The Boring Half Woke Up: Non-AI Growth Accelerates to 20% and the Stock Reprices 23% in a Session

Results vs. Consensus

Q3 2025 Scorecard

Datadog is not broadly modelled below the revenue and non-GAAP EPS lines, so those are the only two rows where a reliable published consensus exists. The company's own guidance, issued thirteen weeks earlier, is the more informative benchmark for the rest of the print.

Versus Company Guidance (issued August 7, 2025)

The operating-income beat is more than three times the revenue beat in percentage terms. That is the signature of a company whose cost base was set against a materially lower revenue plan and which did not spend the upside as it arrived.

Year-Over-Year Comparisons

Quarter-Over-Quarter Comparisons

The sequential margin recovery flatters the quarter somewhat. Datadog's DASH user conference ran in June, which loads a concentrated block of sales and marketing spend into Q2 and makes Q2 the structural low point for operating margin. The cleaner comparison is the year-over-year one, where non-GAAP operating margin is still 166bps lower because non-GAAP operating expenses grew 32% against 28% revenue growth.

Quality of Beat

Platform & Cohort Performance

Datadog reports a single revenue line and does not disclose product-level or geographic segments, so the analytically useful decomposition is the one management supplies verbally: the split between AI-native customers and everyone else, and the product-density metrics that measure how deeply the platform is embedded.

Revenue Cohort Mix

The non-AI-native cohort

This is the cohort that determines whether Datadog is a durable compounder or an AI-cycle derivative. It grew 20% year-over-year, up from 18% in Q2, and management said sequential usage growth from existing non-AI customers was the highest in twelve quarters. Twelve quarters reaches back to late 2022, which covers the entire cloud-optimization slowdown. Management further volunteered that the acceleration continued into October, which is the only forward-looking datapoint in the release that is not a guidance range.

"As a notable inflection, we saw acceleration of year-over-year revenue growth across our non AI customers. And the sequential usage growth for non-AI existing customers was the highest we have seen going back 12 quarters." — Olivier Pomel, Co-Founder and CEO

Assessment: A two-point acceleration in a cohort this large is worth more to the equity than a five-point acceleration in the AI cohort would be, because it is the part of the base with fifteen years of retention history behind it. The honest caveat is that one quarter of sequential usage strength inside a consumption model is not yet a trend, and management explicitly refused to underwrite duration when asked.

The AI-native cohort

The group now spans more than 500 companies, about 100 of which spend over $100,000 annually and 15 of which spend over $1M. At 12% of revenue it has doubled its share in four quarters. Management's framing is that this cohort is a leading indicator rather than a category, and that the metric will lose relevance as AI workloads spread into the general customer base.

"In Q3, this group represented 12% of our revenue, up from 11% last quarter and about 6% in the year ago quarter. I will note that over time, we think this metric will become less relevant as AI usage in production broadens beyond this group of customers." — David Obstler, CFO

Assessment: The broadening is real and reduces single-name risk at the margin, but the disclosure is deliberately structured to obscure the thing investors most want to know. Management reports AI-native growth "excluding our largest customer," which is an admission that including it would tell a different story. Until the concentration is sized, this cohort should be valued at a discount to the rest of the revenue base rather than a premium.

Platform Adoption Density

Assessment: The density metrics are the strongest part of the disclosure and the least discussed. The 2+ band is saturated and moves a point a year, but the 4+, 6+ and 8+ bands each added four to five points. Product attach is where a usage-based model converts into switching cost, and the shape of this table is what justifies a premium multiple over a single-product monitoring vendor.

Key Operating Metrics

Assessment: Every bookings-based indicator on this table is running ahead of the 28% revenue growth rate: billings at 30%, RPO at 53%, current RPO in the low 50s, security ARR in the mid-50s. Net revenue retention is the one that is not. That gap is the trailing-twelve-month mechanic doing what it always does, and it means the reported 120% understates current-quarter expansion. The corollary is uncomfortable and worth holding onto: the same mechanic will overstate expansion on the way down.

Key Topics & Management Commentary

Overall Management Tone: Management was the most unguarded it has sounded in several quarters, treating a two-point acceleration in the non-AI-native cohort as a return to the company's historical normal rather than as a surprise, and going as far as apologising for over-delivering. Analyst pushback was narrow and clustered on two things the company would not quantify: the economics of the renewed largest AI-native contract, and whether the new agent products will be priced as a line item or absorbed into platform consumption. On both, the response was framework-level rather than numeric, and those are the only two places the call was less than convincing.

