Palantir FY2026 Q2 Earnings: 93% Growth and 149% U.S. Commercial Surge Force an 82% Full-Year Guide

On August 3, 2026, Palantir reported second-quarter revenue of $1.935 billion, up 93% year-over-year — faster than the 85% it posted in Q1, which was itself its fastest ever — and U.S. commercial revenue of $764 million, up 149% year-over-year. Management then raised full-year 2026 revenue guidance to $8.15 billion, implying 82% growth for the year, up from the 71% guide it issued three months ago. A company guiding to 82% growth at an $8 billion run rate is not decelerating into scale. It is accelerating through it.

The Headline Numbers​

Palantir's fiscal year is the calendar year; Q2 2026 ended June 30, 2026. Every figure below is GAAP, from the 10-Q filed August 4, 2026.

The three lines that matter are not the headline revenue number alone but the relationship between it and the two cost lines above net income. Revenue grew 93%. Cost of revenue grew 63%. Operating expenses (R&D plus SG&A) grew roughly 39% combined. When revenue grows more than twice as fast as any cost line, operating income does what it did here: it more than tripled, from $260 million to $912 million. Palantir did not cut its way to a 47% operating margin. It grew its way there, spending more in absolute dollars on every line — R&D up $50 million, SG&A up $156 million — and watching revenue leave each of them behind.

Then there is the tax line. Palantir earned $1.081 billion before tax and paid $15.4 million of it — an effective rate of 1.4%, identical to Q1. The 55.1% net margin exceeds the 47.1% operating margin by eight full points, and every one of those points comes from two non-operating facts: a balance sheet that pays interest income and a tax shield that makes almost all of the profit tax-free. Both demand a closer look before anyone annualizes that net margin.

Revenue Deep Dive: Two Segments, One American Story​

Palantir reports two segments. Both grew spectacularly, and both expanded contribution margins — but one is now pulling away.

Commercial grew 110%. Government grew 79%. That gap — 31 points — is the thesis being repriced in real time. For years the bear case on Palantir was that it is a government contractor with a software valuation; the rebuttal was that Commercial would converge toward Government over time. This quarter Commercial is 49% of revenue, up from 45% a year ago, and its contribution margin (78%) now exceeds Government's (71%). Commercial is not merely catching up. It is becoming the more profitable half of the company.

Geography sharpens the point further.

U.S. revenue more than doubled. Non-U.S. revenue grew 33% — solid in isolation, and dwarfed by what happened domestically. Within the U.S., commercial revenue hit $764 million (+149% year-over-year) and government revenue $809 million (+90% year-over-year). U.S. commercial alone contributed more incremental revenue this quarter than the entire company generated two years ago.

The correct one-sentence description of this quarter is not "Palantir grew 93%." It is: American enterprises adopted Palantir's AI platform faster than the company could hire to support them, and U.S. commercial is now a $3 billion annualized business growing at triple digits. That is an extraordinary strength and a meaningful concentration risk, and the ledger records both without editorializing.

The Margin Story​

Six periods, straight from the ledger:

Read the operating margin column downward: -26.7%, -8.5%, 5.4%, 10.8%, 31.6%, 46.2%, 47.1%. The inflection from negative to positive happened in FY2023, but the slope changed in FY2025. Palantir added 21 points of operating margin between FY2024 and FY2025 and another 15 points in the first half of FY2026. Gross margin has been stable in the low-to-mid-80s throughout; the leverage is almost entirely below the gross line — operating expenses that scale more slowly than revenue.

In every profitable period, net margin exceeds operating margin. For most companies the relationship runs the other way — interest and taxes shave the net margin below operating. Palantir inverts it, for two reasons visible as exactly two lines in the ledger:

1. No debt and a growing cash pile. About $9.4 billion in cash, equivalents, and short-term Treasuries generates interest income ($77.5 million this quarter) instead of interest expense. The balance sheet is a profit center.
2. An effective tax rate of 1.4%. Q2 paid $15.4 million on $1.081 billion of pre-tax income. Q1 paid $12.2 million on $889 million. The rate has been flat at 1.4% for two quarters running, down from 4–8% in FY2023–FY2024.

In FY2023 interest income alone ($132.6 million) exceeded income from operations ($120.0 million) — the cash was more profitable than the business. That era is decisively over: operating income of $912 million this quarter is nearly 12 times interest income.

The One Big Question: What Happens When the 82% Guide Meets Reality?​

The other number that moved more than revenue was guidance, and the relationship between the two is what makes this quarter worth lingering on.

