Datadog Guided to a Floor, Not a Forecast: Upgrading on a 14% Sell-Off and the Best Operating Quarter in Coverage

Results vs. Consensus

Q3 2026 Scorecard

Versus Company Guidance (issued August 6, 2026)

This is the fifth consecutive quarter of clearing the revenue guide ceiling by between 3.8% and 4.7%, so the beat magnitude landed precisely inside the established pattern. The operating-income beat did not: 23.8% above the ceiling is more than double the 9.4% of Q2 and the widest in this coverage window. A company that guided to a 23% to 24% operating margin delivered 27.9%.

Year-Over-Year Comparisons

GAAP operating income of $74.7M is the highest the company has reported in at least the sixteen quarters back to Q4 2022, and the 6.2% GAAP operating margin clears the prior best of 4.7% from Q4 2023. The non-GAAP operating margin expanded 451bps while revenue grew 35.2% and operating expenses grew 23.8%, which is the cleanest expression of operating leverage in this coverage window. The single blemish is free cash flow margin, down 322bps year-over-year, and the cause is working capital rather than spending (see Quality of Beat below).

Nine-Month Comparisons

The nine-month view is the fairer read on cash conversion and it is far less alarming than the quarter: a 60bps margin decline against a 322bps quarterly decline. Capital intensity of 3.7% of nine-month revenue sits below the 4.0% of a year ago, and the full-year guide for capital expenditures plus capitalized software was cut to 3% to 4% of revenue from the 4% to 5% guided in May. The deterioration is in receivables, payables and deferred revenue, not in what the company is buying.

Quarter-Over-Quarter Comparisons

The sequential dollar add of $75.6M is well below Q2's record $115M. It is also the second-largest sequential add in company history, and Q3 absorbed a new element the company described on this call: several large AI accounts renewed early in the quarter, with one more in October, at higher commitments and lower unit rates, which pulls revenue down in the near term while raising committed dollars. The $380M sequential increase in remaining performance obligations is where those dollars went.

Quality of Beat

Revenue. High quality on every decomposition available. Growth accelerated on the core cohort excluding AI customers, accelerated again excluding the largest customer, and accelerated in the share contributed by new logos, which reached roughly 30% of year-over-year growth against 25% a year ago. Gross revenue retention held in the mid-to-high 90s and net revenue retention held in the low 120s. The core did not need the largest account: the customer reduced usage in Q3 exactly as pre-announced in August, and the business still beat the ceiling by 4.55%. The company declined to say how much of that beat came from the account billing above the Q3 assumption. The one honest qualification is that the reduction landed while the account was still billing above its minimum, so the Q3 print does not test the floor the Q4 guide now assumes.

Margins. The most improved line in the print, and improved for structural reasons. Non-GAAP gross margin of 81.2% recovered 160bps sequentially and is now flat year-over-year, which ends the three-quarter erosion sequence that this coverage has flagged in each of the last two notes. Non-GAAP operating expenses grew roughly 25% against 26% last quarter and 32% a year ago, extending the deceleration we identified in August as the first real operating leverage. The combination produced 451bps of year-over-year margin expansion and a 501bps sequential expansion, and the full-year operating margin guide moved to 24.3% from the 23% set in August, roughly 22% in May and roughly 21% in February. That is the third consecutive quarter in which the company raised its own full-year profitability target, on a print the market treated as bad news.

EPS. Operationally driven, with the usual dilution drag. Non-GAAP EPS grew 47.3% against 61.2% non-GAAP operating income growth. The 14-point gap is a 3.6% increase in diluted shares to 375.0M plus an other-income line growing more slowly than operating income. Share count remains a real and rising cost: Q4 guidance assumes approximately 379M diluted shares against 365.5M in Q4 2025, and most of the increase sits in basic shares rather than in convertible dilution: Q3 basic weighted shares rose 10.0M year-over-year, of a 13.0M rise in diluted shares. Stock-based compensation fell to 19.8% of revenue from 22.6%, a third consecutive quarter below 20%.

Cash flow. The weak spot, and it needs to be stated plainly because we gave this pillar a clean bill of health last quarter on the same metric moving the other way. Free cash flow margin of 20.9% is 322bps below the year-ago quarter, and operating cash flow grew only 15.3% against 35.2% revenue growth. The cause is working capital, not capital intensity: deferred revenue drew down $37.3M this quarter against a $7.0M build a year ago, a $44.3M adverse swing, and accrued expenses swung $33.6M against. Capital expenditures plus capitalized software grew only 4.8% and the full-year guide was cut. Trailing-twelve-month free cash flow margin is 26%. The deferred-revenue drawdown is the item worth tracking, because it is the same phenomenon visible in billings growth decelerating to 29.9% from 38%.

Platform & Cohort Performance

Revenue Cohort Mix

The core cohort

The series we have published since initiation now reads 18%, 20%, 23%, mid-20s, high-20s, and high-20s plus roughly a percentage point, which management characterises as the fifth consecutive quarter of acceleration. The increment this quarter was the smallest of the run. Whatever the count, the pillar we opened at initiation has now been confirmed at every single grading.

"Revenue growth from our non AI customers accelerated for the 5th straight quarter was in the high-20s percent year over year. Up about 1 percentage point from last quarter up from about 20 percent in the year ago quarter." — Olivier Pomel, Co-Founder and CEO

Pressed on whether the run can extend another quarter, management declined to forecast and instead defended the level rather than the slope.

Assessment: The honest read is that the slope is flattening even as the level holds. A one-point increment after a five-point run is deceleration in the second derivative, and management's answer ("it is a very high growth rate already, and we feel pretty good about it") is a level argument, not a slope argument. That is appropriate, and it is also the first quarter in which this metric stopped surprising upward. The pillar is intact; its rate of improvement is not.