1. The Non-AI-Native Inflection

The quarter's defining disclosure is that the 88% of revenue that has nothing to do with AI-native startups accelerated. Growth in that cohort went from 18% to 20% year-over-year, and management characterised sequential usage growth from existing non-AI customers as the strongest in twelve quarters. Pressed repeatedly on why now, management gave a three-part answer: a demand environment that is no longer actively deteriorating, two years of accumulated sales-capacity investment beginning to convert, and a set of newer products reaching commercial maturity at the same time.

"So it is broad-based. And I think, again, speaks to a couple of things. It speaks to the fact that, in general, the demand environment is good. Though I would say, there's been a very, very high growth of hyperscaler revenue like over the past an acceleration for the hyper scalers in general. A lot of that is GPU related, but the growth we're seeing here and the exception we're seeing here is largely not GPU-related." — Olivier Pomel, Co-Founder and CEO

The CFO reinforced that the strength was not concentrated in any one customer size band, which is the pattern that would suggest a single large deal rather than a demand shift.

"Yes, I want to -- I'll add that it's across the customer base, enterprise SMB. And when we look at it, it's not just an AI SMB. If you remove the AI companies, you still see a strengthening SMB demand cycle going on. And unlike in previous periods, it also is across spending ranges." — David Obstler, CFO

Assessment: The breadth claim is the part that matters, because a broad-based usage acceleration across spending bands is very difficult to fabricate and very difficult to reverse in a single quarter. This is the strongest evidence in the print that the 2023 to 2024 cloud-optimization drag has fully cleared.

2. The Largest AI-Native Contract Was Renewed at a Lower Price

Management disclosed a nine-figure annualized expansion with an unnamed leading AI company, its largest AI-native customer, framed as securing "better economics for a higher commitment with an early renewal." An analyst immediately identified the tension: a higher commitment normally carries a volume discount, so "better economics" for whom? The answer confirmed the discount runs to the customer.

"Finally, we signed a 9-figure annualized expansion with a leading AI company. This company has been a long-time Datadog customer and has expanded their usage of multiple products, securing better economics for a higher commitment with an early renewal." — Olivier Pomel, Co-Founder and CEO

Asked directly what "better economics" meant, management described the mechanic without hedging: the customer commits to more volume, receives a lower unit rate, and therefore pays less per unit of usage than before.

Assessment: Two things are true at once and the market is currently pricing only the first. The contract extension removes a large binary risk that had been hanging over the name, which is unambiguously good. But a repriced renewal on a customer large enough to warrant its own "excluding our largest customer" disclosure line is a mathematical headwind to reported revenue for the next several quarters unless that customer's usage grows fast enough to outrun its own price cut. Management never sized either side of that equation, and this is the single largest unquantified item in the print.

3. The AI-Native Cohort Is Broadening

Beyond the largest customer, the cohort is developing a middle. There are now more than 500 AI-native companies in the group, roughly 100 spending above $100,000 annually and 15 spending above $1M. Management described the composition as spanning model vendors, video and audio generation, coding assistants, infrastructure layers and pure AI applications, and pushed back on the idea that these are all recently founded companies.

"It's actually fairly broad. So there is model vendors, there's models -- model that can be the lens model that can be video, it can be sound generation, it can be all of the various parts of the stack you see as independent companies." — Olivier Pomel, Co-Founder and CEO

Assessment: Fifteen customers above $1M is a meaningful second tier and materially better than a cohort that is one whale plus noise. The concentration risk is falling in shape even as the cohort's revenue share rises. Investors should still assume the top customer is a large share of the 12% until told otherwise.

4. Security Is the Second Growth Engine Nobody Is Underwriting

Security ARR growth accelerated into the mid-50s percent year-over-year from the mid-40s the prior quarter, with acceleration in every individual security product. The specific unlock management identified is Cloud SIEM entering large deals, which it attributes to product maturity, the cost efficiency of running SIEM workloads on top of Flex Logs, and a channel investment made roughly six months earlier.

"Security ARR growth was in the mid-50s as a percentage year-over-year in Q3, up from the mid-40s we mentioned last quarter. We're starting to see success in including Cloud SIEM in larger deals, and we'll get back to that in a bit in our customer examples." — Olivier Pomel, Co-Founder and CEO

Assessment: Security growing at roughly double the corporate rate, off a base large enough to appear in seven-figure deal narratives, is the most under-appreciated line in the story. It is also the cleanest strategic answer to the bear case that observability is a maturing category, because it converts the same telemetry pipeline into a second budget line at a customer.