Palantir raised the full-year guide by roughly $500 million three months after setting it. At an $8.15 billion midpoint, the implied full-year growth rate moved from 71% to 82% — an 11-point increase mid-year, at a scale where most software companies guide conservatively and lower over time. The Q3 guide alone ($2.16 billion) implies roughly 75% year-over-year growth next quarter.

This is not how enterprise software companies normally guide. A $500 million raise in one quarter at 80%+ growth is a statement that the demand signal is so far ahead of the original forecast that incrementalism would have been misleading. The footnotes make the signal more specific: U.S. commercial revenue is now guided to exceed $3.424 billion for the year (+134% year-over-year), closed U.S. commercial TCV reached $2.132 billion in Q2 (+153% year-over-year), and remaining deal value for U.S. commercial sits at $6.238 billion (+124% year-over-year). The pipeline is growing faster than the revenue it has already produced.

The question this creates is not whether FY2026 will be an outstanding year — at 82% growth, it already is — but what the FY2027 comparison base looks like. Palantir is setting a bar for itself that adds nearly $4 billion of incremental revenue in a single year. Repeating that feat in FY2027 would require growing off an $8.15 billion base at a rate that the law of large numbers has historically denied to every software company at this scale. The one exception to that law is the company currently breaking it, and the FY2026 numbers are the evidence. Whether the exception survives its own comparison base is the question that will define the next four quarters.

A related, more mechanical question sits inside the income statement itself: the tax rate.

Pre-tax income grew roughly 7x from FY2023 to the first half of FY2026 on an annualized basis. The quarterly tax bill barely moved. Palantir's 10-Q explains why: its effective rate differs from the U.S. statutory rate "primarily due to foreign income taxed at different rates, non-deductible stock-based compensation, other non-deductible expenses, and valuation allowances recorded on its deferred tax assets." In plain terms — Palantir spent years losing money, built up deferred tax assets it did not expect to use, wrote them down with a valuation allowance, and is now shielding income with those historical losses.

That creates a two-step future: a one-time, non-cash gain when the valuation allowance is released (a quarter with a spectacular and economically meaningless EPS), followed by a permanent step-up in the tax rate toward 21%. At a 21% rate, Q2's $1.081 billion of pre-tax income would have carried roughly $227 million of tax instead of $15 million — cutting net income by about 20%. The accumulated deficit on the balance sheet and the tax shield are related — both are consequences of the same lost decade — but they are not the same number and will not expire on the same day.

Tracking a $9 Billion Company in Plain Text​

Double-entry bookkeeping has an unglamorous virtue: nothing can be quietly left out, because the books do not balance until every dollar is placed. Modeling Palantir in Beancount makes the tax anomaly impossible to skim past, because the tax line sits in the same transaction as the revenue.

In Beancount, income is a negative (credit) amount and expenses are positive (debit) amounts; net income is the balancing figure. Palantir reports in thousands, so this ledger uses three decimal places — every number below is exact to the filed dollar-thousand.

; FY2026 Q2 Income Statement — three months ended June 30, 2026
; Revenue: 1,935,464 | CoR: 296,870 | R&D: 192,513 | SG&A: 534,077
; OtherNet: 169,341 (net other INCOME) | Tax: 15,383 | Net Income: 1,065,962  ($ thousands)
; Check: -1935.464 + 296.870 + 192.513 + 534.077 - 169.341 + 15.383 + 1065.962 = 0 ✓
 
2026-06-30 * "Palantir Technologies Inc." "FY2026 Q2 Income Statement"
  Income:Revenue                              -1935.464 MUSD  ; revenue earned (credit)
  Expenses:CostOfRevenue                        296.870 MUSD  ; cost incurred (debit)
  Expenses:ResearchAndDevelopment               192.513 MUSD  ; cost incurred (debit)
  Expenses:SellingGeneralAdministrative         534.077 MUSD  ; S&M 339,500 + G&A 194,577
  Income:OtherNet                              -169.341 MUSD  ; net other income (credit): interest income 77,505 + other 91,836
  Expenses:IncomeTax                             15.383 MUSD  ; 1.4% effective rate on 1,081,345 pre-tax
  Equity:Adjustments                           1065.962 MUSD  ; net income offset

Notice what the sign convention exposes. Income:OtherNet is negative — it is income, not expense. It contributes $169.3 million of profit that the operating business did not produce. And Expenses:IncomeTax, at 15.383, is the smallest number in the transaction after R&D — smaller than cost of revenue, smaller than SG&A, smaller than every operating cost. A tax bill that is the smallest line in the income statement of a 55% net-margin company is easy to miss in a press release. In a ledger it is simply sitting there, in line order.

