Datadog Earnings (August 6): A Beat, a Raise, and the Stock Fell 17% Anyway

Show 1 earlier update

TL;DR

• Datadog reports Q2 2026 results Thursday, August 6, before the open, with the call at 8:00am ET. It is the highest-variance print of the week's back half: the last report moved the stock roughly 40% against an 11.5% implied move.
• Consensus sits at the very top of the company's guide: $0.58 of adjusted EPS (guide: $0.57-0.59) on $1.08 billion of revenue (guide: $1.07-1.08 billion), up about 30% from $828 million a year ago. A "meet" is a miss versus positioning.
• The full-year guide hides the real argument: $4.30-4.34 billion of revenue and $2.36-2.44 of EPS implies the second half earns roughly what the first half earned while revenue keeps growing near 30%. Either management is sandbagging again, or margins are about to absorb spend.
• The concentration story: analysts estimate OpenAI at roughly $170 million of revenue, about 60% of Datadog's AI-native cohort, and one preview framed a potential "$150 million hole" if it walks. Any commentary on AI-native customer commitments matters more than the quarter.
• The street is crowded and conflicted: BofA named DDOG its top software pick and Cantor raised its target to $327 on Monday, while Bernstein downgraded in July even as it raised its target to $226 on "exuberant expectations". The stock closed Monday at $273.60, with options pricing about 13%. (Correction, logged in the August 14 scorecard: this page's trade log carried a stale $254.79 spot that never matched Monday's actual close; the entries below now use the correct figure. The scoring conclusion is unchanged, since $273.60 is also what the piece's own "fell 17%" figure above already implies.)

What Time Is Datadog's Earnings Report?

Thursday August 6, before the open, with the conference call at 8:00am ET. Airbnb and The Trade Desk report after that day's close; the full slate is in the week-ahead hub.

The Board

Consensus at the top of the guide, a 13% implied, and a 40% precedent. The reaction hinges on positioning more than the print.

The May Precedent Hangs Over Everything

In May, Datadog beat its guide ($1.006 billion, up 32%, with $0.60 against $0.51 expected) and the stock did something software stocks are not supposed to do: it rose roughly 40% in a session against an 11.5% implied move. That print is why today's 13% implied is not generous, it is scar tissue. It is also why consensus has crawled to the absolute top of the guided ranges: everyone is positioned for the sandbag-and-beat again, which mechanically converts an ordinary in-line quarter into a sell-off.

The full-year frame sharpens it: $4.30-4.34 billion and $2.36-2.44 of EPS implies H2 earnings roughly flat against H1 while revenue grows near 30%. Nobody on the buy side believes that arithmetic at face value. Thursday reveals whether it was conservatism (again) or a genuine spend cycle.

The OpenAI Concentration Question

The uncomfortable estimate circulating: OpenAI at about $170 million of annual revenue, roughly 60% of the AI-native customer cohort, with a bear framing of a "$150 million-plus hole" if it migrates to in-house tooling. These are analyst estimates, not disclosures, but the direction of the risk is real and management knows the question is coming. Alongside it, watch the boring compounding metrics that actually carry the model: $100k+ customers (4,550, up 21% at last count) and AI-workload monitoring adoption.

The Options Angle

About 13% implied, after a realised 40%. This is the single clearest expression this week of the July calibration lesson: realised moves have been beating implied all season, and Datadog is the name where implied has already been humiliated once.

• Selling premium here is the worst trade on the week's board. Documented 3.5x overshoot, crowded positioning, binary concentration question.
• The straddle is the one pre-print volatility buy we take this week. Paying ~13% when the same setup delivered ~40% three months ago, with consensus perched at the top of the guide so that any deviation in either direction gaps, is the rare case where the house calibration lesson and the specific name's history point the same way. Logged below; scored against Thursday's close.

Trade log

#	Stance	Structure	Strikes and expiry	Cost or credit	Spot at writing	Implied move	Breakeven
1	Buy	Long straddle into the print	~$255 line, Aug 7 weekly	Live chain not sourced; costed at the quoted ~13% implied	$273.60, Aug 3 close (corrected)	~13%	needs a move beyond ~13%; May precedent was ~40%
2	Pass	Short premium into the print (any structure)	Aug expiries	Not sourced	$273.60, Aug 3 close (corrected)	~13%	scored on whole position; 40% precedent

The One-Line Read

Datadog beat its own guide by 3% in May and the stock moved 40%, so with consensus now sitting exactly on the top of the new guide and options asking only 13%, Thursday is less a test of the business, which keeps compounding, than of whether lightning-in-positioning can strike the same name twice in one summer.