The Customer We Flagged in November Finally Showed Up: Datadog's Best Quarter Ever Loses 19% to One Account

Results vs. Consensus

Q2 2026 Scorecard

Versus Company Guidance (issued May 7, 2026)

This is the fourth consecutive quarter of beating the guide ceiling by between 3.8% and 4.7%, and the third of clearing it on operating income by roughly double the revenue margin. Nothing about the reported quarter broke the pattern.

Year-Over-Year Comparisons

The year-over-year comparison flatters slightly because Q2 2025 is the weakest comparative available: it carried the DASH user conference at a 19.8% operating margin and a 20.0% free cash flow margin. The half-year view below is the fairer read of the trend.

First-Half Comparisons

Quarter-Over-Quarter Comparisons

The Q2 operating margin expanded 71bps sequentially while absorbing roughly $15M of DASH conference cost that management had flagged in the prior guide. Stripping that out, underlying sequential margin expansion was closer to 200bps.

Quality of Beat

Platform & Cohort Performance

Revenue Cohort Mix

The core cohort

Five consecutive quarters of acceleration: 18%, 20%, 23%, mid-20s, high 20s. This is the pillar we established at initiation and it has been confirmed at every single grading. Management characterised the driver as existing customers rather than new logos, and argued the pace is sustainable precisely because it is well below the 2021 peak.

"revenue growth for our non-AI customers also accelerated again this quarter to the high 20s percent year-over-year, up from the mid-20s last quarter and 18% in the year ago quarter." — Olivier Pomel, Co-Founder and CEO

Assessment: The sustainability argument is the more interesting part. Management's point is that a customer base absorbing high-20s growth has room to keep absorbing it, whereas the 2021 cohort had to digest much larger increases across consecutive years and then optimized them away. That is a specific, falsifiable claim about why this cycle should not end the way the last one did.

The AI cohort, and the account inside it

The cohort reached 750 customers with 31 above $1M and 8 above $10M, up from 22 and 5 in Q1, and now includes hyperscalers running in-house AI labs. Within it sits the account that defined this print.

"we signed a 9-figure renewal with a leading AI company. This longtime, very large customer uses 17 Datadog products to enable unified visibility on production workloads at a very large scale, albeit with a user reduction starting in Q3, which we considered in our guidance" — Olivier Pomel, Co-Founder and CEO

Assessment: The cohort around this account keeps strengthening, and eight customers above $10M is a genuinely diversified top tier compared with the single-whale structure of a year ago. But the revenue-share percentage has now been withheld for three consecutive quarters, which means the market cannot size either the account or the reduction. The company disclosed the fact and withheld the magnitude, which is precisely the combination that produces a 19% day.

Platform Adoption Density

Assessment: The 10-or-more band nearly doubled from 7% to 13%, the steepest relative gain in any density band we have tracked. Set against that, total customer count grew only about 200 sequentially, the weakest net addition in this coverage window, which management attributed to churn at the free-to-paid boundary with minimal revenue effect. The composition is what matters: the paying base deepened materially while the very long tail did not.

Key Operating Metrics

Assessment: RPO deserves attention. Total RPO has now been essentially flat in dollars for three consecutive quarters at $3.46B, $3.48B and $3.47B, with the growth rate falling from 52% to 43%. Management has consistently told investors not to read RPO as a demand signal, and the argument has been right on the way up. It should be applied symmetrically now that the number has stopped growing, but a flat commitment book alongside 38% billings growth is a tension nobody raised on the call.

Key Topics & Management Commentary

Overall Management Tone: Deliberately front-footed about bad news, which is a change. The largest customer's usage reduction was volunteered in the prepared remarks rather than extracted in Q&A, and management framed the decision to fully de-risk the guide as protecting the rest of the story from being overshadowed. On everything else the posture was the most confident of the four calls we have covered, with unusually specific claims about sustainability. The reticence was confined to magnitude: the company will now say that the reduction happened and will not say how large it is, and the two metrics withheld since last year stayed withheld.