5. New-Logo Bookings Doubled, and the Land Motion Is Finally Contributing

New-logo annualized bookings more than doubled year-over-year and set a company record, driven by larger average land sizes in enterprise. The share of year-over-year revenue growth attributable to new customers rose to about 25% from 20% in Q2.

"New logo annualized bookings more than doubled year-over-year and set a new record driven by an increase in average new logo land size, particularly in enterprise. We believe we are starting to see the benefits of our growth of sales capacity." — David Obstler, CFO

Management also described restructuring sales compensation specifically to counteract the gravitational pull of the land-and-expand model, including carving out multi-year plans for large targets whose sales cycles exceed a single comp year.

Assessment: For a company whose growth has been overwhelmingly expansion-driven, a step-change in new-logo contribution genuinely improves the durability of the model, because it widens the base that future expansion compounds on. The comp-plan detail is the sort of unglamorous operating disclosure that tends to be predictive.

6. Bits AI: Strong Product Signal, No Monetization Answer

The agent products announced at the June user conference are in preview with thousands of customers onboarded. Management's evidence is behavioural rather than financial: customers run a two-hour, twenty-person incident bridge, then discover the agent reached the same conclusion three minutes in. A security agent in preview is credited with helping win large Cloud SIEM deals by clearing roughly 90% of alert triage automatically.

"We also started having some customers buy and come to it to just to show value and to make sure we're on to the right product mix. And so we feel good that this is something that is high quality and we can monetize." — Olivier Pomel, Co-Founder and CEO

Assessment: The product signal is credible and the competitive differentiation claim is unusually strong for a company that normally understates. The monetization answer is not there yet, and management said plainly it has not decided whether to charge for the agents directly or let them drive platform consumption. In a usage-based model that ambiguity is less damaging than it would be in a seat-based one, but it does mean Bits AI cannot yet be underwritten as a 2026 revenue line.

7. Platform Density and the 1,000-Integration Milestone

Datadog passed 1,000 integrations this quarter. The commercially relevant statistic is not the total but the usage: the average customer runs more than 50 integrations, and customers spending over $1M run more than 150. Digital experience products, spanning real user monitoring, synthetics and product analytics, were disclosed for the first time as exceeding $300M in ARR, with product analytics alone adopted by more than 1,000 customers.

"Our 32,000 customers use more than 50 integrations on average, while customers spending over $1 million annually with us use more than 150." — Olivier Pomel, Co-Founder and CEO

Assessment: An integration count is a vanity metric; 150 live integrations inside a single large account is a rip-out cost. This is the mechanical explanation for gross revenue retention sitting in the mid-to-high 90s through an entire optimization cycle, and it is the foundation of the bull case on durability.

8. Gross Margin: 81.2% and Management Still Says "Plus or Minus 80%"

Non-GAAP gross margin came in at 81.2%, up from 80.9% sequentially and roughly flat against 81.1% a year ago. Management attributes the resilience to a multi-year cloud-efficiency engineering program, and offset it against the volume discounts larger customers receive. Asked how to think about the trajectory, the CFO declined to move the long-standing framing.

"I'll repeat what we've always said that we've been running the company with a gross margin plus or minus 80%, we've given that range and not changed it, and we watch it." — David Obstler, CFO

Assessment: Read this as guidance not to model gross-margin expansion. Datadog's engineering efficiency work is funding the discounts it gives large customers rather than dropping to the margin line, which is a reasonable trade but means the entire operating-leverage case has to come from opex ratios. That matters when opex is currently growing faster than revenue.

9. RPO Grew 53%, and Management Told You Not to Care

Remaining performance obligations reached $2.79B, up 53% year-over-year, with current RPO growth in the low 50s. Billings grew 30%. Both are far ahead of the 28% revenue growth rate. Management attributed the acceleration to bookings strength generally rather than to the large AI renewal, and repeated its standing position that these metrics are noisy.

"We continue to believe that revenue is a better indication of our trends in our business than billings and RPO." — David Obstler, CFO

Assessment: The caveat is intellectually honest and has been consistent for years, which earns it credibility. It is also convenient in a quarter where RPO grew nearly twice as fast as revenue. The right reading is that a 53% RPO print does not translate into 53% revenue growth, but a company genuinely worried about forward demand does not typically post one.

