model is a commodity and the durable asset is the pipeline that keeps retraining it on the customer's own operational data inside the Ontology. Examples included a 30-billion-parameter post-trained model deployed at a Pacific Command exercise with better task performance than a frontier model, and a defense contractor reporting a 95% cost reduction in preliminary results.

"The specific post-trained model is a commodity item. The compounding value is in the post-training pipeline that is constantly fed from your operational exhaust captured in the Ontology that drives automated post-training runs, model checkpoints, evaluations and deployment across yesterday's, today's and tomorrow's open-weight ecosystem." — Shyam Sankar, Chief Technology Officer

Assessment: The conversion anecdotes are the bull case's best answer to the bookings number, because they are quoted in annualized terms and describe exactly the short-first-contract pattern management claims. They are also hand-picked. The strategic argument is coherent and the revenue supports it. The risk we flagged in August still applies: post-training and hosting customer workloads are more infrastructure-heavy activities than Palantir's historical software, and Topic 7 shows the cost is now visible.

6. Net Dollar Retention at 165%

Net dollar retention rose 800 basis points to 165%, the fourth consecutive increase on our record (134%, 139%, 150%, 157%, 165%). In August the chief executive said the metric would move higher as dormant Foundry accounts migrated onto the newer stack, and it did. Top-20 customers averaged $146M of trailing-twelve-month revenue, up 76% from $83M.

"Net dollar retention was 165%, an increase of 800 basis points from last quarter." — David Glazer, Chief Financial Officer

Assessment: This is the strongest counterweight to the bookings slowdown. A customer base expanding at a 65% net rate before any new logos is not a base with a demand problem. We keep our long-run retention assumption at 135–150% and note that this commitment from August was met in full.

7. Gross Margin Steps Down Again; the 10-Q Names Cloud Hosting

Adjusted gross margin fell to 85.1% from 86.3% in Q2 and 87.9% in Q1; GAAP gross margin fell to 83.4% from 84.7%. On the call, the chief financial officer stated the 85% figure and moved on. The 10-Q gives the cause. Cost of revenue rose $176M year over year, and $135M of that was third-party cloud hosting. The same line rose $89M year over year in Q2 and about $39M in Q1. R&D carried another $21M of incremental hosting cost in Q3.

"Adjusted gross margin, which excludes stock-based compensation expense, was 85% for the quarter." — David Glazer, Chief Financial Officer

In August management described the hosting cost as a trade made for one government customer to speed time-to-value and widen the workflow footprint. The incremental hosting cost has now grown from $39M to $89M to $135M across three quarters, equal to about 5.8% of Q3 revenue. Whether that is one customer scaling or the arrangement spreading to others, the filing does not say.

Assessment: In August we said "a second consecutive quarter of gross margin compression without explanation" would make us reconsider the rating. That condition fired only in part. The compression came, but it did not come without explanation: the 10-Q names the cause, although the call did not. And the explanation is the structural version, not the one-off version: a hosting cost line that keeps scaling. Adjusted gross margin is still up 130 basis points year over year, and operating margin held, so nothing is broken. But this is now a trend, and we lower our gross margin assumption to 84–86%.

8. A Smaller Raise With a Steeper Fourth Quarter

FY26 revenue guidance moved to $8,530–8,534M from $8,150–8,158M, a $378M raise at midpoint (+4.6%) to +91% growth. That is smaller in percentage terms than the 6.5% raises in May and August. The composition matters more than the size, though. The raise is the $151M Q3 beat plus a $227M lift to the implied Q4, so the Q4 guide of $2,649–2,653M implies +14.6% sequential growth and +88% year over year. The August Q3 guide implied only +11.7% sequentially. Management is guiding to a steeper ramp, not a flatter one.