1. The Largest Customer Reduced Usage, and the Guide Absorbs All of It

Management disclosed a nine-figure renewal with its largest AI customer alongside a usage reduction beginning in Q3, and said the full-year guide has been fully de-risked for that account. The disclosure came in the prepared remarks from both the CEO and the CFO.

"Regarding our largest customer, we have seen a usage reduction, which is incorporated in our Q3 and full year 2026 guidance. As Olivier noted, this customer has recently renewed with us." — David Obstler, CFO

Pressed on whether the reduction reflects a lower unit price on a new commitment or genuine downsell, management declined to separate the two and explained the choice to over-discount.

"we did see a reduction in usage, and we took the liberty to fully derisk the guidance for the rest of the year with respect to that customer. And again, the reason for that is we don't control what's happening to a specific customer, but we do have a great amount of control on what's happening to everything else in the business, and the business is booming, and we don't want that to overshadow basically the acceleration we see pretty much everywhere else in the business." — Olivier Pomel, Co-Founder and CEO

Assessment: This is the risk we opened at initiation in November, when the same account was renewed at a lower unit rate and we wrote that management's hedge, that consumption growth would outrun the price cut, was an assumption rather than a disclosure. That assumption has now failed. The credit due to management is that it volunteered the fact, over-discounted the guide rather than under-discounting it, and did so in prepared remarks. The debit is that four quarters of refusing to size the account means investors still cannot model what a "fully de-risked" number actually removes.

2. The Best Operating Quarter in the Company's History

Setting the customer aside, the print was exceptional. Revenue grew 35.6% with sequential dollars added of $115M, sequential growth of 11.4% was the fastest since Q2 2022, and the core cohort accelerated a fifth consecutive time.

"Within that, our 11% quarter-over-quarter revenue growth is the highest since Q2 2022. And our quarter-over-quarter revenue added of $115 million is a record by a significant margin." — David Obstler, CFO

Assessment: A record sequential dollar add is the hardest metric in this business to manufacture, because it is immune to base effects. A record sequential ARR add in Q1 and a record sequential revenue add here is the operating signal that the Q3 guide obscures.

3. What the Q3 Guide Actually Implies

The Q3 revenue guide of $1.135B to $1.145B implies 1.7% sequential growth at the midpoint, against the $115M record just delivered. Year-over-year it implies 28% to 29% against 35.6% reported. An analyst asked directly how much of that step-down is the customer.

Q: "given your commentary around how strong the core is and that your largest customer was not additive to growth here in 2Q, should we assume that if we ex out the sequential downtick in that customer that the core guide would have been probably 300 or 400 basis points higher?"
— Brad Reback, Stifel

A: "Well, I can't speculate. But what I will say is, look, the business overall is growing at the same rate. if you exclude that customer, as I said. And the business has been accelerating overall."
— Olivier Pomel, Co-Founder and CEO

Assessment: The question proposed a 300 to 400bps quantification and management declined to confirm or deny it. On the company's own pattern of beating the quarterly guide ceiling by roughly 4%, a Q3 print near $1.19B would represent about 34% growth, which is close to the reported rate. The market is pricing the guide at face value plus a fear that the reduction is larger or more permanent than described, and management has given it nothing with which to argue otherwise.

4. Operating Leverage Finally Appeared

Operating expenses grew 26% year-over-year against 31% in Q1 and 36% in the year-ago quarter, while revenue grew 35.6%. That is the first quarter in this coverage window where revenue growth exceeded operating expense growth by a wide margin, and it produced 307bps of year-over-year operating margin expansion despite the June conference.

"Q2 OpEx grew 26% year-over-year versus 31% last quarter and 36% in the year ago quarter. We held our DASH conference -- user conference in June, and as expected, the event cost about $15 million." — David Obstler, CFO

Assessment: For four quarters we have written that the operating leverage story was on hold while the company deliberately reinvested an accelerating top line. This is the quarter it stopped being on hold. Full-year operating income guidance of $1.01B to $1.03B at a 23% margin is up from about 21% guided in February, and the underlying spending discipline is now visible rather than promised.

