Best Quarter in Years, Worst Entry Point in Months: Downgrading Datadog to Hold on a 31% Session

Results vs. Consensus

Q1 2026 Scorecard

Versus Company Guidance (issued February 10, 2026)

The 4.7% revenue beat against the guide ceiling is the third consecutive quarter in a tight band, after 4.1% in each of Q3 and Q4. The operating-income beat of 9.0% is roughly double the revenue beat, the same asymmetry seen in every quarter of this coverage.

Year-Over-Year Comparisons

Two lines on this table settle debates we have been carrying. Stock-based compensation grew 19.8% against 32.2% revenue growth, taking the ratio to 19.6% from 21.6%, its first move below 20% in the period we have followed. Against that, capitalized software grew 85.7% and is the single largest reason free cash flow grew only 18.3% while operating cash flow grew 23.2%.

Quarter-Over-Quarter Comparisons

The sequential margin and cash-flow declines are seasonal rather than structural. Q1 carries the annual reset of payroll taxes on equity vesting and a full quarter of the prior year's hiring, and Q4 is the seasonal high point. The comparison that matters is year-over-year, where non-GAAP operating margin expanded 35bps. That said, free cash flow margin fell on both bases, which is not seasonal.

Quality of Beat

Platform & Cohort Performance

Revenue Cohort Mix

The core cohort

The pillar that has carried this thesis since initiation accelerated a fourth time. Management gave the figure as the mid-20s percent, against 23% in Q4 and 19% in the year-ago quarter, and again attributed it to breadth rather than to any single driver.

"non-AI customer revenue growth accelerated again this quarter to mid-20% year-over-year up from 23% last quarter and 19% in the year ago quarter. We think this is a sign of strong continued cloud migration, greater adoption of our products, and customers have all kinds accelerating their use of AI." — Olivier Pomel, Co-Founder and CEO

Assessment: Four consecutive accelerations from 18% to the mid-20s is the strongest evidence available that the 2023 to 2024 optimization drag has not merely cleared but reversed. The one qualification is that the third of management's three attributed causes, customers accelerating their own AI use, blurs the very cohort distinction the metric exists to preserve. The core is increasingly AI-exposed even where the customers are not AI natives.

The AI-native cohort

The cohort now includes 22 customers spending above $1M annually, up from 19, and management disclosed for the first time that five spend above $10M. Two new land deals were signed with the AI research divisions of two of the world's largest technology companies, both for training workloads.

Assessment: The $10M tier is the disclosure that matters. Five customers above $10M means the cohort has a genuine top end that is not one account, which is a materially different risk profile from the single-whale structure we flagged at initiation. The revenue-share percentage remains withheld for a second consecutive quarter, so the concentration cannot be measured directly, but the shape keeps improving.

Product Portfolio (first-time disclosure)

Assessment: This is a better disclosure than last quarter's three-pillar ARR figures because it sizes the tail rather than the head. Eight products above $50M against 26 total says the platform's growth has been carried by roughly a third of the portfolio, and that the attach metrics investors watch are being driven by products that are individually still small. It is a credible bull argument and it is also unfalsifiable, since no product-level growth rates were given.

Platform Adoption Density

Assessment: The 8-or-more band widening seven points, from 13% to 20%, matches the widest density gain we have tracked. The AI-integration statistic is the more strategically interesting one: the 20% of customers sending AI data represent roughly 80% of ARR, which means AI adoption is concentrated in exactly the accounts that already spend the most. That is the shape you want if the thesis is that AI adds a layer to existing spend rather than replacing it.

Key Operating Metrics

Assessment: Net revenue retention finally moved. In Q3 management explained that the trailing-twelve-month construction meant an accelerating business would show up in the metric with several quarters of lag, and that is precisely what has now happened. Current RPO growth re-accelerating to the mid-40s from about 40% is the second confirming signal, and it partially answers the deceleration we flagged as unexplained last quarter. Security ARR growth remains undisclosed for a second quarter, which is now a pattern rather than an omission.