"We are also raising our full year 2026 revenue guidance midpoint to $8.532 billion, representing 91% growth year-over-year, a 9-point increase over our full year 2026 revenue guidance from last quarter." — David Glazer, Chief Financial Officer

Assessment: If the historical 6–8% beat holds in Q4, revenue lands near $2.81–2.86B, which is +100% to +104% year over year, and FY26 finishes near $8.69–8.74B. In August we wrote that a third raise would put FY26 "near $8.6B"; the guide is $8.53B and our estimate is $8.60–8.75B. The guidance philosophy is intact. That is why the downgrade in this note rests on visibility and price, not on the near-term numbers.

9. The Expense Ramp Is Not One Quarter

Adjusted expense of $887M rose 19.6% sequentially and 53% year over year, about $19M above the roughly $868M the Q3 guide implied, on new-graduate hiring and AI platform investment. The Q4 guide implies adjusted expense of about $1,009M, another 13.8% step. Over two quarters adjusted expense will have risen about 36% from Q2's $741M. Revenue is growing faster, so the guided Q4 adjusted operating margin of 61.9% is the highest of the year, and the full-year guide moved to 61.5% from 60.0%.

"Q3 adjusted expense was $887 million, up 20% sequentially and 53% year-over-year, primarily driven by the continued investment in our AI platform and technical hiring as we welcomed our latest class of new grad hires." — David Glazer, Chief Financial Officer

Assessment: August's commitment to watch was whether the Q3 ramp was one quarter or a reset. It is a reset, and the company can afford it: operating margin did not fall in Q3 and is guided up in Q4. On a GAAP basis, operating expenses fell to 36.5% of revenue from 49.2% a year ago. We raise our FY26 adjusted operating margin estimate to 61.5–62% and leave the long-run range at 60–64%.

10. Cash Conversion Softens as Receivables Build

Cash from operations of $1,205M and adjusted free cash flow of $1,216M were both flat on Q2 while revenue grew $378M. Receivables rose $398M in the quarter. The full-year adjusted free cash flow guide rose to $4.8–5.0B from $4.5–4.7B, which after $3.36B over nine months implies $1.44–1.64B in Q4. The company ended the quarter with $10.4B of cash, cash equivalents and U.S. Treasuries and no debt.

"In the third quarter, we generated $1.205 billion in cash from operations and $1.216 billion in adjusted free cash flow, representing margins of 52% and 53%, respectively." — David Glazer, Chief Financial Officer

Assessment: A billing-timing quarter, not a quality problem: the 10-Q attributes the nine-month operating cash flow to revenue growth "partially offset by timing of billings to and payments from customers." The Q4 implied range would be a record. But the free cash flow margin fell 10.5 points in a quarter, and capital deployment is still absent: repurchases were $1.5M for the first nine months of the year.

11. Below the Line: The Equity Marks Reverse

In August we noted that marks on equity holdings had added $0.03 to GAAP EPS and warned that marks reverse. In Q3 other income swung to a net expense of ($25.8M) from a $91.8M gain in Q2. The 10-Q discloses $25M of net unrealized losses on publicly-traded equity securities in the quarter, most of which are subject to short-term sale restrictions, and sizes the book: $185M of publicly-traded equity securities and $217M of privately-held equity securities at September 30.

Assessment: The equity book is about $0.4B against $10.4B of cash and Treasuries, so it is small, and the drag this quarter was about one cent. The more useful point is that Q3 EPS of $0.44 was earned with below-the-line items working against it, so the operating EPS trajectory is better than the headline $0.41 to $0.44 step suggests.

12. Government: Maven Past 100,000 Users, and the Deterrence Pitch

U.S. government revenue grew 96%, and the call's government content was operational: an Army prime agreement for 8 TITAN systems, Maven above 100,000 users across all services, a 25,000-strong builder community, and the DORA disconnected-edge device. The question that reached government asked about the chief executive's letter describing Palantir as part of an "anti-war movement"; the answer was a deterrence argument rather than a business update.