5. Cash Conversion Recovered, Against the Prevailing Narrative

Free cash flow of $278.7M grew 68.6% with the margin up 485bps year-over-year to 24.9%. First-half free cash flow of $567.8M is a 26.7% margin against 25.8% a year ago. Capital expenditures fell to $9.9M from $15.2M and capitalized software rose to $27.3M from $19.6M.

Assessment: We downgraded this pillar to Emerging last quarter on an 86% jump in capitalized software and a 337bps year-over-year margin decline. Combined capital intensity grew just 7.1% this quarter and the margin comparison reversed. Note that the widely circulated "29% to 25%" framing compares Q1's seasonal peak to Q2's conference quarter, which is not the comparison that tells you anything about the trend.

6. RUM Crossed $200M ARR Growing Above 50%

Real User Monitoring passed $200M in ARR and accelerated to more than 50% year-over-year growth, disclosed for the first time. Management tied the acceleration to customers pairing RUM with the newer Product Analytics product.

"RUM, or Real User Monitoring, now exceeds $200 million in ARR and accelerated at its scale to over 50% growth year-over-year." — Olivier Pomel, Co-Founder and CEO

Assessment: A product accelerating at $200M of scale is unusual and it validates the portfolio disclosure from last quarter, where management said 18 of 26 products were early in their lifecycle and each could exceed $100M. RUM is the worked example of that claim actually happening.

7. DASH Delivered Autonomy, Not Just Assistance

The June conference produced more than 100 new capabilities, the substantive theme being a shift from AI that assists engineers to AI that closes the loop: Bits AI can now create and maintain monitors, identify root cause, recommend and implement fixes, operate within guardrails, learn from prior incidents, and detect symptomatic behaviour before escalation. Bits Release validates deployments, Bits Code generates fixes, and Bring Your Own Cloud lets customers keep data in their own infrastructure.

Assessment: The product direction directly answers the disruption thesis that de-rated this stock last winter, by making Datadog the agent rather than the thing agents replace. Bring Your Own Cloud is the underrated announcement: it addresses data-residency objections that previously disqualified the platform from certain regulated and sovereign workloads, and the company cited its largest-ever BYOC win displacing a legacy logging tool at petabyte scale.

8. The Automation Paradox, Asked and Answered

An analyst put the sharpest version of the bear case directly: if Bits AI automates the workflows that historically drove consumption, does Datadog cannibalise its own usage-based revenue?

"if we provide more value, we get more -- as I was saying earlier on the call, like we sell more software by helping customers make more money or save money or both. And I think if we can automate more and let them do more, we'll provide more value. That's as simple as that. I think the future of observability is not just observing, it's fixing." — Olivier Pomel, Co-Founder and CEO

Management added the observable rebuttal: customers using Bits AI deploy more of the product, create more dashboards and alerts, and add more users.

Assessment: The empirical claim is the useful half and it is checkable over time. The theoretical answer, that value creation and consumption move together in a usage model, is correct in the long run but does not exclude a transition period where automation reduces retained data volumes before it increases them. This is the single most important open question in the multi-year thesis and it now has a management answer rather than an evasion.

9. Gross Margin Declined for a Third Consecutive Quarter

Non-GAAP gross margin of 79.6% is down from 80.2% in Q1, 81.4% in Q4 and 80.9% a year ago. Management repeated the standing framework and explicitly reaffirmed the long-run range.

"As we've discussed in the past, our gross margin varies from quarter-to-quarter with investments into innovations for our customers, offset by efficiency efforts. There's no change in our expectations for gross margin, which has been in the 80% plus or minus range historically." — David Obstler, CFO

Assessment: A cumulative 180bps of sequential erosion from the Q4 peak, alongside three consecutive year-over-year declines, is a trend rather than variation, and it coincides with the onboarding of hyperscale training workloads and petabyte-scale logging deals. Management has now been asked about gross margin in each of the last three quarters and has given the same framework answer each time. The reaffirmed range is a floor commitment, not an explanation.

10. Dilution Is Now a Real Line Item

Diluted shares reached 371.0M against 358.7M a year ago, up 3.4%, and Q3 guidance assumes approximately 378M with the full year at approximately 376M, against roughly 372M guided in February. Stock-based compensation fell to 19.6% of revenue from 21.8%.