Key Topics & Management Commentary

Overall Management Tone: The most confident of the four calls we have covered, and unusually willing to volunteer superlatives that can be checked against future prints: highest first-quarter sequential dollar add ever, strongest sequential usage growth since early 2022, all-time record ARR added, record new-logo bookings and record average land size. The forward-looking commentary shifted from defending the category against agentic disruption, which dominated the prior call, to describing AI as an additional secular growth driver alongside cloud migration. The only reticence was on the two metrics that remain withdrawn and on the sequential decline in gross margin, which drew a single sentence of explanation and no follow-up.

1. A Fourth Consecutive Core Acceleration, and the First $1B Quarter

Revenue reached $1,006.4M, the first billion-dollar quarter in company history, with total ARR passing $4B. Growth accelerated to 32% from 29% in Q4 and 25% a year ago. The core cohort excluding AI natives reached the mid-20s. Management framed the sequential mechanics in records rather than percentages.

"Our 6% quarter-over-quarter revenue growth is the highest for Q1 and since 2022. And our $53 million quarter-over-quarter revenue added is the highest ever for Q1. That included the strongest quarter of sequential usage growth from existing customers since the first quarter of 2022. We also delivered an all-time record for sequential ARR added to the quarter." — David Obstler, CFO

Assessment: Sequential dollar records are harder to game than growth percentages because they do not benefit from an easy comparison. An all-time record ARR add in a first quarter, which is seasonally the weakest, is the strongest single operating datapoint in the four quarters we have covered.

2. The Full-Year Raise Is Mostly Forward, Not Pass-Through

FY26 revenue guidance moved to $4.30B to $4.34B from $4.06B to $4.10B, a $240M increase at the midpoint. The Q1 beat versus the guide midpoint was $50.4M, leaving roughly $190M of genuine forward raise. Operating income guidance rose $100M at the midpoint and EPS guidance rose 13.2%.

"for fiscal 2026, we expect revenues to be in the range of $4.3 billion to $4.34 billion, which represents 25% to 27% year-over-year growth. Non-GAAP operating income is expected to be in the range of $940 million to $980 million, which implies an operating margin of 22% to 23%." — David Obstler, CFO

Assessment: Three months ago this company guided FY26 to 18% to 20% growth and we argued in print that the guide should be read as an instrument rather than a forecast. It has now been raised to 25% to 27% after a single quarter. The operating margin guide moving up to 22% to 23% from about 21% also confirms the pattern we described: the guided margin is depressed by the conservative revenue guide, and rises as revenue is de-risked.

3. Training Became a Real Market

The company landed seven-figure and eight-figure annualized deals with the AI research divisions of two of the world's largest technology companies, both for hyperscale training workloads, using the GPU monitoring product that went generally available this quarter. Management was explicit that this is a change from a year ago.

"Interestingly, last year, when we reported earnings, we said we're mostly interested in inference workloads and training is not a real market for us yet. Now we actually see training becoming a market." — Olivier Pomel, Co-Founder and CEO

The reasoning offered is that training has moved from an artisanal research activity conducted by a handful of firms to a production workload with uptime economics, where a failed run costs a week of competitive position.

Assessment: This is a genuine TAM addition rather than a reframing. At initiation, management told us GPU monitoring generated no significant revenue and called it a future opportunity. Two quarters later the product is generally available and has anchored a seven-figure and an eight-figure landing. That is the free option in our Bull #4 pillar beginning to pay, and it is the first time that pillar has produced revenue evidence rather than usage metrics.

4. Hyperscalers Are Buying, Which Is the Hardest Case

Beyond the AI research divisions, management said it landed parts of hyperscalers during the quarter, organisations with both the balance sheet and the engineering depth to build internally.

"The thing that's also interesting in particular this quarter is that we also landed some large parts of hyperscalers. And hyperscalers typically have a culture of building everything themselves. And the certainly have the balance sheet and the human capital to support some of that build-out. Like if there was ever a set of companies for whom it makes sense to do it themselves, and we do those companies." — Olivier Pomel, Co-Founder and CEO

Assessment: Last quarter the counter-evidence to the build-versus-buy thesis was a frontier AI lab. This quarter it is hyperscalers, which is a stronger case still because they sell competing observability tooling of their own. Two consecutive quarters of the most-capable possible buyers choosing to buy is close to the practical limit of what can be demonstrated on this question.