"In Q3, the Army awarded Palantir a prime agreement to produce 8 TITAN systems." — Ryan Taylor, Chief Revenue Officer and Chief Legal Officer

Assessment: The government franchise is in the best shape of our coverage period, and four consecutive quarterly accelerations make it the line we have under-estimated most. Our new range still assumes Q4 growth slows to +79% to +89% on the higher Q4 2025 base, so the risk to it is to the upside. It also has the lowest contractual visibility, which matters more now that the commercial backlog has slowed.

13. Three Questions, and a Closing Grievance

For the second consecutive quarter the analyst session ran to three questions, and none addressed bookings, gross margin, cash conversion, guidance or the 18-month growth commitment from August. The chief executive's closing remarks thanked early retail holders and then went further.

"And part of the delivery here, part of what motivates me is helping to make sure we are fighting for you and, quite frankly, making sure that the people that made fun of you and made fun of us are suffering." — Alexander Karp, Chief Executive Officer

Assessment: The format means the one adverse disclosure of the quarter went unexamined on the call, and investors had to read the 10-Q to learn why gross margin fell. That is a cost to shareholders, not to the company. The closing remark adds nothing to the investment case and keeps the CEO-posture risk on our list.

Guidance & Outlook

The FY26 framework has now been raised three times since February: $7,182–7,198M, then $7,650–7,662M, then $8,150–8,158M, now $8,530–8,534M. In growth terms that is +61%, +71%, +82% and +91%. The third raise is the smallest of the three in dollars and percent, but it carries the largest sequential step in the implied quarter that follows. The company is still guiding to a number it is confident of and beating it by six to eight points.

Implied quarter-over-quarter ramp: the Q4 midpoint of $2,651M is +14.6% on Q3's $2,313M, above the +11.7% the August guide implied for Q3 and below the +19.5% just printed. Year over year the guide steps from +96% actual to +88% guided. A 6–8% beat on the Q4 guide would print +100% to +104%, so on the reported numbers the deceleration is still at least a quarter away.

Street at: pre-print consensus for FY26 sat near $8.22B, so the new guide is about 4% above the Street for the year. For Q4, the one pre-print estimate we tracked was near $2.46B, which puts the new guide roughly 8% above it. The one pre-print FY27 estimate we tracked was near $12.36B, or about +45% on the new FY26 guide midpoint. That estimate does not reflect the chief executive's August 18-month growth sentence, which would have implied a far higher number.

Guidance style: unchanged and conservative. Three consecutive quarterly beats of 6–8% over the company's own revenue guide, after a beat of just under 6% in Q4 2025, and three raises to the FY26 framework in three quarters after the same pattern through FY25. We continue to treat the guide as a floor. Our base case is a 3–5% beat in Q4, which gives FY26 revenue of $8.61–8.67B. Our full FY26 range of $8.60–8.75B runs from a beat of about 2.5% to one of about 8%, the top of the historical pattern.

Analyst Q&A Highlights

The session again ran to three questions. Two answers produced distinct disclosures, so they appear as separate entries below.

Why the Contracts Are Getting Shorter: the Bootcamp Sales Motion

The opening question asked how the Sovereign AI bootcamps convert attendees into customers. The answer gave the field-level version of the duration explanation: prospects arrive frustrated with what they have, decide faster, start with much larger initial commitments, and sign shorter terms that management expects to expand.

Q: "So my question is, with the Sovereignty AI Bootcamps, I mean, I've been to -- it seems like they're really becoming huge for converting customers. Can you just walk through, like, Alex, like you talk to so many of these customers. Like, do they come into a boot camp, just they're not -- they're trying to understand their AI path and then they leave and they're actually a customer?"
— Daniel Ives, Yorkville Ives

A: "And so what, like, Ryan and I are seeing on the front line in commercial is this much quicker interaction, much quicker to deployment, shorter duration of the contract, but then an ability to expand that contract dramatically."
— Alexander Karp, Chief Executive Officer

Assessment: This is the most important sentence in the Q&A for the bookings debate, and it came unprompted. It confirms that shorter terms are a feature of the new sales motion rather than a one-quarter accident. That supports the demand reading of the TCV figure and weakens the visibility reading at the same time: if short initial terms are the new pattern, backlog coverage will run structurally lower than it did in the first half.