The balance-sheet line that carries the story is retained earnings:

2026-06-29 pad Equity:RetainedEarnings                    Equity:Adjustments
2026-06-30 balance Equity:RetainedEarnings                    1629.973 MUSD  ; ACCUMULATED DEFICIT (debit balance)

That balance is positive, and in Beancount's sign convention a positive equity balance is a debit — a hole. Every other company in the Open Ledger carries retained earnings as a credit. Palantir carries a $1.63 billion accumulated deficit: after more than twenty years, its lifetime losses still exceed its lifetime profits, though the deficit has fallen by more than $1 billion in the first half of FY2026 alone (from $3.56 billion at FY2025 year-end). At roughly $1 billion of net income a quarter, the deficit crosses zero sometime in the first half of 2027 — the quarter Palantir will finally, cumulatively, have made money.

The Multi-Year Arc​

Read the growth column downward: +23.6%, +16.7%, +28.8%, +56.2%, +84.7%, +92.9%. Palantir decelerated to 17% in FY2023 and has accelerated every year and every quarter since, at ever-larger revenue. Companies at $4 billion of revenue do not normally re-accelerate; they grind downward as the law of large numbers catches them. The FY2023 trough is the proof that the recent numbers are not a long-running trend line extended — they are a regime change, and the second straight quarter above 80% is what makes it credible rather than a one-off.

The deficit column is the counterweight. It peaked at $5.86 billion at the end of FY2022 and has fallen every period since — the ledger's memory of a decade in which this company burned cash on a scale it is only now finishing repaying. In the first six months of FY2026 alone, the deficit fell by $1.93 billion. One more year at this pace erases it entirely.

The Verdict: Bull vs. Bear​

Bull Case

• Growth is accelerating at scale: 16.7% -> 28.8% -> 56.2% -> 84.7% -> 92.9%. Re-acceleration at $4B+ of revenue is rare and cannot be explained by an easy base.
• U.S. commercial is the engine: $764M (+149% YoY), 78% contribution margin, $2.1B of TCV (+153%), $6.2B of remaining deal value (+124%) — the pipeline is growing faster than current revenue.
• Real operating leverage, not cost-cutting: revenue +93% while every expense line still grew (CoR +63%, R&D +35%, SG&A +41%); total operating income tripled.
• A Rule of 40 score near 155% (93% growth + 62% adjusted operating margin) — the metric was designed to be hard to hit at 40.
• Fortress balance sheet: zero debt, ~$9.4 billion in cash and Treasuries, $1.22 billion of adjusted free cash flow in the quarter (63% margin).
• Stock-based compensation as a share of revenue is compressing, and the commercial/government mix is shifting toward the higher-margin segment.
• Guidance raise of $500M in one quarter (+11 points to 82% full-year growth) signals demand well ahead of the original forecast.

Bear Case

• The 1.4% effective tax rate is temporary by construction. Normalizing toward 21% would have cost roughly $227M this quarter instead of $15M — about 20% of net income.
• The net margin is flattered by the balance sheet: $169.3M of interest and other income is 16% of pre-tax income and has nothing to do with selling software.
• GAAP operating margin is 47%, not the headline 62% — the gap is ~$265M of stock-based compensation, a real cost that adjusted figures exclude.
• Extreme geographic concentration: the U.S. grew 115% while non-U.S. grew 33%. This is a bet on one country's AI adoption cycle.
• Government is still 51% of revenue, and most of it is U.S. government — exposed to appropriations, procurement cycles, and political risk.
• The FY2027 comparison base is now brutal: repeating 80%+ growth off $8.15 billion requires almost $7 billion of incremental revenue that the law of large numbers has denied to nearly every software company at this scale.
• The company has never, cumulatively, earned a dollar: a $1.63B accumulated deficit is the arithmetic record of how expensive it was to get here.

Our Take

The operating story here is real and is being under-described by the bears: a company at this revenue base does not re-accelerate from 17% to 93% growth by accident, and it did it while letting every cost line grow — the hard way and the durable way. The U.S. commercial surge (149%) is not a mix-shift story but a demand-creation story; Palantir is selling a product that a year ago many enterprises had not budgeted for at all. The Rule of 40 score near 155% is not a fluke of one metric — it is what a genuinely high-gross-margin business looks like when demand arrives faster than its cost base.

But investors reading a 55% net margin should understand they are reading three things stacked: a 47% operating margin, a balance sheet that pays them $169 million a quarter, and a tax rate that is 1.4% for reasons that expire. Strip the last two back to normal — a statutory tax rate and no interest windfall — and Palantir is a ~47%-operating-margin software company growing 93%. That is still an outstanding business, and it is a materially different one from the 55% net-margin machine the headline implies. The ledger's job is to make sure you never confuse the two, and the accumulated deficit still sitting on the balance sheet as a debit is there to remind you which one arrived first.