5. Net Revenue Retention Rose for the First Time

Trailing-twelve-month net revenue retention moved to the low 120s from about 120%, the first increase since we initiated. Current RPO growth re-accelerated to the mid-40s from about 40%, and billings grew 37% against 34% in Q4.

"Our trailing 12-month net revenue retention percentage was in the low 120%, up from about 120 last quarter and our trailing 12-month gross retention percentage remains in the mid- to high 90s." — David Obstler, CFO

Assessment: In Q3 the CFO explained that the trailing-twelve-month construction would show the acceleration with a lag of several quarters. That prediction has now been borne out, which is a small but real credibility deposit. The corollary we flagged at the time still stands: the same lag will overstate expansion on the way down.

6. Four Products Reached General Availability, Including the AI Agents

GPU Monitoring, the Datadog MCP Server, Bits AI Security Analyst and Experiments all reached general availability in the quarter, alongside FedRAMP High certification for the government business. The usage metrics on the AI products compounded sharply: SRE agent investigations more than doubled from December to March, LLM Observability spans nearly tripled sequentially, and MCP server tool calls quadrupled sequentially.

"We now have over 6,500 customers sending data for 1 or more of AI integrations. Though this is only 20% of total customers, they represent about 80% of our ARR." — Olivier Pomel, Co-Founder and CEO

Assessment: The shipping cadence is impressive and the usage growth rates are steep, but this is now the third consecutive quarter with adoption counts and no revenue figure for any AI product. The Bits AI monetization question first asked at our initiation remains unanswered two quarters after general availability. Usage metrics compounding at 2x to 4x per quarter off small bases is a weaker signal than it appears.

7. Gross Margin Fell 124bps Sequentially, Explained in One Sentence

Non-GAAP gross margin was 80.2%, down from 81.4% in Q4 and 80.3% a year ago. The explanation offered was the standing framework rather than anything quarter-specific.

"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." — David Obstler, CFO

Assessment: The year-over-year figure is essentially flat at 11bps lower, so on the company's own "plus or minus 80%" framing there is nothing anomalous here. What is worth watching is the coincidence of a 124bps sequential gross margin decline with the onboarding of hyperscale training workloads, which are the most infrastructure-intensive revenue the company has ever taken on. No analyst connected the two on the call.

8. Free Cash Flow Growth Lagged Revenue by 14 Points

Free cash flow grew 18.3% against 32.2% revenue growth, and the margin fell 337bps year-over-year to 28.7%. Operating cash flow grew a healthier 23.2%. The gap is capitalized software development costs, which rose 85.7% to $34.2M. Asked about capital intensity given the growth in telemetry volumes, management pointed to its cloud-hosted model.

"sort of the investments we're making right now, you we run most of our workloads on cloud, meaning you'll see all of that in OpEx, nothing CapEx. So we have low CapEx." — Olivier Pomel, Co-Founder and CEO

Assessment: The answer addresses property and equipment, which is indeed small at $11.4M, and does not address capitalized software, which is three times larger and grew 86%. Capitalizing development spend raises non-GAAP operating income and lowers free cash flow simultaneously, so a rising capitalization rate makes the margin story look better than the cash story. This is now the metric we would most like a fifth quarter of data on.

9. The Largest-Customer Conservatism Was Maintained

The FY26 guide continues to apply an explicit extra discount to the largest customer, the same construction disclosed last quarter that our Q4 note showed implies roughly flat growth from that account.

"our guidance philosophy overall remains unchanged. As a reminder, we base our guidance on trends observed in recent months, and apply conservatism on these growth trends. In addition, as with last quarter, we are applying a higher degree of conservatism to our largest customer." — David Obstler, CFO

Assessment: Maintaining the carve-out while raising the total guide by 5.9% means the raise came entirely from the rest of the business. That is the better composition. It also means the account remains a live uncertainty two quarters after we first flagged it, and the company has now chosen to discount it twice without ever sizing it.

10. Sovereignty and the Public Sector Became Investment Lines

The company received FedRAMP High certification, announced a planned UK data center, and described bring-your-own-cloud products that run on customer infrastructure. Management framed data residency and AI sovereignty as a rising source of customer demand requiring geographic and certification investment.