Bandwidth, Not Demand, Is the Stated Constraint

Later in the same answer, management said the company does not have the capacity to serve all the inbound demand and must re-resource internally to focus on the customers ready to buy. This repeats August's statement that delivery capacity, not demand, binds growth, but this time without the partner-led delivery plan that accompanied it.

Q: "Like, do they come into a boot camp, just they're not -- they're trying to understand their AI path and then they leave and they're actually a customer?"
— Daniel Ives, Yorkville Ives

A: "What we have to do internally is figure out how to re-resource our company to focus exclusively on those people because we don't have even the bandwidth to do that."
— Alexander Karp, Chief Executive Officer

Assessment: If bandwidth binds, then bookings in any one quarter partly measure delivery capacity rather than demand, which is another reason not to read Q3's TCV as a demand signal. It also means growth depends on scaling delivery, and the August commitment to scale through partners received no update.

The Labs: Competitor, Partner or Market-Maker

A question asked how management persuades executives who still buy AI directly from frontier labs that the orchestration layer should come from Palantir. The answer was that Palantir does not need to persuade them: demand exceeds capacity, the labs are expanding the market, and the relationship is symbiotic. Management said the only actively competitive effort is educating clients on the risks of closed models without an application layer.

Q: "When you get a CEO that's still unconvinced, what kind of analogies and metaphors can you use to help them understand the importance of this diffusion layer, orchestration harnesses coming from you and not directly from the lab? What kind of examples do you have to provide them to understand the risk?"
— Gil Luria, D.A. Davidson

A: "The weird thing here, though, is we really cannot satiate demand. So this is a market that is growing so rapidly. We don't need the people who aren't convinced. We really don't."
— Alexander Karp, Chief Executive Officer

Assessment: The answer is consistent with the numbers but avoids the question. It describes the frontier labs as partners in a growing market rather than as rivals for the same enterprise budget. That holds while the market grows faster than any one vendor can serve it. It is not a durable competitive answer, and it is now the frame through which the company addresses all competition, having not mentioned data-platform vendors on its calls or in its releases for seven quarters.

Why the Workflow, Not the Knowledge Worker, Is the Unit of Value

Near the end of the call, the technology chief returned unprompted to the labs question with the most specific version of Palantir's competitive argument: productivity gains for individual workers do not by themselves produce a return, while automating an enterprise workflow does, and the workflow's data exhaust is what allows customers to post-train a model better than the frontier for their task.

A: "I mentioned this earlier in my comments, but it really is the case that I think one of the conceits that the labs don't have right and that perhaps the customers who don't quite yet realize that the return is not there, is that the atomic unit of the enterprise is not the knowledge worker. You can make people more productive and generate no actual return on investment."
— Shyam Sankar, Chief Technology Officer

Assessment: This is a better answer to the competitive question than the one given in the moment, because it names a specific economic difference between selling seats of intelligence and running workflows. It is also the argument that requires Palantir to host and post-train inside customer environments, which ties the competitive moat directly to the cost line that pressured gross margin this quarter.

The Government Business as Deterrence

The final question asked what the chief executive's "anti-war movement" framing in the shareholder letter means for the government business. The answer was a philosophical case for deterrence and for leadership in AI against China, with no new business detail.

Q: "Alex, in your shareholder letter, you said that you see Palantir as the pioneers of a genuine and viable anti-war movement. Can you expand on what that means and what it might mean for your government business?"
— Brad Zelnick, Deutsche Bank

A: "It's this product for -- our products are built to avoid war, meaning the adversary knows that they will have such pain inflicted upon them that, in fact, it's not worth going to war and also to create social cohesion because we're going to bring back our troops healthier, safer, non-damaged and that makes our society much more effective."
— Alexander Karp, Chief Executive Officer

Assessment: A coherent framing for the defense franchise and no new information for the model. The government segment's results speak for themselves, so the cost of the answer was the opportunity: this was the last question, and bookings and margins went unasked.