10. Guidance Philosophy: The Conservatism Is Structural

The Q4 guide of $912M to $916M implies roughly 24% year-over-year growth and about 3.2% sequential growth off the Q3 base. For reference, the Q3 guide implied 2.7% sequential growth and the quarter delivered 7.1%. Management explicitly described its method as basing guidance on recent observed trends and then applying conservatism on top.

"First, our guidance velocity overall remains unchanged. As a reminder, we based our guidance on trends observed in recent months and imply conservatism on these growth trends." — David Obstler, CFO

On the specific question of Q4 seasonality and the calendar placement of holidays, the CFO said the day-by-day usage pattern is modelled explicitly and incorporated as in prior years.

Assessment: The guide should be read as a floor, not a forecast. But there is a trap in mechanically applying the Q3 beat to Q4: the Q3 guide was set before management could see the non-AI acceleration, whereas the Q4 guide is set with it in hand and with October trends already known to be strong. The conservatism buffer is probably thinner this time, not thicker.

11. The GAAP Business Still Loses Money

Datadog reported a GAAP operating loss of $5.8M against operating income of $20.3M a year ago. The delta is stock-based compensation, which reached $200.6M, or 22.6% of revenue, growing 41% year-over-year against 28% revenue growth. Total GAAP operating expenses grew 34.4%. Management framed the opex trajectory as deliberate and decelerating.

"Our Q3 OpEx grew 30% -- 32% excuse me, year-over-year, down from 36% last quarter. We continue to grow our investments to pursue our long-term growth opportunities, and this OpEx growth is an indication of our execution on our hiring plan." — David Obstler, CFO

Assessment: The deceleration in opex growth is real and welcome. The compensation structure is the harder issue: at 22.6% of revenue and growing 13 points faster than revenue, SBC funds roughly 80% of reported operating cash flow. Dilution has been modest so far at 1.2% year-over-year on diluted shares, which is genuinely good discipline, but that is a function of a rising share price absorbing the grants. Nobody asked about the GAAP path on this call, and nobody volunteered one.

Guidance & Outlook

The FY25 revenue raise of $71M at the midpoint decomposes into roughly $35M of realised Q3 beat and roughly $36M of genuine forward raise. That split matters: management did not simply pass through the quarter and hold the rest flat, which is the more conservative pattern this company has often followed. Roughly half the raise is a real upgrade to the Q4 outlook.

The operating-income raise is more aggressive still. FY25 non-GAAP operating income went up $67M at the midpoint against a $71M revenue raise, an incremental flow-through of about 94%. Management is guiding almost the entire revenue upside to fall to the operating line in Q4, which is a meaningful signal about hiring pacing into year-end.

Implied Q-over-Q ramp: The Q4 midpoint of $914M requires 3.2% sequential growth off Q3's $885.7M. Q3 itself delivered 7.1% sequential against a guide that implied 2.7%. Applying the same gap mechanically would put Q4 above $945M, but the Q4 guide was set with October's strength already visible and the holiday usage lull is a genuine seasonal drag that management models explicitly.

Street at: Consensus sat near $887M for Q4 going into the print, so the guide midpoint lands roughly 3% above where the Street was. The FY25 EPS guide of $2.00 to $2.02 is a 10.7% raise from the prior $1.80 to $1.83 range.

Guidance style: Structurally conservative and consistently so. The company entered 2025 guiding to roughly 19% revenue growth and is exiting it guiding to 26%. Every guide this year has been a floor. The forward risk is that the Street has now fully internalised that pattern and models the beat rather than the guide, which removes the cushion that made prior quarters easy.

Other guidance notes: FY25 net interest and other income of approximately $170M, cash taxes of $10M to $20M, a 21% non-GAAP tax rate applied for 2025 and going forward, and capital expenditures plus capitalized software together at 4% of revenue.

Analyst Q&A Highlights

Whether GPU Monitoring Is Contributing to the Acceleration

The call opened with the question that mattered most for interpreting the beat: is any of this growth coming from monetizing GPU infrastructure, the workload category that has driven the hyperscalers' own acceleration? Management's answer was an unusually clean denial, and it reframed the entire quarter.