Assessment: FedRAMP High opens federal workloads that were previously inaccessible, and management confirmed it has been building channel and sales capacity ahead of the certification rather than after it. This is a multi-year revenue line with a slow ramp and no near-term modelling implication, but it widens the addressable market in a direction that is structurally insulated from the agentic-disruption debate.

11. Operating Expense Growth Re-Accelerated to 31%

Q1 non-GAAP operating expenses grew 31% year-over-year against 29% in Q4 and 29% in the year-ago quarter, which management again framed as execution against the hiring plan rather than cost pressure. The FY26 operating margin guide nonetheless moved up to 22% to 23% from about 21%.

Assessment: Non-GAAP opex growth of 31% against 32% revenue growth is the definition of running flat, and it is a deliberate choice to reinvest an accelerating top line rather than harvest it. The FY26 margin guide rising while opex growth accelerates tells you the revenue guide moved further than the spending plan did, which is the same mechanic that produced 22.4% delivered against about 20.9% initially implied in FY25.

Guidance & Outlook

Decomposition of the raise: the FY26 revenue midpoint rose $240M while the Q1 beat versus the guide midpoint was $50.4M. Roughly $190M, or 79% of the increase, is a forward raise applied to Q2 through Q4 rather than a pass-through of the quarter. That ratio is what distinguishes this from a routine beat-and-raise.

Implied Q-over-Q ramp: The Q2 midpoint of $1.075B requires 6.8% sequential growth, which management quantified as $64M to $74M of sequential dollars. Q2 2025 grew 8.6% sequentially, so the guide is below last year's sequential rate despite a stronger entry point. The Q2 operating margin guide of 21% to 22% includes roughly $15M of DASH user-conference cost, which is the same June seasonal load that made Q2 2025 the margin trough.

Guidance style: Unchanged and now four quarters observed. Q3, Q4 and Q1 each beat the quarterly guide ceiling by between 4.1% and 4.7%. The FY26 guide has already been raised from 18-to-20% growth to 25-to-27% growth after one quarter, and the explicit largest-customer discount remains in place on top.

Other guidance notes: Cash taxes of $30M to $40M for FY26, a 21% non-GAAP tax rate maintained, and capital expenditures plus capitalized software of 4% to 5% of revenue. No free cash flow guidance was given.

Analyst Q&A Highlights

Whether the Explosion in Machine-Generated Code Converts to Platform Activity

The call opened by asking whether the volume of code now being produced by coding agents is actually reaching production, which is the precondition for it generating observability demand. Management confirmed it is seeing the conversion at every layer.

Q: "is there any way to conceptualize the growth in the sheer raw volume of, code is being produced in the world today due to adoption of code generators such as Quad code and Codex and cursor, because they seem to be developing the capability to take on full projects and some of the charts are showing these capabilities are just exponentially exploding upward in a straight line. I'm wondering how much of that code is going into production and therefore, driving activity for Datadog."
— Mark Murphy, JPMorgan

A: "We definitely think and see that the there's many more applications being created. There's going to be way more complexity in production. We see some of that happening already today. Some of those new applications are getting into production, they're finding users. We see some signs of that at every layer of our platform."
— Olivier Pomel, Co-Founder and CEO

Assessment: The bear case on this stock holds that agents reduce the need for observability. Management's answer inverts it: agents increase the ratio of production code to human comprehension, and observability is what closes that gap. The mid-20s core acceleration is consistent with that reading, which is the first time the argument has had a matching number behind it.

Whether Silicon Heterogeneity Is a Tailwind

A follow-up asked whether the proliferation of custom accelerators alongside conventional CPUs creates monitoring complexity that favours a vendor-neutral platform. The answer said yes and then disclosed a more consequential change about training workloads.