What They're NOT Saying

1. A duration-adjusted bookings figure. On the Q2 call management gave TCV growth on a dollar-weighted duration basis (+129% total, +271% U.S. commercial). The Q3 release still defines the measure, but the release, the call and the 10-Q report no Q3 figure on it, and no aggregate ACV. This is the one number that would confirm or refute the shorter-duration explanation for 6% TCV growth.
2. Why gross margin fell, on the call. The 10-Q names third-party cloud hosting (+$135M year over year in cost of revenue, up from +$89M in Q2). The call gave the 85% figure without a cause. Neither the call nor the 10-Q says whether the hosting still covers the one government customer named in August or more, whether it is priced to recover cost, or where it settles.
3. The 18-month growth commitment. In August the chief executive committed to growing "at a rate equal or above" the U.S. commercial pace for 18 months. The release, the call and the 10-Q neither repeated nor clarified it. U.S. commercial grew 148% and the company 96%, and the Q4 guide implies +88%.
4. Capital allocation, for a fourth consecutive quarter. $10.4B of cash and Treasuries, a guide to $4.8–5.0B of free cash flow, no debt, and $1.5M of share repurchases in nine months according to the 10-Q. No plan for the cash was discussed.
5. An FY27 framework. None again in the release, the call or the 10-Q. With contracts now shorter and backlog coverage lower, a framework would carry more information than it did when coverage was 8x.
6. Receivables. Receivables rose 26.8% in a quarter in which revenue rose 19.5%, and free cash flow margin fell from 63% to 53%. Neither was mentioned on the call; the 10-Q attributes cash flow to billing and payment timing in general terms.
7. The partner-led delivery plan. August's call described scaling delivery through third-party partners. This quarter management said the company lacks the bandwidth to serve demand and must re-resource internally. The release, the call and the 10-Q gave no update on partners.
8. International. Both international lines decelerated year over year (commercial +26% to +21%, government +42% to +31%). On the call, management said it was better to avoid talking about Europe.
9. Data-platform competition. Seventh consecutive quarter in which neither the call nor the release mentioned data-platform competition from hyperscalers or data-cloud vendors. The competitive conversation is entirely about the frontier labs, which management describes as partners.

Market Reaction

• Pre-print setup: PLTR closed at $178.34 on November 2, up 0.3% year to date against an S&P 500 up 10.2%, down 11.0% over the trailing twelve months and down 5.5% over the trailing 30 days. The stock set a record close of $209.05 on October 9, 28.5% above the $162.66 at which we published in August, then gave back 14.7% into the print, including 9.2% in the two sessions immediately before it (-6.3% on October 30 and -3.1% on November 2). The 52-week closing range was $107.27 to $209.05.
• After hours, November 2: the stock traded down about 3% on the release and call.
• Reaction session, November 3: opened at $168.18 (-5.7%), traded a range of $165.25 to $176.54, and closed at $175.40, down 1.6% or $2.94. The close was 6.1% above the session low. The S&P 500 rose 0.4%.
• Volume: 53.1M shares against a 30-day average of 22.3M, or 2.4x normal.
• Follow-through, November 4: opened at $164.95, traded $164.60 to $177.43, and closed at $171.70, down 2.1% on a day the S&P 500 rose 1.6%. Over the two sessions after the print the stock fell 3.7% while the index rose 2.0%.
• Where it leaves the stock: $171.70 is down 3.4% year to date, 17.9% below the October record close and 5.6% above our August publication price. Market capitalization on 2,573.9M diluted shares is approximately $441.9B.

The size of the move was modest; its direction is the information. Palantir beat its own guide by 7.0%, inside the band that has held for three quarters, raised the full year for the third time, and guided Q4 roughly 8% above the one pre-print estimate we tracked. On the August pattern that should have been enough. It was not, because the setup was different and because the one genuinely new piece of information in the release was negative. The stock entered the print near flat for the year after a 28.5% run to a record and a 14.7% retreat from it, so holders were sitting on gains and the reported beat was already the base case. The new information was the bookings line. The price action over the two sessions, a gap down, a sharp intraday recovery, then a decline on a strong market day, describes buyers who still want the revenue story and sellers who are using strength to reduce exposure to a slower backlog.