Assessment: A rising share price mechanically increases if-converted dilution from the 2029 convertible notes, so a share count climbing while stock-based compensation falls as a percentage of revenue is largely a function of the stock having doubled. It is nonetheless a real drag: non-GAAP EPS grew 41.3% against 56.6% operating income growth, and roughly a third of that gap is the step from net income to per-share.

11. Enterprises Are Now the AI Story, Not Just the AI Natives

Management argued the core acceleration is itself AI-driven, with conventional enterprises modernising and moving workloads to cloud in order to adopt AI, and separately reported an explosion in traffic to the AI-first product surfaces.

"we see it manifest itself in just more transformation, more cloud adoption, more workloads, more modernization from customers. And that's what drives the majority of the known AI customer acceleration." — Olivier Pomel, Co-Founder and CEO

Assessment: This is analytically double-edged. It is the bullish explanation for why the core keeps accelerating, and it also means the ex-AI-customer metric no longer isolates non-AI demand, because the cohort's own growth is increasingly AI-driven. The clean cohort distinction that made this metric so useful for the past five quarters is eroding just as the company stops disclosing the AI cohort's revenue share.

Guidance & Outlook

Decomposition of the raise: the FY26 revenue midpoint rose $140M while the Q2 beat versus the guide midpoint was $46.5M. The residual $93.5M is a forward raise applied to Q3 and Q4 after absorbing the largest customer's reduction. In other words, the rest of the business was raised by more than $93.5M and the customer took some of it back.

Implied Q-over-Q ramp: The Q3 midpoint of $1.14B requires only 1.7% sequential growth. Backing Q3 out of the full-year midpoint leaves roughly $1.19B for Q4, which is 4.6% sequential growth and about 25% year-over-year. The shape is a Q3 air pocket followed by partial recovery, which is what a step-down at a single large account looks like when the rest of the base keeps compounding.

Guidance style: Unchanged, with one addition. The company has beaten its quarterly guide ceiling by between 3.8% and 4.7% in each of the last four quarters, and the FY26 guide has been raised twice, from 18-to-20% growth in February to 25-to-27% in May to roughly 30% now. Layered on top, for the first time in full, is an unquantified de-risking of a single named-but-unnamed account.

Other guidance notes: Cash taxes of $30M to $40M, a 21% non-GAAP tax rate maintained, and capital expenditures plus capitalized software of 4% to 5% of revenue. Still no free cash flow guidance.

Analyst Q&A Highlights

Whether the Reduction Is Price, Churn or Downsell

The first question of the call went straight at the disclosure and asked management to separate three very different explanations. It did not get the separation.

Q: "I was wondering if you could share any additional details in terms of the new contract? Was it a similar duration? And in terms of the lower usage, is that a function of the customer getting lower unit price because of making a new commitment? Or is there some churn downsell that we're seeing through not only for Q3 but for the balance of the year?"
— Sanjit Singh, Morgan Stanley

A: "as usual, we don't want to comment too much on any specific customer because we also don't really control what's happening with any specific customer. We wanted to be transparent about this on the call because we did see a reduction in usage, and we took the liberty to fully derisk the guidance for the rest of the year with respect to that customer."
— Olivier Pomel, Co-Founder and CEO

Assessment: The three explanations offered in the question have materially different implications. A lower unit price on a larger commitment is benign and was already disclosed in Q3 2025. A usage decline is neutral-to-negative depending on cause. A downsell is a competitive event. Management answered at the level of "a reduction in usage" without distinguishing, and the market defaulted to the worst of the three. Volunteering the fact while withholding the taxonomy is what made this a 19% session rather than a 5% one.

How the Rest of the Business Is Growing Without It

The most load-bearing claim of the call was volunteered at the end of the opening exchange on the largest customer, and it is the reason the stock did not fall further.