The AI cohort, and the account inside it

The cohort broadened materially: more than 850 AI customers against roughly 750 a quarter ago, 39 of them above $1M of annual spend against 31, and all 15 of what management calls the top 15 AI leaders now customers. The company also disclosed that roughly a third of its entire customer base, and roughly three quarters of its $100,000-plus customers, use at least one Datadog AI product. On the largest account specifically, the disclosure moved in a direction that was better than feared.

"As we communicated in August, our largest customer reduced their usage in the third quarter, while their usage remained above their minimum commitment." — David Obstler, CFO

Later in the call the CFO strengthened that from "above" to "substantially above," and then conceded the arithmetic consequence for Q4 directly.

"We said we would be conservative in Q3. We said conservative relative to what they were spending. And in Q4, we have been explicit and said minimum commitment. It is true that if they spend more than that, that would be an implicit beat, mathematically." — David Obstler, CFO

Assessment: This is the most useful disclosure on the call and the market appears to have ignored it. The Q3 guide was set conservatively against observed spending and produced a 4.55% beat. The Q4 guide is set at a contractual floor, which is a structurally more conservative assumption than the one that just produced a 4.55% beat. Management has not quantified the gap, which is the legitimate complaint, but it has told investors the direction and the mechanism. The eight-customer count above $10M was not refreshed this quarter, which is a small step backward in a cohort disclosure that is otherwise richer than last quarter's.

The early-renewal repricing, which is new this quarter

The genuinely new item in the print is not the largest customer. It is a pattern management described at several large AI accounts: renewing early, extending duration, raising the committed dollars, and cutting the unit price. The effect on reported revenue is immediate and negative; the effect on committed revenue is immediate and positive.

"the only thing that has changed was that in sort of Q3 ... we did some renewals extensions with some large and large AI, customers where as normal, we essentially extend and increase the commit but do better pricing for that. And that, that happened in a number of large customers." — David Obstler, CFO

"I could bring up a few of those large customers. Where we saw renewals way earlier than we thought we would. there is another 1 that happened in October that we are now discussing. So these are all good news, but, they do create, like, headwind into the short term into the revenue. And, that is 1 change from last quarter." — Olivier Pomel, Co-Founder and CEO

Management was explicit that the mechanism runs through price rather than demand, and that the commitments went up rather than down.

"But I think it is important to clarify. These are, renewals, but they are at higher levels than their previous commit, which is how they are getting you know, the volume discounts. But we are we are renewing these customers at higher commits." — David Obstler, CFO

Assessment: On the facts disclosed this is the healthiest possible reason for a revenue air pocket: customers growing fast enough to want more capacity before their contracts expire. It also fully explains the otherwise puzzling split in this print, where billings growth decelerated eight points to 29.9% while remaining performance obligations jumped $380M after three flat quarters, even as year-over-year RPO growth slowed to 38% from 43%. Lower unit prices on larger, longer commitments produce exactly that signature. The debit is that it is now the second revenue variable management will describe but not size, and management conceded it adds volatility to quarterly guidance. We would rather own this problem than a demand problem, and we would rather own either one at 14.1x than at 17.5x.

Platform Adoption Density

Assessment: Every density band widened again, and the 10-or-more band exactly doubled year-over-year, from 8% to 16%, after nearly doubling from 7% to 13% last quarter. The infrastructure monitoring disclosure is the new one and it is the most impressive single number in the release: a product line first shipped more than ten years ago crossing $2.0B of annual recurring revenue while growing in the mid-30s, against the $1.6B disclosed in February. Set against that, total customer count added only about 100 sequentially, the weakest net addition in this coverage window for a second straight quarter. The pattern is consistent and by now well understood: the paying base is deepening fast while the free-to-paid boundary churns with negligible revenue effect. The headline customer count has become a low-information metric.

Key Operating Metrics

Assessment: Read billings and remaining performance obligations together and they stop being contradictory. In August we flagged that total RPO had been flat in dollars for three consecutive quarters at $3.46B, $3.48B and $3.47B while billings accelerated to 38%, and we wrote that nobody on the call had raised the tension. This quarter it inverted: RPO jumped $380M to $3.85B while billings decelerated to 29.9%, although year-over-year RPO growth still slowed to 38% from 43%. Both moves are the signature of the early-renewal repricing management disclosed, and both are visible in the deferred-revenue drawdown on the cash flow statement. Current RPO growth decelerating to the mid-30s while duration lengthens says the incremental commitment is longer-dated, which is the same story again. The flat-RPO worry from August is resolved, and it was resolved by the mechanism that produced the weak guide.

Key Topics & Management Commentary

Overall Management Tone: Confident on the business and conspicuously unbothered by the guide, which is a change from the defensive posture of August. Management volunteered the new repricing dynamic unprompted in the first Q&A answer, characterised it as "all good news" with a short-term revenue cost, and twice declined to quantify any of it. Where the call weakened was on the one thing investors most needed, which is any framework for sizing either unquantified variable; the answers there were procedural rather than analytical, and the CFO's appeal to Q4 seasonality is not supported by the company's own revenue record.

1. The Q4 Guide Embeds a Contractual Floor, Not an Expectation

The single most important sentence in the prepared remarks was a change in guidance methodology for one account. In Q3 the company guided conservatively relative to what the largest customer was actually spending. In Q4 it guides to what the customer is contractually obliged to spend.

"Regarding our largest customer, our Q4 guidance contemplates their revenue at their minimum commitment." — David Obstler, CFO

In the quarter just reported, that same customer spent substantially above its minimum. Management also disclosed what the rest of the business is guided to do without it: fourth-quarter revenue growth excluding the largest customer is implied in the low-to-mid 30s, against 25% to 26% for the total.