Q: "We've talked about GPU monetization versus CPU monetization. So how closer are we to the point where you can confidently expand and get your share of the customer wallet when it comes to whether it's training workload, inferencing workload on the GPU clusters, which are becoming more prevalent and increasingly a larger part of the compute build-out in the future?"
— Kasthuri Rangan, Goldman Sachs

A: "Yes. So we have products that are getting into the market now for GPU monitoring. But these don't generate any significant revenue yet. So all the revenues we've shared, like the acceleration, et cetera, that's not related to us capitalizing more on GPUs, that's a future opportunity."
— Olivier Pomel, Co-Founder and CEO

Assessment: This is the most valuable exchange on the call and it cuts both ways. It removes the concern that the beat is a leveraged bet on the AI capex cycle, since none of the acceleration comes from GPU monitoring. It simultaneously establishes a free option that is not in any model. A management team inclined to promote would have described GPU monitoring as "early but ramping"; describing it as generating no significant revenue is the more credible posture.

What Is Actually Driving the Enterprise Re-Acceleration

The dominant line of questioning on the call was why the non-AI cohort turned now, after eight quarters of deceleration and stabilization. Management gave a three-part decomposition rather than a single cause, and notably declined to claim that cloud migration itself had accelerated.

Q: "Olivier, I wanted to talk about some of those enterprise trends you're seeing in sort of your non-AI cohort. What do you sort of put the improved performance in growth this quarter on? You mentioned that the sales productivity or the benefit from some of the sales investments starting to come online. Is there sort of an uplift in sort of the cloud migration trends as you're starting to see enterprise build more AI applications."
— Sanjit Singh, Morgan Stanley

A: "Yes. I'd say there's 3 parts to it. One part is the demand environment is not -- is positive in general. I don't know that we see massive acceleration of cloud migration, but at least the environment is not pushing the other way. We know which happens from time to time. So that's point number one. Point number two is we've been growing sales capacity quite a bit, and we've created new go-to-market motions to go after the kind of customers who were not getting before."
— Olivier Pomel, Co-Founder and CEO

Assessment: The refusal to claim a cloud-migration tailwind is the tell that this answer is honest. Two of the three drivers named are internal and controllable, which is the better outcome for durability: sales capacity and product maturity persist, whereas a macro upswing does not. It also means the acceleration is repeatable only as long as capacity keeps scaling.

What "Better Economics" Means on the Largest AI Contract

The nine-figure renewal drew the sharpest pushback of the call. The question identified the logical problem directly: a larger commitment normally means a volume discount, so describing the outcome as better economics required explanation. The answer confirmed the discount, and confirmed the near-term revenue consequence.

Q: "You had mentioned the expansion of the contract with your largest AI native customer and I believe you said better economics for a higher commitment. Can you speak to that because I would assume a higher commitment would carry a volume-based discount. I'm just trying to understand if. For some reason, if that was not the case here, what did you mean by better economics?"
— Mark Murphy, JPMorgan

A: "Well, the bigger economics means you commit tomorrow, you get a better price. And as we -- remember, we have a usage model. So we charge people every month on what they use at the price we agreed. So if you get better economics, and your usage is somewhat similar month to month, one month and the next that you pay less, but the overall backdrop of our business is increased consumption."
— Olivier Pomel, Co-Founder and CEO

Assessment: Management answered the question honestly and in doing so disclosed a forward revenue headwind on its single largest AI customer. The final clause is the hedge that the whole thesis on this cohort rests on: consumption growth is expected to outrun the price cut. That is an assumption, not a disclosure, and it is unverifiable from outside because the company will not size the customer.

Why Net Revenue Retention Did Not Move

With bookings, RPO and usage all accelerating, the flat 120% net revenue retention figure was the obvious inconsistency. The question anticipated its own answer, and management confirmed the mechanic rather than reaching for a better explanation.

Q: "David, on net retention rates, why aren't we necessarily seeing more upward pressure on the metric, just given the strength of expansionary bookings that you alluded to in the quarter from the installed base. And I mean I suspect it's because it's a trailing 12-month metric. But any directional color you can just share on that."
— Fatima Boolani, Citi

A: "Yes, you've nailed it. It's a trailing 12 months, it's a number that's rounded, it has the dynamics that you might expect in that the growth of the non-AI natives has been, as we mentioned, a combination of landing and expanding at higher rates than we've seen in recent quarters. So if that continues as you go into a trailing 12-month metric, you see a directional movement."
— David Obstler, CFO

Assessment: A straightforward confirmation that reported net revenue retention lags the business by up to a year. The practical implication for modelling is that this metric should be treated as a confirming indicator with three or four quarters of delay, never as a leading one, in either direction.