Q: "I was wondering if you could share any additional details in terms of the new contract? ... And in terms of the lower usage, is that a function of the customer getting lower unit price because of making a new commitment? Or is there some churn downsell that we're seeing through not only for Q3 but for the balance of the year?"
— Sanjit Singh, Morgan Stanley

A: "The last thing I will say because I know so it's on people's minds is if you back out our largest customer from our growth, you get pretty much the same growth rate as the rest of the business has been accelerating very steadily. Actually, we've seen, I think, now 5 quarters of continuous acceleration from the rest of the business, and we feel very good about the -- what we see in the market."
— Olivier Pomel, Co-Founder and CEO

Assessment: Two distinct statements are bundled here and only one is verifiable. That the business ex-largest-customer grows at "pretty much the same rate" as reported is a claim investors cannot check, because the account has never been sized. That the rest of the business has accelerated for five consecutive quarters is checkable, and it matches the ex-AI-customer series we have tracked since initiation. The second claim is the one to lean on.

Why the Q3 Sequential Guide Is Below Historical Q3 Patterns

An analyst noted that the Q3 guide sits below the company's own sequential precedent and asked what assumptions produced it. The answer confirmed the cause without quantifying it.

Q: "I would love to get more color on your Q3 guide specifically. It seems like it's a little below your sequential levels of how you've guided your previous Q3s. I would love to just hear more about what trends you're seeing going into Q3 and maybe what some of the assumptions of the guide are."
— Tracy Prachef (for Eric Heath), KeyBanc Capital Markets

A: "I think it's similar to the methodology. We take what we see and provide some conservatism. And I think we had mentioned in the script that we've been renewed our largest customer, but we've seen us declines relative to the previous quarter. We said that. So that's all taken into consideration in trying to develop a guidance that is consistent with the methodology of conservatism that we've used as a public company."
— David Obstler, CFO

Assessment: The useful content is the confirmation that nothing else changed in the guidance methodology, so the entire deviation from the sequential pattern is attributable to one account plus the usual conservatism. That is a narrower and more benign reading than the price action implies, and it is also the reading management would give whether or not it were true.

Whether the New Contract Is Structurally Stickier

A question sought any structural feature of the renewal, duration, rate card or product breadth, that would raise confidence in the account's stability. The answer was generic by design.

Q: "On the largest customer renewal, is there anything you can tell us around maybe just any changes around the duration or anything that makes this new contract maybe a little stickier in terms of the discounted rate card, the amount of products that they're able to kind of use for better value, anything that increases the conviction level around stickiness?"
— Aleksandr Zukin, Wolfe Research

A: "I won't comment on this other than to say that most of our enterprise customers, as we talked about for a long time, have annual plus and then the pricing is generally volume-based pricing. So I would say, overall, our customers transact with us in that way. And then we have that level of commitment."
— David Obstler, CFO

Assessment: The one substantive detail sits elsewhere in the prepared remarks: the account runs 17 Datadog products. That is a deeper deployment than all but a handful of customers and it is the strongest available argument that the reduction is a usage adjustment rather than the start of an exit. Management chose not to make that argument explicitly, which was a missed opportunity on a day when it needed one.

Whether Automating the Workflow Cannibalises the Consumption Model

The most consequential long-run question of the call asked whether Bits AI eats the activity that generates Datadog's own revenue.

Q: "thinking that Bits AI is going to be increasingly automating activities that historically has created observability workflows. I'm curious and really wonder how do you ensure that greater automation that might be driven by Bits AI doesn't eventually reduce the volume of activity that traditionally drove Datadog consumption?"
— Koji Ikeda, Bank of America

A: "Like when they use Bits AI, they use more of our product. They deploy more of it. They create more dashboards and alerts and everything else. They have more users inside of our product, like it's not a zero-sum game."
— Olivier Pomel, Co-Founder and CEO

Assessment: The answer is empirical rather than theoretical, which makes it the better kind. Customers adopting the agent products expand their footprint rather than contracting it, at least so far and at least at current adoption levels. The caveat is that the observed population is early adopters, who are the least likely group to reduce usage, so this evidence will need to be reproduced at scale.

Whether the Non-AI Acceleration Is Sustainable

With the core cohort five quarters into an acceleration, a question asked what would keep it going and whether budget scrutiny would hit non-AI spend first. Management's answer invoked the 2021 comparison directly.