Assessment: The gap between a floor and an expectation is the entire valuation question on this print. If the largest customer bills anywhere above its minimum in Q4, as it did in Q3, the guide beats mechanically, and management said so. Applying the company's own five-quarter pattern of clearing the ceiling by roughly 4% to the $1.20B top end implies about $1.248B, which would be 4.3% sequential growth and 30.9% year-over-year. We do not underwrite that as a base case, but it is the arithmetic the market declined to perform.

2. The Guide Would Be the First Sequential Decline in Company History

At the $1.195B midpoint, Q4 revenue would be 0.17% below Q3. We pulled the full quarterly revenue series to test what that means in context. Every quarter from Q4 2018 through Q3 2026 grew sequentially, so a declining quarter would be the first in the company's public record. More tellingly, the fourth quarter has been among the strongest sequential quarters of each year.

The 2020 and 2021 Q2-to-Q3 rates are shown as n/a because we could not tie them to filings on hand. The Q3-to-Q4 rates for those years come from the full quarterly revenue series.

Management offered seasonality as one of three explanations for the shape.

"I do wanna remind everybody that all things being equal, the sequential between Q3 and Q4, tends to be less because of, the late December effect. When our usage based products and vacations so and some code freezes." — David Obstler, CFO

Assessment: This is the one management claim on the call that the record contradicts. Fourth-quarter sequential growth exceeded third-quarter sequential growth in 2020, 2021, 2023 and 2025, and matched it in 2022 and 2024. There is no "late December effect" visible in eight years of revenue. That matters analytically, and in the company's favour: if seasonality explains none of the deviation, then essentially all of the roughly seven-point gap to the Q4 pattern is the two disclosure choices, the minimum-commitment assumption and the repriced renewals. A guide depressed by stated assumptions is a different object from a guide depressed by demand, and only the first one mean-reverts.

3. Early Renewals Are the New Variable, and They Are Growth-Funded

Management introduced a dynamic that did not exist in the August call: large AI customers coming back before their contracts expire to raise commitments in exchange for better unit pricing. The question that drew out the mechanics got an unusually clear answer.

"Yeah. that is that is totally what happened. there is a few other things that typically happen with situations ... So there is a bit of account optimization there when we work with them. But the tech DCV is on the background of growth. These customers are growing, and they are deploying much larger workloads, and they are looking at doubling, tripling, quadrupling their footprint with us over time." — Olivier Pomel, Co-Founder and CEO

Asked whether these renewals, like the largest customer, are also guided at their minimums, the CFO said they are guided like the broader customer base: take the recent organic growth rate of the customer and discount it. That places them on a discounted-growth basis rather than at a contractual minimum, the treatment reserved for the largest account.

Assessment: Two things are true at once. The mechanism is benign and arguably bullish, because a customer renews early only when it expects to outgrow its commitment, and the commitments went up. And the disclosure is worse than last quarter, because the model now carries two unquantified revenue variables where it carried one. Note the asymmetry this creates in the guide: the repricing headwind is fully loaded into Q4 revenue, while the volume growth that justified the lower price arrives later. That asymmetry unwinds in the company's favour over the following two to three quarters.

4. The Operating Beat Was the Widest in This Coverage Window

Non-GAAP operating income of $334.3M exceeded the guide ceiling by 23.8%, against a revenue beat of 4.55%. Operating margin of 27.9% came in roughly 400bps above the guided range. Two things produced it: gross margin recovering 160bps sequentially, and non-GAAP operating expense growth decelerating again to roughly 25%.

"Q3 OpEx grew 25 percent year over year versus 26 percent last quarter and 32 percent in the year ago quarter. And Q3 operating income was $334 million or 28 percent operating margin, compared to 23 percent last quarter and 23 percent in the year ago quarter." — David Obstler, CFO

The full-year operating margin guide moved to 24.3% at the midpoint, from 23% in August, roughly 22% in May and roughly 21% in February.

Assessment: For four quarters this coverage wrote that the operating leverage story was deferred while the company reinvested an accelerating top line. Last quarter it stopped being deferred and this quarter it compounded: 451bps of year-over-year expansion and 501bps sequentially, with the full-year guide raised for a third consecutive quarter. A 24% full-year operating margin with revenue growing 32% is a materially better financial profile than the one we were valuing in May, and the market paid nothing for it on the day.

5. Gross Margin Recovered, Reversing Three Quarters of Erosion

Non-GAAP gross margin of 81.2% is up 160bps sequentially from 79.6% and now flat against the year-ago quarter. This closes a sequence we flagged in each of the last two notes: 81.4% in Q4 2025, 80.2% in Q1, 79.6% in Q2, and three consecutive year-over-year declines.

"through a combination of the penetration of our platform as well as our engineering we have been, you know, plus or minus in this 80% range. I think you see in this quarter, we ended up around 81. And so this is common practice for us." — David Obstler, CFO

Assessment: We treated the three-quarter decline as a trend rather than variation and said the "varies quarter to quarter" framework was a floor commitment rather than an explanation. The framework has now been vindicated on its own terms, and we should say so: the erosion reversed within one quarter and without a change in the explanation. The watch item closes. The open question that remains is the one management still has not addressed, which is why the margin moved 180bps down over two quarters and then 160bps back up while the workload mix shifted toward training and petabyte-scale logging.

6. Cash Conversion Gave Back the Q2 Gain, on Working Capital

Free cash flow of $250.6M at a 20.9% margin is 322bps below the year-ago quarter, after a quarter in which the margin rose 485bps year-over-year. The cash flow statement locates the shortfall precisely: deferred revenue drew down $37.3M against a $7.0M build a year ago, and accrued expenses swung $33.6M against. Capital intensity moved the other way, with capital expenditures plus capitalized software up only 4.8% and the full-year guide cut to 3% to 4% of revenue from 4% to 5%.