Whether the Agent Products Become a Revenue Line in 2026

The second unresolved item on the call was how the new agent portfolio gets priced. Management was candid that the decision has not been made, and framed the question as one about AI packaging generally rather than about one product.

Q: "Bits AI seems like a really exciting thing out of Dash. And I know it's still in preview, but you mentioned there's a lot of interest there. So I'm just curious how you think about the agentic opportunity with Bits AI. How meaningful this can be for 2026 as a differentiator versus competition and also as a revenue contributor?"
— Eric Heath, KeyBanc

A: "In terms of the impact for next year, on the packaging side, I'm not completely sure yet whether the biggest impact will be seen from what we charge Bits AI itself or for the rest of the platform, that it gets benefits from the differentiation of Bits AI. I think that's more of a broader question of packaging and monetization of AI."
— Olivier Pomel, Co-Founder and CEO

Assessment: In a seat-based model this answer would be a problem. In a consumption model it is closer to a non-issue, because an agent that increases telemetry ingestion monetizes whether or not it carries a price tag. The genuine risk is the opposite one, largely unexamined on the call: agents that resolve incidents faster could reduce the volume of data customers feel compelled to retain.

Contract Structure and the RPO Contribution of the AI Renewal

A precise attempt to bound the AI renewal by working backwards through the reported RPO acceleration got no numeric answer, though management did decouple the RPO figure from the contract.

Q: "Are you able to clarify whether that was a 1-year deal or multiyear? And then related to that, David, what is the contribution to CRPO from that deal, which I presume landed in your CRPO number. If it is a 1-year deal does the entirety of that contract contribute to the sequential CRPO performance in the quarter?"
— Karl Keirstead, UBS

A: "I think in this case, we've mentioned this current and the total is roughly the same, and there wouldn't be anything in that contract that would have been materially around of those numbers. Those numbers, I think we mentioned are produced from the bookings growth more generally and not from that particular contract."
— David Obstler, CFO

Assessment: The refusal to disclose the term is expected and consistent with policy. The substantive concession is that the 53% RPO growth is not an artifact of one large renewal, which materially strengthens the forward-demand read. Investors got the more useful of the two answers even though they did not get the one they asked for.

How Long the Non-AI Acceleration Can Persist

The final substantive question sought a duration estimate on the quarter's central datapoint. Management declined to give one, and the refusal was framed around the structural unpredictability of a consumption model rather than around any specific concern.

Q: "Just the duration of this acceleration of the non-AI cohort. It seems like from all your forward-looking metrics, whether it's billings, RPO, CRPO. Those were, again, really, really strong how long do you think we should think about the duration of this trend of this non-AI acceleration?"
— Aleksandr Zukin, Wolfe Research

A: "Well, we're a consumption business. So we -- the hardest thing to understand is what the future is going to look like for consumption. The way I would say it is we feel very good about it at the midterm, long term. Now with ebbs and flow in any given month or quarter, that's harder to tell."
— Olivier Pomel, Co-Founder and CEO

Assessment: The honest answer, and the one that should temper how much a single quarter's sequential usage figure is extrapolated. A consumption model with mid-to-high-90s gross retention has a very stable floor and a very unstable slope. Management is confident about the floor and explicitly refuses to underwrite the slope, which is exactly the distinction the market collapsed when it repriced the stock 23% in one session.

What They're NOT Saying

1. The size of the largest AI-native customer: The company reports AI-native growth "excluding our largest customer" and reports the cohort at 12% of revenue, but never bridges the two. Four separate questions probed the contract's size, term and RPO contribution; none produced a number. The construction of the disclosure implies the customer is large enough to distort the cohort's growth rate on its own.
2. The revenue impact of the repriced renewal: Management confirmed the customer now pays a lower unit rate but gave no indication of magnitude or timing. A price reduction on a nine-figure annualized contract is a material forward item that is currently unmodellable from the outside.
3. Any framework for 2026: Normal for a Q3 call, but with the Street now anchoring on an exit rate rather than a guide, the absence of even a directional 2026 comment leaves an unusually wide range of reasonable models.
4. Bits AI pricing: Explicitly undecided by management's own admission. The product is described as materially better than anything else in the market, and the company has no plan yet for capturing that.
5. GPU monitoring revenue: Disclosed only as "not significant." No customer count, no ARR figure, no ramp timeline for a product category management positions as a major future opportunity.
6. The path to GAAP profitability: Not raised by management and not asked by any analyst, despite a GAAP operating loss and stock-based compensation at 22.6% of revenue growing 41% year-over-year. This is a topic the company has been allowed to leave alone for several years.
7. Why free cash flow margin fell 535bps: Free cash flow grew only 5% against 28% revenue growth, and the call contained no discussion of the working-capital reversal or the 50% increase in capex plus capitalized software that caused it.
8. Net revenue retention beyond one rounded figure: The company gives a single rounded trailing-twelve-month number with no cohort split, which makes it impossible to see whether the AI-native and non-AI-native cohorts are expanding at similar rates.