The reaction also confirms the valuation point we made in August. A stock that falls on a 96% growth print with a raised guide is a stock in which the multiple, not the operating results, sets the price. That is what "valuation requires near-perfect execution" means in practice: the execution was close to perfect on the income statement, and imperfect on one leading indicator, and the price weighed the leading indicator more.

Street Perspective

Debate: Is the Bookings Slowdown Duration Noise or a Demand Signal?

Bull view: the shortfall is arithmetic. Shorter initial terms mechanically cut TCV while annual spend is unchanged or higher, the converted deals were quoted in ACV terms that are very large, Q3's U.S. commercial TCV still beats every quarter before Q2, and net dollar retention of 165% plus a Q4 guide of +14.6% sequentially are not what a demand problem looks like.

Bear view: management gave the duration-adjusted number on the Q2 call, when it was flattering, and the release, the call and the 10-Q give none for Q3. Remaining deal value grew 2%, RPO was flat, and coverage fell sharply. Even if demand is fine, shorter contracts mean FY27 revenue must be won rather than harvested from backlog, and that lowers the visibility the multiple is built on.

Our take: the bull side has the better argument on demand and the bear side has the better argument on visibility. We think most of the TCV gap is duration. But a stock at 32x forward revenue is priced on visibility, not on demand, and on that question the bear side scored this quarter. A duration-adjusted figure that we can verify would settle it.

Debate: If a 96% Quarter Cannot Lift the Stock, What Can?

Bull view: one session of mild weakness after a 28.5% run to a record is consolidation, not a verdict. FY26 is now guided above $8.5B, the beat pattern implies Q4 growth near 100%, and the one pre-print FY27 estimate we tracked, near $12.4B, looks low again. On FY27 numbers that keep rising, the multiple compresses by itself.

Bear view: the stock trades at roughly 51x current-year revenue and 88x current-year free cash flow. The operating cadence is fully known and fully priced. With a beat-and-raise already the base case, the asymmetry runs downward: any quarter that is merely excellent, or any leading indicator that wobbles, takes the stock lower, as this one did.

Our take: the bear framing is closer to right over twelve months. We expect estimates to keep rising and the business to stay exceptional. But on the same FY27 revenue estimate the stock is more expensive now (32.0x EV) than when we declined to add in August (30.3x), and the leading indicator is weaker. Rising estimates are now needed simply to hold the price, which is a Hold profile.

Debate: Is Hosting Cost a Temporary Drag or a New Margin Structure?

Bull view: adjusted gross margin is still up 130 basis points year over year, operating margin held at 61.7% and is guided higher in Q4, and contribution margin rose to 73%. Hosting customer workloads buys stickiness and wider deployments. A software company running 85% gross margins and 62% operating margins has no margin problem.

Bear view: the incremental hosting cost rose from $39M to $89M to $135M year over year across three quarters and is spreading into R&D. The product direction (post-training, model factories, hosting inside customer boundaries) makes this cost structural. Two consecutive sequential declines, explained only in the 10-Q, are how margin resets begin.

Our take: this is a real trend rather than a one-quarter artifact, and we lower our gross margin assumption to 84–86%. It does not yet threaten operating margin, which depends on expense leverage the company is still delivering. The risk is the combination: lower gross margin, an expense base guided up 36% over two quarters, and revenue visibility that just fell. Any one is manageable; together they narrow the room for error.