"Quarterly, the timing of collections and payments vary. And on a trailing 12 month basis, free cash flow margin was 26 percent." — David Obstler, CFO

Assessment: We moved this pillar to Contained in August on a 485bps year-over-year improvement, so symmetry requires moving it back on a 322bps deterioration, and we are doing that. Two qualifications matter. The nine-month margin is down only 60bps, and the component that worried us originally, capitalized software and capital expenditures, improved on both the actual and the guide. The deferred-revenue drawdown is not a collections problem; it is the accounting shadow of the same repriced renewals that depressed the revenue guide. That makes it informative rather than alarming, but it does belong in the pillar.

7. Scout 1.0 Makes Datadog a Model Trainer

The company introduced Scout, a reasoning model it trained itself on in-house distributed reinforcement learning infrastructure, optimized for site-reliability tasks and now powering Bits investigations. The claim is specific and checkable in principle: frontier-level performance on alert investigations at a fraction of frontier inference cost.

"Last month, we introduced Scout, our own reasoning model specifically optimized for SRE tasks ... First, it performed on par or even better than frontier models. On alert investigations. At a fraction of the inference cost. This is valuable, as it allows Bits to investigate many more alerts, start earlier, and go deeper on incidents As an always on first responder for our customers." — Olivier Pomel, Co-Founder and CEO

Asked how the proprietary-versus-open-weight mix evolves and what it means for margin, management described a three-tier approach: frontier models where only they suffice, open-weight models where they are good enough, and post-trained proprietary models where the company's own data creates an efficiency edge.

Assessment: This is the most consequential product disclosure of the quarter and it is underappreciated. The Adaptive ML acquisition closed in June and has produced a shipped model in one quarter. Strategically it converts Datadog's data asset into a cost advantage rather than only a product advantage, which is the first credible answer to the question of what prevents a frontier lab from building observability agents directly. It also has a gross margin dimension: if the agent layer runs on a cheap in-house model rather than purchased frontier inference, the unit economics of AI features improve structurally. We see no revenue attached to this yet and we are not modelling any.

8. Bits AI Gets a Packaging Model, Deliberately Priced Near Cost

Five quarters after this coverage first asked how Bits would be monetized, the company disclosed the model: a shift from charging per investigation to AI credits spanning the agent's full set of use cases. Management also disclosed the pricing philosophy, which is the more revealing half.

"we also have a model where we charge for AI credits specifically for what Bits does on top of that. And the model we have chosen is a model where we stay fairly close to inference cost there so that we can pass on the savings ... In terms of what becomes the dominant monetization, it is hard to tell. And we do not really care. Again, we are fully user based." — Olivier Pomel, Co-Founder and CEO

Management explained the choice by reference to the pace of change: pricing is held near inference cost so that efficiency gains in the underlying models pass through to customers rather than accruing to Datadog, on the view that investigation volumes rise by orders of magnitude while unit costs fall by orders of magnitude. Bits was described as the fastest-adopted product in company history, with Bits testing reaching general availability in October.

Assessment: The question we have been asking for five quarters is answered, and the answer is that Bits is deliberately not a profit centre. Pricing near inference cost makes it an adoption vehicle for platform consumption, which is coherent with a usage-based model and mildly deflationary for blended gross margin at the AI layer. That reframes what to look for: the metric that matters is not Bits revenue but whether Bits-adopting customers expand platform consumption, which is the empirical claim management made last quarter and which still has not been quantified. We are retiring the "Bits revenue disclosure is overdue" complaint and replacing it with a request for the attach data.

9. GPU Monitoring Is Explicitly Not in the Numbers

Asked how much of infrastructure monitoring's mid-30s growth comes from AI infrastructure, management gave an answer that cuts against a conclusion this coverage drew two quarters ago.

"Today, that growth is not actually related directly to GPUs. The you know, we. GPUs are still largely unseen for our customers on training. We have products there, but we do not monetize those products widely yet. And so for us, that is an opportunity we have in front of us. that is not something that is baked into the current growth and the current numbers." — Olivier Pomel, Co-Founder and CEO

Against that, the usage signals are extraordinary: GPU devices monitored up eightfold so far this year, spans sent to agent observability up ninefold, and MCP tool calls up fiftyfold.

Assessment: A correction to our own record is required. In May we upgraded the AI-observability pillar from At Risk to On Track on what we called the first hard revenue evidence, citing GPU monitoring anchoring seven- and eight-figure training landings. Management has now said plainly that GPU products are not widely monetized and are not in the growth. We are keeping the pillar On Track, because the evidence base has broadened considerably in other directions, but the revenue rationale we published in May does not hold and the pillar now rests on adoption breadth and a disclosed packaging model rather than on revenue. Read constructively, an eightfold increase in monitored GPU devices that generates no meaningful revenue is unpriced optionality rather than a failure. It is also, for now, unbankable.

10. Distribution Through the Frontier Labs

Agent observability is now available in the Anthropic marketplace, and Datadog is a launch partner for OpenAI's new enterprise marketplace. Customers can buy the product through the model providers they already work with.

Alongside that, the AI cost-management disclosure was the most striking competitive datapoint in the prepared remarks: beyond the three major clouds, AI providers are now the largest single category of spend that Datadog customers track in cloud cost management, with the number of customers tracking AI costs up more than elevenfold and more than a third of cloud-cost-management customers doing so.

Assessment: Two separate strategic points sit here. The marketplace listings are a channel that did not exist a year ago and they place Datadog inside the procurement path of the labs rather than beside it. The cost-management data is the better insight: if AI providers are already the largest tracked spend category after the hyperscalers, then Datadog is instrumenting the AI budget itself, which is the position from which the next cycle's consolidation budgets get spent. Neither is monetized enough to model.