Market Reaction

• Pre-print setup: DDOG closed at $154.98 on November 5. The stock was up 8.5% year-to-date against the S&P 500's 15.6%, up 20.7% over trailing twelve months, and up just 0.3% over the trailing 30 days. The 52-week closing range going in was $87.00 to $168.65, so the shares entered the print roughly 8% below their own 52-week closing high and materially behind the index for the year.
• Reaction session (November 6, before the open): Opened at $178.89, a 15.4% gap, traded between $177.57 and $194.87, and closed at $190.82. That is a 23.1% single-session gain, or $35.84 per share, and a close well above the prior 52-week closing high of $168.65.
• Volume: 21.3M shares against a 30-day average of 3.7M, or 5.7 times normal. The session low of $177.57 sat just 0.7% below the open, so the gap held all day.
• Relative move: The S&P 500 fell 1.1% on the same session, making the relative move roughly 24 points.

The size and shape of the move are both informative. A 15.4% opening gap that then extends another 7.7 points into the close is not a short squeeze pattern; it is real institutional repositioning through the session, and the 5.7x volume confirms it. What repriced was not the beat, which was large but broadly in character for a company that has beaten its guide all year. What repriced was the removal of two overhangs at once.

The first overhang was the fear that Datadog had become an AI-native derivative with a decaying core. The 20% ex-AI growth rate, up from 18%, retired that argument in a single line of the prepared remarks. The second was contract risk on the largest AI-native customer, which the nine-figure early renewal removed. A market that had been discounting the stock for both simultaneously had to unwind both discounts at once, which is how a good quarter becomes a 23% day.

The pre-print setup explains the magnitude. The shares had gone essentially nowhere for a month and trailed the index by seven points year-to-date, so positioning going into the print was light and expectations were low. That asymmetry is now gone. The stock closed the session 13% above its prior 52-week closing high, and the next print will be judged against a very different bar.

Street Perspective

Debate: Is the Non-AI Acceleration Durable or a Single-Quarter Artifact?

Bull view: Three independent forward indicators corroborate the usage data. New-logo bookings doubled, RPO grew 53% and current RPO grew in the low 50s, and the acceleration was described as broad-based across enterprise, SMB and every spending band. Management confirmed the trend continued into October. Coordinated fabrication across that many metrics is not plausible, and the drivers named are internal and repeatable.

Bear view: This is one quarter of sequential usage inside a consumption model, measured against a twelve-quarter comparison window chosen by the company. Management refused to underwrite duration when asked directly. The most likely explanation is a normal cyclical recovery in cloud spend that will show up at every infrastructure vendor, in which case Datadog captured a beta upswing and the market paid an alpha multiple for it.

Our take: The bull side has the better of this, but not by as much as a 23% move implies. The breadth evidence across spending bands is genuinely hard to explain as noise, and the fact that management attributed it to sales capacity and product maturity rather than to a macro upswing is the more credible framing. The correct posture is to treat 20% as a real new floor for the core while declining to extrapolate further acceleration, which is roughly what management itself said.

Debate: Should the AI-Native Cohort Carry a Premium or a Discount?

Bull view: The cohort doubled its revenue share in four quarters and is broadening fast, with 15 customers now above $1M and roughly 100 above $100,000. These are the fastest-growing companies in the economy and Datadog is embedded in their production stacks early. The cohort is a call option on AI application deployment that costs nothing to hold.

Bear view: Twelve percent of revenue sits in a customer group that includes at least one account large enough to require its own exclusion line, whose contract was just repriced downward, and whose spending is funded by venture and strategic capital rather than operating cash flow. This is the lowest-quality revenue in the company and it is growing fastest, which flatters the blended growth rate while degrading its durability.

Our take: The bear case is underweighted in the current price. The broadening is real and reduces concentration in shape, but the repricing of the largest contract is a disclosed, unquantified, forward revenue headwind on the single biggest account. Until the company sizes it, this cohort deserves a discount to the core multiple, not a premium. The stock currently gets credit for both cohorts at the same rate.