Q: "when you look across the Amazon with Trinium and Graviton and Google with TVs, Microsoft has launched the myosilicon. It looks like that is starting to explode. In our understanding is that trying to monitor the mixed environment is a lot more difficult than if you just have a uniform fleet of Intel and AMD chips ... Can you speak to whether that trend is giving you some tailwinds?"
— Mark Murphy, JPMorgan

A: "I think the heterogeneity of the silicon is definitely a trend that plays in our favor there. The more heterogeneous, the more you need someone else to make sense of everything for you ... Interestingly, last year, when we reported earnings, we said we're mostly interested in inference workloads and training is not a real market for us yet. Now we actually see training becoming a market."
— Olivier Pomel, Co-Founder and CEO

Assessment: The heterogeneity argument is sound but small; the training disclosure buried in the same answer is the material one. A company that told investors twelve months ago that training was not a market for it has now landed seven-figure and eight-figure training deals. That is a TAM expansion arriving faster than the roadmap implied.

Build Versus Buy, as It Actually Plays Out

A question about the persistent industry noise around open-source observability alternatives drew a description of the consolidation motion, and then the hyperscaler disclosure that is the strongest available counter-evidence on the topic.

Q: "if I listen to you in your prepared remarks, there's a lot of like consolidation that people try to do open source tooling and then realize they kind of needed to come to you and come back. On the other hand, in the industry, we still have a lot of like noise around that level. How do you see it in real life."
— Raimo Lenschow, Barclays

A: "these customers have 4 or 6, 7, 15, and 25 different things, and different pockets in the organization, and different business units, and it's a huge mass. And they come to us, they can unify all that. They get better results because all of the data is in one place ... And also they save money because they don't have all these pockets in efficiency everywhere."
— Olivier Pomel, Co-Founder and CEO

Assessment: The consolidation economics are the durable part of this business: a customer running 25 tools who moves to one platform reduces spend and increases Datadog's share of it simultaneously. The CFO's supporting point, that the product-density metrics are the observable proof of consolidation, is the right way to audit the claim from outside.

Why Training Is Inflecting Now, and What It Attaches At

A follow-on asked management to explain the timing of the training opportunity and to size observability spend as a proportion of AI workload spend. The first half was answered clearly; the second was not.

Q: "Why do you think the training opportunity it's happening now or inflecting now? ... how do we think about the attach rate on trading versus inference of observability?"
— Gabriela Borges, Goldman Sachs

A: "now it's turning into production. It's turning into something that many more companies are doing. It's scaling by orders of magnitude, and it's becoming something that has to be on all the time, reliable and every minute you lose is or whether every fellow you have in your training around is a week you give away to the competition."
— Olivier Pomel, Co-Founder and CEO

Assessment: The qualitative explanation is convincing: training stopped being a research activity and became a production workload with downtime costs, which is exactly when observability becomes purchasable. The attach-rate question went unanswered beyond a reference to the 6,500-customer integration count, so there is still no way to size AI-driven observability demand from the outside.

What Underpins the Largest Sequential Dollar Guide in Company History

The Q2 guide implies a sequential dollar increase materially larger than anything the company has previously delivered. The question asked what supports that confidence, and specifically whether it depends on two large AI research customers ramping.

Q: "Even if you put up a modest beat on that guide, it's going to be by order of magnitude, the largest sequential dollar at I think, in the company's history. And I just wanted to unpack what's giving you that confidence? And in particular, is there anything interesting to call out, David, in terms of the ramp of a couple of the larger research labs, one of which renewed with you guys in the fourth quarter, another one just landed."
— Karl Keirstead, UBS

A: "As you know, we're recurring revenue model. So the biggest indication of in the near term of the next quarter is the ARR growth in the previous quarter. And when we said we had a record. So essentially, at the bedrock of this is sort of the run forward of ARR that we've already signed. The ARR add was very broad-based and was not very concentrated."
— David Obstler, CFO

Assessment: The most useful answer on the call. It establishes that the Q2 guide rests on ARR already contracted rather than on assumed new bookings, and the CEO's follow-on that the quarter would still have set an ARR-add record excluding the largest Q4 landing directly rebuts the concentration concern the question raised. This is the answer that makes the raise credible.

What the Category Looks Like When Agents Do the Investigating

A question about the long-run product and pricing shape of observability in a world where agents rather than engineers consume the platform drew an unusually candid admission about forecast horizons.