11. The Enterprise Land Motion Widened Beyond Technology

The deal list was unusually diversified away from the AI-native cohort that has driven the last two years. A five-year seven-figure annualized deal with a United States federal defense agency starting on nine products, following FedRAMP High certification earlier in the year and with Impact Level 5 accreditation in progress. Seven-figure annualized deals with two of Japan's largest banks, both adopting the full platform. A seven-figure annualized deal with a home health care software company consolidating four tools. A seven-figure expansion with an online entertainment company growing to 23 products. And an eight-figure expansion with the AI division of one of the world's largest technology companies, nearly tripling its commitment two quarters after landing.

"And by the way, those wins I mentioned, they do not they do not even register into the revenue yet, you know, because these are these are plans, but these are not customers that are ramped users yet." — Olivier Pomel, Co-Founder and CEO

Assessment: The public-sector and foreign-bank wins are the ones that matter for the medium-term thesis, because they are regulated, slow-moving, multi-year budgets that are structurally insensitive to the AI capital-spending cycle the market is currently worried about. The explicit note that none of these wins is in revenue yet is the counterweight to the weak Q4 guide, and it is the kind of detail that gets lost on a day when the stock falls 14%. Federal and Japanese banking are also the two hardest places to displace an incumbent once installed.

Guidance & Outlook

Decomposition of the raise. The full-year revenue midpoint rose $60M while the Q3 beat against the guide midpoint was $57.1M. Only about $2.9M of the raise is therefore forward. That is a sharp break from the pattern: roughly $93.5M of Q2's $140M raise was forward, and roughly 79% of Q1's $240M raise was. On operating income the forward component is larger, about $10.7M of an $80M raise, and on EPS about $0.05 of $0.22. In plain terms, the company passed the quarter's beat through to the year and raised the forward revenue outlook by essentially nothing, while raising the forward profit outlook modestly.

Implied sequential path. The Q4 midpoint of $1.195B requires 0.17% sequential contraction, against +7.6%, +6.9%, +7.7% and +7.5% in the four prior fourth quarters. At the company's established pattern of clearing the quarterly ceiling by roughly 4%, Q4 would print near $1.248B, which is 4.3% sequential and 30.9% year-over-year. The distance between those two outcomes, roughly $53M of revenue, is the entire argument over this print.

Street at. Consensus entering the print sat near $1.202B for Q4 revenue and $0.66 for Q4 EPS. The revenue guide midpoint is therefore about 0.6% light and the EPS guide is roughly 6% ahead of where the street sat. Describing this as a guidance miss is a stretch on the numbers; the market reacted to the sequential shape, not to the gap against consensus.

Guidance style. Unchanged in method and more conservative in one input. Management restated that it guides on recent observed trends with conservatism applied, and then applied a stricter assumption to the largest account than it did a quarter ago, moving from "conservative relative to what they were spending" to the contractual minimum. The full-year revenue guide has now been raised at every quarter of 2026, from 18% to 20% growth in February, to 25% to 27% in May, to roughly 30% in August, to 31.9% now.

Other guidance notes. Full-year net interest and other income of approximately $200M, cash taxes of $30M to $40M, and a 21% non-GAAP tax rate maintained for 2026 and forward. Guidance for 2027 was explicitly deferred to the year-end call, including any framework for the largest customer. Still no free cash flow guidance, now a fifth consecutive quarter.

Analyst Q&A Highlights

What Changed in the Q4 Assumptions Versus Ninety Days Ago

The first question of the call asked whether the outlook reflected a changed view of the business or simply the minimum-commitment methodology. The answer produced the only genuinely new disclosure of the quarter, volunteered rather than extracted.

Q: "I want to start with the Q4 guide. If we went back 90 days ago, have anything changed respect to underlying assumptions between the different cohorts, so between the core business and the sort of largest customer versus 90 days ago, or is this outlook, just sort of a reflection of that sort of minimum commitment methodology that you are applying in Q4?"
— Sanjit Singh, Morgan Stanley

A: "the only thing that has changed was that in sort of Q3 ... we did some renewals extensions with some large and large AI, customers where as normal, we essentially extend and increase the commit but do better pricing for that."
— David Obstler, CFO

A: "there is another 1 that happened in October that we are now discussing. So these are all good news, but, they do create, like, headwind into the short term into the revenue. And, that is 1 change from last quarter. Other than that, like, everything else played out pretty much to where we were expecting."
— Olivier Pomel, Co-Founder and CEO

Assessment: The most valuable exchange on the call. Management said that exactly one thing changed in the forecast inputs, that it is a price effect on growing accounts rather than a demand effect, and that everything else "played out pretty much to where we were expecting." An investor who believed the Q4 guide signalled deteriorating demand had that reading refuted in the first answer and the stock still fell 14.4%.

How to Model the Sequentials From Here

A question acknowledged that the cohort's growth has made sequential modelling unreliable and asked when the pattern normalises. The answer identified the controlling variable and conceded it is outside the company's control.

Q: "what can you tell us about how to think about modeling the sequentials in this business? Is it that we get through the next, call it, 2 quarters and then we are back to steady state? Is it that maybe we see another step function down every year in the sequential because you have this earlier renewal dynamic, which again is a really good place to be because you are getting these really large deals that the only way to get better pricing is because the dollar volume is going up."
— Gabriela Borges, Goldman Sachs

A: "So the customer in Q3 spent substantially above the minimum commitment. So 1 thing that is very important in the sequentials is what happens there ... essentially, 1 thing that will create less of a sequential guidance or actual would be that number stabilizes and is not running above the minimum commitment."
— David Obstler, CFO

Assessment: The phrase "substantially above the minimum commitment" is the most load-bearing disclosure of the call and it was given away in a modelling question. It converts the Q4 guide from a forecast into a floor with a known direction of error. The question also framed the repricing correctly, which is that a customer only earns volume pricing by raising volume.