Model Update Needed

Valuation: at the November 4 close of $171.70 and 2,573.9M diluted shares, market capitalization is approximately $441.9B. Net of $10.4B of cash, cash equivalents and U.S. Treasuries, enterprise value is approximately $431.5B. That is 50.6x the FY26 revenue guide midpoint of $8.532B and 32.0x our FY27 revenue estimate of $13.5B. On free cash flow, the FY26 guide midpoint of $4.9B puts the stock at 88x. For a like-for-like comparison with August, the same calculation at our August publication price of $162.66 gives 30.3x our current FY27 estimate. The stock is about 1.7 turns more expensive on FY27 revenue than at the point where we said we were holders, not buyers.

Price target: $190, down from $200. Range $140 to $250, from $150 to $260. The method is unchanged: enterprise value to our FY27 revenue estimate, plus net cash, across about 2.60B diluted shares. Our FY27 estimate rises slightly to $13.5B from $13.4B. The target multiple falls to 36x from 38x because the evidence that justified 38x, stable 8x-plus backlog coverage while revenue compounded, weakened this quarter. At the old 38x the target would be about $201, so the reduction is entirely the multiple. The upside case of $250 assumes FY27 revenue of $14.5B at 44x, reflecting duration-adjusted bookings that prove strong and renewed coverage. The downside case of $140 assumes FY27 revenue of $12.5B at 28x, reflecting the bookings slowdown carrying into revenue in the second half of 2027. The $190 target implies +10.7% from $171.70, against +23% when we published in August.

Thesis Scorecard Post-Earnings

The August Checklist, Graded

Our August conditions are graded as written. The add conditions (a re-test toward $135–145 without thesis impairment, or a print that resolved the gross margin question and clarified the growth commitment) were not met: the post-print low was $164.60, gross margin fell again, and the release, the call and the 10-Q did not address the commitment. In August we wrote: "Our conditions for reconsidering the rating are a second consecutive quarter of gross margin compression without explanation, or a Q3 revenue print that fails to beat the guide by at least the historical 6%." The guide-beat condition did not fire (7.0% against a 6% threshold). The gross margin condition fired only in part. The compression came, for a second consecutive quarter, but the 10-Q names third-party cloud hosting as the cause, so it did not come without explanation. The call gave no cause, and the cost the 10-Q names is growing, not passing. We have reconsidered anyway, and we say plainly that the downgrade does not rest on this trigger. It rests on the bookings slowdown and on valuation, as set out in the Action paragraph below.

Overall: the operating thesis is intact and the visibility thesis weakened. Three bull pillars were confirmed outright, including a government franchise that keeps accelerating and a Rule of 40 that keeps expanding. Against that, the margin pillar stayed at risk with a second gross margin decline, and one bear point escalated: shorter contracts and stalled backlog growth bring the deceleration question back toward FY27. Valuation remains the binding constraint, and the reaction to this print is the clearest evidence yet that it binds.

Action: downgrade to Hold from Outperform, price target $190 (+10.7%). We made the case for Outperform in February on a de-rated entry point and a structural margin step-up, and held it in May and August on durability and visibility. The entry point is gone: the stock is above our August price and more expensive on the same FY27 estimate. The margin step-up has given back two quarters of gross margin. The visibility argument rested on backlog coverage that fell this quarter. Our August gross margin trigger fired only in part, because the 10-Q names the cause, so this downgrade does not rest on it. It rests on the bookings slowdown and on valuation: 32.0x EV to our FY27 revenue estimate against 30.3x on the same basis in August, when we said hold a full position and do not add. What remains is an exceptional business at a price that leaves about 11% to our target, roughly a market return with above-market risk. We would move back to Outperform on either of two conditions: U.S. commercial RDV coverage back above 7.5x together with U.S. commercial TCV growth above 50% on the company's dollar-weighted duration basis, in a release, call or filing we can verify, or a pullback to $140 or below (our downside case) without impairment to the thesis. We would move to Underperform on any one of three: Q4 revenue below $2,705M (a beat of less than 2%, below our estimate range), a sequential decline in U.S. commercial remaining deal value, or adjusted gross margin below 83%. If an upgrade condition and a downgrade condition fire in the same quarter, they offset and the rating stays at Hold.