Debate: Does Valuation Still Work After a 23% Session?

Bull view: Growth accelerated, the FY guide moved up 10.7% on EPS, incremental flow-through on the raise was roughly 94%, and gross retention is in the mid-to-high 90s. A durable 25%-plus grower with 81% gross margins and this much product attach has historically cleared 20x forward revenue, and the multiple compresses fast if 2026 growth lands anywhere near the current exit rate.

Bear view: The shares closed at roughly 19x enterprise value to FY25 revenue and about 75x our FY25 free cash flow estimate, on a business that lost money at the GAAP operating line and where 16% of non-GAAP EPS is interest income on the cash pile. Free cash flow margin fell 535bps year-over-year. The entire re-rating happened in one session, which means the inflection is fully priced before the second confirming quarter exists.

Our take: The bear case on price is stronger than the bear case on business quality, and those are different arguments. This is an excellent business that just got 23% more expensive in six and a half hours. The multiple now requires the non-AI acceleration to persist and the AI cohort to outgrow its own price cut. Both may happen; neither is yet demonstrated. That is the definition of a Hold rather than a criticism of the company.

Model & Valuation Framework

Model Update

Valuation

At the November 6 close of $190.82 and roughly 350M shares outstanding, market capitalization is approximately $66.8B. Netting $4.14B of cash and marketable securities and adding back the $1.0B of 2029 convertible notes gives an enterprise value near $63.7B.

Fair value framework: Applying a 14x to 17x multiple to our FY26 revenue estimate of $4.24B produces an enterprise value range of $59.3B to $72.0B, or $178 to $215 per share after adjusting for net cash and the converts. The midpoint of roughly $197 implies about 3% upside from the $190.82 close.

Valuation impact: Our 12-month fair value estimate is $197, a range of $178 to $215. The shares closed the reaction session inside that band and close to its midpoint. The 23.1% move consumed essentially the entire re-rating that the improved fundamental trajectory justifies, which is why the operational upgrade to our estimates does not produce a positive rating.

Thesis Scorecard Post-Earnings

This is our initiation on Datadog, so the scorecard below establishes the pillars we will grade against in subsequent quarters rather than scoring a standing thesis.

Overall: The operating thesis is stronger than we expected coming into this print, and three of four bull pillars were confirmed on their first grading. The offsetting development is that the equity repriced the entire improvement inside a single session, and two of three bear points concern items management chose not to address at all.

Action: Hold. Establish the position framework and wait for the second confirming quarter on ex-AI growth, or for a valuation reset, before paying up.

Bottom Line: Initiating at Hold

Datadog delivered the best quarter in its recent history and the market paid for all of it before lunch. Revenue of $885.7M beat consensus by 3.9% and cleared the company's own guidance ceiling by $34.7M, non-GAAP operating income beat the guide high end by 15.2%, and the full-year EPS guide went up 10.7%. Those are the surface facts, and they are not the reason the stock moved 23.1%.

The reason is that the 88% of revenue nobody was excited about accelerated. Growth excluding AI-native customers went from 18% to 20%, sequential usage in that cohort was the strongest in twelve quarters, and the acceleration was corroborated by record new-logo bookings, 53% RPO growth and a rising contribution from new customers. Management attributed it to controllable internal factors rather than to a macro upswing, which is the more durable explanation and the more credible one. Alongside that, the early renewal of the largest AI-native contract removed a binary risk that had been suppressing the multiple all year.

Three things keep this at Hold rather than Outperform. The renewal that removed the contract risk did so by cutting the customer's unit price, an admitted forward revenue headwind on the single largest AI account that management declined four opportunities to size. Free cash flow margin fell 535bps year-over-year while stock-based compensation reached 22.6% of revenue and funded roughly 80% of operating cash flow, and neither development was raised on the call by anyone. And the shares now trade at 18.8x enterprise value to this year's revenue and roughly 75x our free cash flow estimate, having closed 13% above their prior 52-week closing high.

We are constructive on this business and expect to own it at some point. The bull pillars we are establishing here were confirmed on their first grading, which is not common. But the entire re-rating that a durable 25% grower deserves was delivered in one session, ahead of the second quarter of evidence that the inflection is a trend rather than a print. Our fair value estimate of $197 sits about 3% above the close. We initiate at Hold and will upgrade on a second confirming quarter or a better price, whichever arrives first.