Whether the Repricing Is a Temporary or Recurring Drag

A follow-on exchange tested the exact shape of the headwind: lower price on existing volume now, growth on top later. Management confirmed the model and added the account-optimization detail.

Q: "if I think about the customers renewed early, so they get better pricing ... but it kind of creates a short term headwind because the same volume is now on a lower price. But then the volume grows. And from then on, in a couple of quarters, we should see like the normal progression. Is that kind of in the right thing to think about?"
— Raimo Lenschow, Barclays

A: "Yeah. that is that is totally what happened ... But the tech DCV is on the background of growth. These customers are growing, and they are deploying much larger workloads, and they are looking at doubling, tripling, quadrupling their footprint with us over time."
— Olivier Pomel, Co-Founder and CEO

Assessment: An unambiguous confirmation that the drag is a one-time price reset on a growing base, with normal progression resuming within a couple of quarters. This is the single best reason to look through the Q4 guide, and it was obtained by an analyst proposing the answer rather than by management offering it. The caveat is that management also disclosed another such renewal in October, so the reset is staggered rather than concluded.

How Much of the Q3 Beat Came From the Largest Customer

A question went directly at the decomposition that would let investors size the account. It was refused, but the refusal carried a concession.

Q: "I know you mentioned your largest customer did spend above their commit in Q3. There any way to tell us how much of the beat relative to your guidance in Q3 because of the large customer spending above the commit rate versus the rest of the business, depending on how you forecasted them in Q3?"
— Samik Chatterjee, JPMorgan

A: "We are not-- we are not answering. We are not being specific, but it is true. It-- It is true that if you are basically guiding, like, below their-- We said we would be conservative in Q3 ... And in Q4, we have been explicit and said minimum commitment. It is true that if they spend more than that, that would be an implicit beat, mathematically."
— David Obstler, CFO

Assessment: Five quarters of asking how large this account is have produced five refusals, and this is the most frustrating one because the company clearly has the number and used it to set the guide. The concession is nonetheless material: the CFO stated on the record that Q4 beats if the customer spends above its minimum, which it did last quarter. That is as close to a pre-announced beat as a company can get without quantifying it.

Whether the Largest Customer's Consumption Pattern Changes the Long-Term View

A question tried to convert "substantially above minimum" into a forward argument about new use cases at the account. Management declined to project and restated its standing position.

Q: "on your largest customer, you used the word substantially above their minimums, which I find interesting. And I know you are assuming, you know, only the minimums for Q4, but philosophically, looking at this customer and their recent consumption patterns, does this perhaps change your opinion about how they might use Datadog longer term, maybe new use cases above and beyond what they are using today that could make that growth opportunity really interesting for this customer?"
— Matthew Hedberg, RBC

A: "I think that is true for any customer that, you know, there are many different use cases. For that in particular, we. As we said all along for many quarters, we do not control. We have the contract, the renewal, and we are not in the business of projecting that."
— David Obstler, CFO

Assessment: The non-answer is consistent rather than evasive, and consistency has value here: the company has declined to project this account in every quarter since the risk emerged, including the quarters when projecting it would have helped. What is new is that the account is now running above a floor rather than below an expectation, which is a better place for the same silence.

Whether Enterprise Token-Cost Pressure Is Driving the Renewals

A question probed whether the early renewals were really budget distress among conventional enterprises grappling with unplanned AI spend. The answer separated the two cleanly.

Q: "you mentioned in your comments that a lot of customers are grappling with higher token costs, higher AI spend, and are trying to get it under control ... And to the extent that any of the customers that early renewed to try to get price discounts were not AI natives. They were traditional enterprises. Do you think that might be because they are trying to grapple with budget pressure?"
— Karl Keirstead, UBS

A: "in terms of the early renewal, that we see, they are not directly related to that. I think they come from large customers. That are either AI labs or that are focused on the. That are tech companies in general ... And that is that is does not really have anything to do with their own token consumption. It has more to do with their own growth as a business."
— Olivier Pomel, Co-Founder and CEO

Assessment: The right question to ask and a clear answer. The repricing is concentrated in AI labs and technology companies and is driven by their own growth, not by enterprise budget stress. Had the answer gone the other way, the Q4 guide would signal a broad-based pricing problem rather than a cohort-specific one, and our rating action would be different.

How Much of Infrastructure Monitoring's Growth Is AI Infrastructure

A question sought to attribute the infrastructure monitoring suite's mid-30s growth at over $2.0B of annual recurring revenue, and asked how sustainable it is as GPU efficiency improves. The answer removed GPUs from the growth attribution entirely.

Q: "guys disclosed it is now exceeding $2 billion in ARR and growing in the mid-30s. Could you quantify how different that growth rate is from when you last disclosed the scale of this business ... How much of that growth is coming from AI infrastructure and how sustainable is that growth rate as GPU efficiency improves?"
— Koji Ikeda, Bank of America

A: "Today, that growth is not actually related directly to GPUs ... GPUs are still largely unseen for our customers on training. We have products there, but we do not monetize those products widely yet. And so for us, that is an opportunity we have in front of us. that is not something that is baked into the current growth and the current numbers."
— Olivier Pomel, Co-Founder and CEO

Assessment: The most analytically useful answer of the call and the one that should change how the stock is valued. A product line above $2.0B growing in the mid-30s, with no meaningful GPU monetization in it, means the oldest part of the portfolio is compounding on conventional cloud adoption while the AI-infrastructure opportunity sits entirely unpriced. It also retires a conclusion this coverage published in May, which is noted candidly in the thesis scorecard below.

