Palantir Earnings: Revenue Up 93%, Guidance Raised to $8.15 Billion, and the Whisper Bar Cleared With Room to Spare

TL;DR

• Updated August 4: the print is in, and it buried the whisper. Revenue of $1.935 billion, up 93%, against a $1.81 billion consensus that had already parked itself above the company's own guide. Adjusted EPS of $0.41 versus $0.35 expected. That is nine straight beats.
• The year got underwritten, with room to spare. Full-year revenue guidance rose to $8.15-8.16 billion, 82% growth, clearing the street's $7.83 billion whisper bar by more than $300 million. US commercial guidance now calls for at least $3.42 billion, up 134%.
• US commercial revenue grew 149% to $764 million; US government grew 90% to $809 million. GAAP net income crossed $1 billion in a single quarter for the first time, and adjusted free cash flow came in at $1.22 billion.
• The stock rose 12-15% in extended trading, quoted as high as $144.31 against Monday's $125.65 close. That is at, or through, the top of the 10-15% implied move. The straddle this piece passed on is tracking as a miss; scoring below.
• Snap reported the same evening and also beat. The preview framework this piece ran with on Monday afternoon is preserved in the sections below, marked to the result.

The Board

The street stood above the company's guide, and Palantir cleared consensus, guide and whisper in one print.

What Did Palantir Report for Q2 2026?

Revenue of $1.935 billion, up 93% year on year, adjusted EPS of $0.41 against $0.35 expected, and a full-year guide raised to $8.15-8.16 billion.

Hold that against the setup. Management had guided $1.797 to $1.801 billion and consensus had climbed above it to $1.81 billion, which meant printing the top of the company's own range would have scored as a miss. Palantir printed roughly $134 million above its own guide top and about $125 million above consensus, up 19% sequentially on Q1's $1.63 billion. The whisper mechanics that punished Broadcom's +200% AI guide and SK Hynix's record quarter in July resolved the other way here, for the simplest reason: the print did not land on the whisper, it went through it.

The composition matters as much as the headline:

• US commercial: $764 million, up 149% year on year and 28% sequentially. This was the bull case's load-bearing wall: commercial taking the baton before government comps normalise. It did.
• US government: $809 million, up 90%. The engine did not slow while the handoff happened.
• GAAP net income of roughly $1.07 billion, the first billion-dollar GAAP quarter in the company's history, against about $329 million a year ago. Adjusted free cash flow of $1.22 billion. The company put its Rule of 40 score at 155, which is the sort of number that usually contains a typo and this time does not.

The Real Bar Was $7.83 Billion. The Guide Came Back at $8.15

The preview's core argument was that the quarter was not the bar, the year was: the street's whisper case wanted a full-year raise to about $7.83 billion, implying 75% growth, against a prior company guide of $7.65-7.66 billion. A beat with a timid raise was priced to fail.

The raise was not timid. $8.15-8.16 billion, 82% growth, more than $300 million past the whisper case, with Q3 guided to about $2.16 billion, full-year US commercial guided to at least $3.42 billion (up 134%), and full-year adjusted free cash flow guided to $4.50-4.70 billion. The forward-book detail worth checking when the shareholder letter and 10-Q settle is remaining performance obligation, but a guide of this size answers the bookings question at the altitude the market actually trades on.

The Move, Session by Session

Labelled, because this is where earnings coverage usually garbles itself:

• Friday July 31 close: $123.07. The spot this piece's preview quoted "near $123" against.
• Monday August 3 regular session close: $125.65, up 2.1%. The print landed after this close.
• Extended trading Monday evening: up 12-15%, quoted between roughly $140.70 and $144.31 depending on the snapshot's timestamp. That is an after-hours read, not a close.
• Tuesday August 4 is the session that counts. The open strikes the entry for row 3 of the trade log, and the close is what rows 1 and 2 get scored against.

The February Memory, Retired for a Night

The reason this piece refused to treat Monday as a coin flip with a happy bias was the documented failure mode: -18.8% in a month when the market last switched from "pays any price" to "remembers valuation", and a stock still more than a third below its high going into the print. That risk did not vanish, it just was not tested. A 93% growth quarter and an 82% growth guide is precisely the evidence that keeps a hype-sensitive multiple fed. The February playbook goes back in the drawer, unretired, for the next timid guide.

The Options Angle, Scored

The preview logged three positions on Monday afternoon. Here is the original log, followed by how each is resolving.

Trade log

#	Stance	Structure	Strikes and expiry	Cost or credit	Spot at writing	Implied move	Breakeven
1	Pass	Long straddle into the print	$123 line, Aug 7 weekly	Live chain not sourced; quoted implied 10-15%	~$123, Aug 3 quote (session live at writing)	~10-15%	needs more than the implied vs a 7.39% realised average
2	Pass	Short premium into the print (any structure)	Aug expiries	Not sourced	~$123, Aug 3 quote	~10-15%	scored on whole position; consensus-above-guide makes both tails live
3	Conditional	Post-print long (shares or 1-2 month calls) if the full-year guide reaches ~$7.83bn	Struck off the Aug 4 open	Struck off the Aug 4 open	To be struck Aug 4	n/a	Scored against the post-call entry if triggered

Updated August 4, against the extended-hours tape:

• Row 1, the straddle pass, is tracking as a miss. "Palantir's own history says the implied usually overpays here." The extended-hours move ran to the top of the 10-15% implied band, and if Tuesday's close holds anything like it, a straddle bought at the quoted implied roughly breaks even or better and the pass gets scored as a loss. This is the July calibration lesson, again: realised keeps beating implied this season, and "the expected move looks wide" keeps being wrong as a reason to stand aside. Final scoring at Tuesday's close.
• Row 2, the short-premium pass, was right. Any premium seller through the wings is carried out by a 12-15% gap. Scored as a win for the pass.
• Row 3 triggered, with margin. The condition was a full-year raise to about $7.83 billion; the raise printed at $8.15-8.16 billion. Per the row's own terms the entry is struck off the Tuesday August 4 open, and the entry price gets logged in this piece's next update. What was written Monday still holds: the confirmation trade owns a confirmed AI software leader with shares or calls and time on its side, into a week where AMD and Datadog report behind it on the busiest calendar of the quarter.

The straddle pass going in the scorecard as a probable loss, in the same paragraph as a correct short-premium pass, is the whole point of logging both: the two passes were one opinion about direction risk and two different opinions about volatility, and only one of them survived contact with the print.

The One-Line Read

The street parked itself above the company's own number and Palantir cleared the consensus, the guide and the whisper in a single print: 93% growth, the first billion-dollar GAAP quarter, an $8.15 billion year underwritten by management, and the only bad call left on the table was ours, for passing a straddle in a season that keeps paying whoever buys the move.