Q: "I just wanted to sort of get your thoughts on the sustainability and whether this acceleration that you're seeing is driven by some of the new customer logos that you're pointing out or more usage going up? And as CFOs get more sort of cautious around their budgets, do you see more sensitivity around non-AI eventually relative to some of the AI products ..."
— Samik Chatterjee, JPMorgan

A: "For one thing, if you compare to what we have seen in the heady days of 2021 or the growth rates are accelerating, but they're still far below what we were seeing at that time. And so we don't create the same issue of customers having to digest very large increases multiple years in a row. I think in this case, we're very well within the range of sustainability."
— Olivier Pomel, Co-Founder and CEO

Assessment: The strongest sustainability argument management has made in the four calls we have covered. The 2022 to 2024 deceleration was a digestion of 2021 over-provisioning; if current growth is materially below that peak, the same digestion should not recur at the same magnitude. It is a specific and falsifiable mechanism rather than a general assertion of confidence.

Why Net Customer Additions Were the Weakest on Record

Total customers grew only about 200 sequentially, in a quarter that included the DASH conference, which has historically been an accelerant. The question was pointed and the answer located the shortfall at the bottom of the base.

Q: "I wanted to ask about new customer additions. This probably was the weakest quarter I ever remember for you guys, especially in the quarter where you had DASH, where historically DASH has been an accelerant of new customer additions. Any color there would be great."
— Ittai Kidron, Oppenheimer & Co.

A: "Our gross customer additions continue to be strong and on trend line. and that's the vast majority of our revenues. We have at the very low end, the border between free and contract, and that has variability, very low effect on revenues."
— David Obstler, CFO

Assessment: The explanation is consistent with the rest of the disclosure: $100k+ customers grew 22.6% and new-logo bookings more than doubled in enterprise, so the paying base is not what softened. It does mean the headline customer count is now a low-information metric, and it is worth remembering that the same free-to-paid boundary would mask genuine small-customer churn if it ever began.

What They're NOT Saying

1. The size of the largest customer, or of the reduction: Four quarters of questions, now including a quarter in which the company volunteered that the account is shrinking. Investors are told the guide is "fully de-risked" without being told what was removed.
2. Whether the reduction is price, usage or downsell: Asked directly in the first question of the call. Management answered "a reduction in usage" and declined to distinguish among three explanations with very different implications.
3. The AI cohort's share of revenue: Withheld for a third consecutive quarter, in the very quarter it became most decision-relevant.
4. Security ARR growth: Withheld for a third consecutive quarter, despite AI Guard and the security agent being among the headline DASH launches.
5. Why gross margin has fallen three quarters running year-over-year: 180bps of cumulative sequential decline from the Q4 peak, answered each time with the same "varies quarter-to-quarter" framework and never connected to the hyperscale training and petabyte logging workloads onboarded over the same period.
6. Why RPO has been flat in dollars for three quarters: $3.46B, $3.48B, $3.47B, with growth decelerating from 52% to 43%, while billings accelerated to 38%. Not raised by management, not asked by any analyst.
7. Any revenue from the AI product suite: Now three quarters past the first general-availability launches and five since the question was first asked. Adoption is still reported only in usage counts.
8. FY26 free cash flow: Still no guidance, for a fourth consecutive quarter, on the metric that carries the valuation.

Market Reaction

• Pre-print setup: DDOG closed at $283.17 on August 5, up 108.2% year-to-date against the S&P 500's +12.8%, up 106.8% over trailing twelve months, and up 10.3% over the trailing 30 days. The 52-week closing range going in was $102.62 to $288.15, so the shares entered the print within 1.7% of their 52-week closing high.
• Reaction session (August 6, before the open): Opened at $227.45, a 19.7% downward gap, traded between $225.26 and $243.00, and closed at $229.29. That is a 19.0% single-session decline, or $53.88 per share.
• Volume: 15.0M shares against a 30-day average of 4.3M, or 3.5 times normal.
• Relative move: The S&P 500 fell 0.2% on the same session, making the relative move roughly 19 points.