What They're NOT Saying

1. The size of the largest customer, or of its minimum commitment: Five quarters of questions, and this quarter the company used the exact number to build the Q4 guide and still would not share it. The CFO's "we are not answering" is the plainest refusal yet.
2. How much of the Q3 beat came from that customer billing above minimum: Asked directly. Refused directly. This is the one disclosure that would let investors judge whether the Q4 floor is 1% or 5% below the likely outcome.
3. The magnitude of the early-renewal repricing: A brand-new revenue headwind, disclosed qualitatively, affecting "a number of large customers" plus one more in October, with no dollar figure, no count and no duration.
4. The AI cohort's share of revenue: Withheld for a fourth consecutive quarter. Last given at 12% in Q3 2025. The cohort is now the main swing factor in the model and the one metric that would size it is gone.
5. Security ARR growth: Withheld for a fourth consecutive quarter, while management continues to describe security as a full industry rebuild and cites full-platform security lands at the defense agency and elsewhere.
6. The discount applied to the repriced AI renewals: Asked explicitly alongside the 2027 question. The CFO said the early-renewal customers are guided like the broader base, on recent organic growth with a discount applied rather than at a contractual minimum, but gave no size for that discount.
7. Any 2027 framework for the largest customer: Deferred to the year-end call. Given that the account drove both the August and November guides, investors will carry that uncertainty through the February print.
8. The AI customers above $10M of ARR: Disclosed at 8 last quarter and 5 the quarter before, not refreshed this quarter. A small withdrawal, and in the wrong direction for a cohort disclosure that otherwise improved.
9. Any revenue from Bits AI or the broader AI product suite: A packaging model was finally disclosed, which is progress, but adoption is still reported only in usage multiples. The attach claim made last quarter, that Bits users expand platform consumption, remains unquantified.
10. Full-year free cash flow: Still no guidance, a fifth consecutive quarter, on the metric that carries most of the valuation. Capital intensity was guided down, which makes the continued absence harder to explain.
11. Why gross margin fell 180bps over two quarters and then recovered 160bps in one: The "varies quarter to quarter" framework now has to explain a reversal as well as a decline, and it still has not been connected to the workload mix.

Market Reaction

• Pre-print setup: DDOG closed at $220.52 on November 3, up 62.2% year-to-date against the S&P 500's +10.6%, up 40.0% over trailing twelve months, and down 20.2% over the trailing 30 days. The 52-week closing range going in was $102.62 to $293.26, so the shares entered the print 24.8% below their 52-week closing high rather than near it.
• Reaction session (November 4, before the open): Opened at $199.74, a 9.4% downward gap, traded between $186.51 and $207.74, and closed at $188.72. That is a 14.4% single-session decline, or $31.80 per share.
• Volume: 13.2M shares against a 30-day average of 3.1M, or 4.3 times normal.
• Relative move: The S&P 500 rose 1.6% on the same session, making the relative move roughly 16 points.
• Peer reaction: The closest observability and data-infrastructure comparables fell between 2.5% and 3.0% on the session, a small fraction of the move in DDOG itself.

The setup is the most important difference between this print and the last one. In August the shares entered a print up 108% year-to-date and 1.7% below their 52-week closing high, and a single disclosure cost 19%. This time the stock arrived already de-rated: down 20.2% over 30 days from $276.42 on October 5, in a sequence driven by multiple compression across cloud software on AI-spending fears rather than by any change in Datadog's own estimates. The 14.4% decline is therefore a second leg down on an existing de-rating, and the two together have taken the shares 35.6% below their 52-week closing high.

The mechanics of the session are narrow and legible. A company that beat revenue consensus by 4.4%, beat EPS consensus by 26.0%, cleared its own operating income ceiling by 23.8%, raised full-year revenue, operating income and EPS guidance, and cut its capital intensity guide, lost a seventh of its market value. Every part of that decline traces to the shape of one number: a Q4 revenue midpoint 0.17% below the quarter just reported, where the four prior fourth quarters grew between 6.9% and 7.7% sequentially.

That the peer group fell only 2.5% to 3.0% while the S&P rose 1.6% confirms the move was company-specific rather than a sector repricing. It also argues against the simplest bear reading. If the market believed Datadog's guide revealed a broad slowdown in observability or AI-adjacent spending, the comparables would not have been nearly flat. The market read this as a Datadog disclosure event, which is also how management described it.

Volume at 4.3 times normal is heavier than the 3.5 times that accompanied August's 19% session, on a smaller price move. Taken with the gap-and-extend pattern inside the day, where the stock opened down 9.4%, fell to down 15.4%, and closed down 14.4%, the picture is of forced repositioning into a stock that had already lost its momentum ownership during October.

Street Perspective

Debate: Is the Q4 Guide a Forecast or a Floor?

Bull view: The company stated that the guide assumes its largest customer at a contractual minimum, and stated that the same customer billed substantially above that minimum last quarter. The CFO conceded on the record that spending above the minimum produces a beat "mathematically." Applying the company's own five-quarter pattern of clearing the ceiling by roughly 4% implies a Q4 print near $1.248B and 30.9% growth. The guide is a floor with a known direction of error.

Bear view: A floor is only conservative if the floor is far from reality, and nobody outside the company knows where it sits, because the account has never been sized through five quarters of asking. A company that genuinely expected to beat by 4% would not have guided the midpoint below the prior quarter, because it understands perfectly well how that reads. The minimum-commitment language may be prudent framing around a real step-down that the company declines to describe.