The mechanics are unusually clean. A company that beat revenue by 3.8%, beat EPS by 12.1%, raised full-year revenue guidance by $140M and raised full-year EPS guidance by 5.0% lost a fifth of its market value in one session. Every element of that decline traces to a single sentence in the prepared remarks and to the Q3 number that sentence produced.

The setup explains the magnitude. Entering the print the stock was up 108.2% year-to-date and sat 1.7% below its 52-week closing high, so it was priced for continued acceleration rather than for a one-account air pocket. When the Q3 guide implied 1.7% sequential growth against a $115M record just delivered, the market did not have to revise its view of the business to justify a large move; it only had to revise the growth path by a quarter or two. At that starting multiple, that was enough.

Volume of 3.5 times normal is notably lighter than the 5.5 to 5.7 times that accompanied the two big up-sessions in this coverage window. That is consistent with a repricing driven by a specific, bounded disclosure rather than a wholesale change of view, and it leaves the stock at $229.29, still 21.5% above the level at which we downgraded three months ago and 123.4% above the 52-week closing low.

Street Perspective

Debate: Is the Largest-Customer Reduction Bounded or the First Instalment?

Bull view: The account renewed, runs 17 Datadog products, and management chose to fully de-risk the full year rather than guide to hope. Product breadth at that depth is not something a customer unwinds quickly, and a company planning for further declines does not simultaneously raise the full year by $140M. The reduction is a bounded usage adjustment already fully absorbed.

Bear view: The same account was repriced downward in Q3 2025 and has now reduced usage in Q2 2026, so this is the second adverse event in four quarters from one customer that has never been sized. AI-native workloads are exactly where customers have both the technical capability and the cost incentive to bring observability in-house or re-architect it away. One data point is noise; two in the same direction is a trend.

Our take: The bear has the stronger pattern argument and the bull has the stronger structural one. Our reading is that 17 products deployed on production workloads at scale is a genuinely high exit barrier, and that management's decision to over-discount reduces the probability of a further negative surprise from this account this year. But we flagged this account as the largest unquantified risk in November and were told the consumption growth would outrun it. That did not happen, and we are not inclined to accept a second reassurance that cannot be checked.

Debate: Does the Core Justify Looking Through Q3?

Bull view: Five consecutive quarters of core acceleration to the high 20s, a record $115M sequential add, billings up 38%, new logos more than doubling in enterprise, operating expense growth decelerating to 26% and free cash flow margin up 485bps year-over-year. Every part of the business the company controls is compounding faster than a year ago. One customer does not change that.

Bear view: The metric that made the core story checkable is decaying. Management now says the core acceleration is itself driven by customers adopting AI, which means the ex-AI-customer cut no longer isolates non-AI demand. Meanwhile RPO has been flat in dollars for three quarters. If the AI cycle turns, the "core" turns with it, and the disclosure needed to see that coming has been withdrawn.

Our take: The bear's point about metric decay is the sharpest observation available on this print and nobody made it on the call. The core cohort is becoming an AI-exposed cohort by management's own description, at the same time as the company stopped publishing the AI cohort's revenue share. We still think the operating evidence is strong enough to look through a single-quarter air pocket, but the analytical tools for verifying that are getting worse, not better.

Debate: Is a 19% Drop an Entry Point?

Bull view: A 19% decline on a beat-and-raise is the classic overreaction. The business is growing 36%, guiding to roughly 30% for the year with a fully de-risked largest customer, expanding operating margin, and generating a 27% first-half free cash flow margin. Buying quality after a single-issue sell-off is how the February upgrade worked.

Bear view: The stock entered the print up 108% year-to-date and 1.7% off its high. After a 19% decline it is still 21.5% above where it traded three months ago and 123% above its 52-week low, at 17.5x enterprise value to the raised full-year guide and roughly 65x this year's free cash flow. A drop from an extreme is not the same as a discount.

Our take: The bear is right and it is not close. Our February upgrade worked because the stock had de-rated to 12.3x trailing revenue while fundamentals improved. This is the opposite setup: the multiple is higher than at either of our two previous rating decisions, and the de-rating has taken the stock back to roughly where it traded six weeks ago. A 19% headline is not a valuation.

Model & Valuation Framework

Model Update

Valuation

At the August 6 close of $229.29 and roughly 358M shares outstanding (Q2 weighted-average basic was 356.3M), market capitalization is approximately $82.1B. Netting $5.0B of cash and marketable securities and adding back the $1.0B of 2029 convertible notes gives an enterprise value near $78.1B. With the company halfway through its fiscal year, we anchor on next-twelve-month revenue of approximately $5.06B rather than on the calendar year.

The February and May columns are computed on the enterprise value at those closing prices against the same denominators used in each of those notes, and are shown to make the direction of travel explicit: on every measure the stock is more expensive after a 19% decline than it was at either prior rating decision.

Fair value framework: Applying a 14x to 18x multiple to our next-twelve-month revenue estimate of $5.06B produces an enterprise value range of $70.8B to $91.1B, or $209 to $266 per share after adjusting for net cash and the converts. Our central case of 16x gives $237.

Valuation impact: Our 12-month fair value estimate rises to $237 from $205, entirely because the valuation base rolls forward from FY26 to next-twelve-month revenue as the year progresses. The multiple range is unchanged from May. At $229.29 the remaining gap to fair value is roughly 3%, narrower than the 9% that produced the May downgrade.

Thesis Scorecard Post-Earnings

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

Overall: Thesis intact and more nuanced. All four bull pillars are confirmed or unchanged, and the cash-conversion bear point resolved favourably. But the concentration risk that has been the central caveat of this coverage since the first report has now converted from a hypothetical into a disclosed reduction, and it did so in the quarter the stock was most expensive.

Action: Hold. The operating story is stronger than at any point in this coverage and the price does not compensate for the risk that just materialized.

Bottom Line: Maintaining Hold

In November we initiated coverage of Datadog and wrote that the repricing of its largest AI customer was "the single largest unquantified item in the print," that management's hedge was that the customer's consumption growth would outrun its own price cut, and that this was "an assumption, not a disclosure." Nine months later that assumption has failed. The company disclosed on this call that the account has reduced usage beginning in the third quarter and that the remainder of the year has been fully de-risked for it.

Everything else about the quarter was the best this company has produced. Revenue grew 35.6% to $1,121.5M, beating consensus by 3.8% and the guide ceiling by $41.5M. Sequential revenue added of $115M was a company record by a wide margin. The core cohort accelerated for a fifth consecutive quarter to the high 20s. Operating expense growth decelerated to 26% and delivered 307bps of year-over-year operating margin expansion despite the June conference. Free cash flow margin rose 485bps year-over-year, reversing the deterioration we flagged in May, and full-year revenue guidance still went up $140M after absorbing the customer reduction.

Management deserves credit for how it handled the bad news. The reduction was volunteered in prepared remarks rather than extracted in questioning, and the guide was over-discounted rather than under-discounted. What management would not do is size it. Four quarters of questions about this account have produced a confirmed price cut, a confirmed usage reduction, an explicit extra guidance discount applied twice, and still no revenue figure. Investors are being asked to accept that the guide is de-risked without being told what was removed, and that is why a beat-and-raise lost 19% of the equity value in a session.

The rating question is simpler than the narrative. A 19.0% decline reads like an opportunity, but the shares entered the print up 108.2% year-to-date and within 1.7% of their 52-week closing high, and closed at $229.29. That is 21.5% above the $188.73 where we downgraded in May and 77% above the $129.67 where we upgraded in February. On enterprise value to current-year guided revenue the stock is at 17.5x, against 14.6x in May and 10.3x in February. It is more expensive after the drop than at either prior decision.

Our fair value estimate rises to $237 as the valuation base rolls to next-twelve-month revenue, leaving roughly 3% of upside. We would become buyers again in the low $190s, or on a Q3 print that clears the de-risked guide by the customary margin, or on the company finally sizing the account that has now moved this stock twice. Until one of those arrives, we maintain Hold.