Our take: The bull has the better evidence and the bear has the better instinct about incentives. What tips it is the first Q&A answer: management was asked what changed in the assumptions versus 90 days ago and said one thing changed, that it is a price reset on growing accounts, and that everything else "played out pretty much to where we were expecting." That is a specific, falsifiable statement about demand, offered unprompted. We will know in February whether it was true, and the cost of being wrong at 14.1x is far lower than the cost of being wrong at 17.5x was in August.

Debate: Does the Early-Renewal Repricing Mark a Structural Price-Down in the AI Cohort?

Bull view: Customers do not renew early to pay less for the same thing; they renew early because they expect to outgrow their commitment and want better unit economics on the way up. Management confirmed the commitments went up, not down, and the evidence is in the numbers: remaining performance obligations jumped $380M after three flat quarters, in the same quarter billings decelerated. The dollars are committed; only the timing of recognition moved.

Bear view: This is the AI cohort discovering it has pricing power, and it will keep discovering it. A year ago the largest customer was repriced; last quarter it announced a usage reduction; this quarter the repricing broadened to "a number of large customers" with another in October. Each event individually is benign and each is unquantified, and the cumulative effect is a usage-based business whose largest and fastest-growing customers now reset their price cards whenever they scale. That is margin and growth leakage dressed as a volume win.

Our take: The bear's pattern argument deserves more respect than the bulls give it, because large-AI-account contract mechanics determined the guide again this quarter, as they did in August, and we have been told each time that it is good news. But the direction of the commitment is the fact that settles it: these renewals raised committed dollars and lengthened duration, which is visible in the RPO line and is not something a company can fabricate. The honest position is that the AI cohort is a lower-price, higher-volume, higher-volatility business than the core, and that this is now a permanent feature rather than a transition.

Debate: After a 36% Drawdown, Is the Multiple Finally Paying for the Growth?

Bull view: At $188.72 the shares trade at 14.1x enterprise value to the raised full-year guide, below the 14.6x where this coverage downgraded in May and well below the 17.5x of August, on a company whose full-year operating margin guide has risen from 21% to 24.3% across the same period. The stock is 35.6% below its 52-week closing high while estimates have gone up at every quarter of 2026. That combination is a de-rating of the multiple, not of the business.

Bear view: 14.1x current-year revenue and roughly 69x this year's earnings is not cheap on any absolute measure, and the growth rate the market must underwrite is no longer 35% but something closer to the mid-20s once the largest customer and the repriced renewals annualize. A business decelerating from 35% to 25% does not deserve the multiple it carried at 35%, and the stock can be down 36% and still expensive, as it was in August.

Our take: We made the bear's argument in August and it was correct then, so it is worth being precise about why it is weaker now. In August the multiple was higher than at both prior rating decisions while the forward raise was shrinking. Today the multiple is below the level at which we downgraded, the full-year guide has been raised again, the profitability guide has been raised for a third consecutive quarter, capital intensity has been cut, and two of our three open operating worries closed favourably. The growth deceleration the bear describes is real and it is in our estimates: we model roughly 24% for 2027. At 11.9x next-twelve-month revenue for a business compounding in the mid-20s with a 24% operating margin, the price now contains the deceleration.

Model & Valuation Framework

Model Update

Valuation

At the November 4 close of $188.72 and roughly 360M shares outstanding (Q3 weighted-average basic was 358.6M), market capitalization is approximately $67.9B. Netting $5.23B of cash and marketable securities and adding back the $1.0B of 2029 convertible notes gives an enterprise value near $63.7B. With one quarter of the fiscal year remaining, we anchor on next-twelve-month revenue of approximately $5.36B, built from a $1.235B Q4 estimate and roughly 24% growth in 2027.

The August, May and February columns are computed on the enterprise value at those closing prices against the denominators used in each of those notes. The direction of travel is the point: the stock is now cheaper on every measure than at any rating decision since February, while the full-year revenue guide and the full-year operating margin guide have both been raised at all three subsequent quarters.

Fair value framework. We apply 13x to 17x to our next-twelve-month revenue estimate of $5.36B, producing an enterprise value range of $69.7B to $91.1B, or $205 to $265 per share after adjusting for net cash and the convertible notes. Our central case of 15x gives $235.

The multiple range compresses one point from the 14x to 18x used in August, and the compression is deliberate. The near-term revenue path is demonstrably less predictable than we modelled it: the company has guided to what would be the first sequentially declining quarter in its history, and two of the three largest swing factors in the forecast are variables management will describe but not size. Against that, the quality of the business improved on almost every axis we track, which is why the compression is one point rather than three.

Valuation impact. Our 12-month fair value estimate is $235, essentially unchanged from the $237 set in August. The base rolling forward a quarter raises it and the one-point multiple compression offsets that almost exactly. At $188.72 the gap to fair value is roughly 25%, against the 3% that justified maintaining Hold in August and the 9% that produced the May downgrade.

Thesis Scorecard Post-Earnings

Scored against the standing thesis carried since initiation in November 2025.

Overall: Thesis strengthened. Three of four bull pillars confirmed, with one carrying a candid correction to our own prior reasoning. Two of the three operating worries we carried into this print, gross margin erosion and the flat commitment book, resolved favourably and in the same quarter. The risk side got worse in two places that matter, cash conversion and disclosure, and the concentration risk widened in scope from one account to a cohort behaviour. What changed most is the price: the valuation pillar that has been Materializing since May is now Contained, and it is the reason the rating moves rather than the business.

Action: Buy. The quarter measured a business compounding 35% with a 28% operating margin and improving unit economics. The guide is an assumption the company has described as a contractual floor, on an account that billed substantially above that floor three months ago. We are paid roughly 25% to accept that the floor is conservative, at a multiple below the one at which we downgraded this stock in May.