Q: "what is your sort of vision of that -- how that evolves for Datadog, both from a product standpoint and an experience standpoint from a UI perspective, but also like is there going to be immune modalities in terms of pricing when agents are consuming the Datadog platform to a higher degree than engineers do today?"
— Sanjit Singh, Morgan Stanley

A: "it's hard to tell where we're going to be in 4 or 5 years. If you had told me 2 years ago that most engineers would go back to coating in the console. I wouldn't have believed you ... we don't care whether most of the usage is humans, most of usages agents. Our business model lends itself to do pretty well like we are usage-based it doesn't really matter where the is coming from that perspective."
— Olivier Pomel, Co-Founder and CEO

Assessment: The structural point is correct and underappreciated: a consumption model is indifferent to whether the consumer is a human or an agent, which makes Datadog less exposed to the seat-count compression that threatens per-user software. The admission that the five-year shape is unknowable is honest and it is also the reason the multiple is volatile.

Whether Rising Telemetry Volumes Change the Capital Intensity of the Business

With training workloads arriving and telemetry volumes described as growing sharply, a question asked what keeps capital intensity contained. The answer addressed property and equipment and did not address the line that actually moved.

Q: "I wanted to ask you about the structural changes to the capital intensity of the business ... I wanted to get a better understanding of what sort of extrinsic or intrinsic engineering efforts you're undertaking to keep a very efficient CapEx envelope in spite of the fact that it seems like that would increase because of the torrent of telemetry we're seeing on the platform."
— Fatima Boolani, Citi

A: "we run most of our workloads on cloud, meaning you'll see all of that in OpEx, nothing CapEx. So we have low CapEx. If it changes, we'll tell you, like if for some reason, we decide to make different kinds of investments and some of it more front some it more CapEx, we'll tell you, but that's not the case today."
— Olivier Pomel, Co-Founder and CEO

Assessment: Accurate as far as it goes, and incomplete. Purchases of property and equipment were $11.4M, genuinely small. Capitalized software development costs were $34.2M, up 85.7% year-over-year, and are the reason free cash flow grew 14 points slower than revenue. The question was about capital intensity and the answer covered only the smaller half of it.

What They're NOT Saying

1. The AI-native share of revenue: Withheld for a second consecutive quarter after being disclosed at 11% and 12%. The cohort is now described only in customer counts and spending tiers.
2. Security ARR growth: Also withheld for a second consecutive quarter after mid-40s and mid-50s figures in Q2 and Q3 2025. Security products shipped and were discussed; the growth rate was not.
3. The size of the largest customer: A "higher degree of conservatism" has now been applied to it twice without the account ever being sized, and without the magnitude of the discount being quantified.
4. Why capitalized software rose 86%: Not volunteered, and the one question about capital intensity was answered on property and equipment instead. This single line explains most of the 337bps decline in free cash flow margin.
5. What drove the 124bps sequential gross margin decline: Attributed to normal variation between innovation investment and efficiency work, with no quantification, in the same quarter the company onboarded hyperscale training workloads.
6. Any revenue from the AI product suite: Four products reached general availability this quarter and the AI agents have been generally available since December. Adoption is reported in investigation counts, span counts and tool calls. No dollar figure has been given for any of them.
7. The observability attach rate on AI workloads: Asked directly how to benchmark observability spend as a percentage of training or inference spend; the answer redirected to the customer-count metric.
8. FY26 free cash flow: Still no guidance, in a year where capitalized software is guided higher and the FCF margin has now declined year-over-year for three consecutive quarters.

Market Reaction

• Pre-print setup: DDOG closed at $143.71 on May 6. The stock was up 5.7% year-to-date against the S&P 500's +7.6%, up 35.5% over trailing twelve months, and up 23.3% over the trailing 30 days. The 52-week closing range going in was $102.62 to $199.72.
• Reaction session (May 7, before the open): Opened at $187.99, a 30.8% gap, traded between $179.47 and $198.60, and closed at $188.73. That is a 31.3% single-session gain, or $45.02 per share.
• Volume: 26.0M shares against a 30-day average of 4.7M, or 5.5 times normal.
• Relative move: The S&P 500 fell 0.4% on the same session, making the relative move roughly 32 points.

A 30.8% opening gap that holds through the close is the market repricing the entire forward curve, not trading the quarter. The mechanism is straightforward: three months ago the company told investors FY26 would grow 18% to 20%, and this morning it told them 25% to 27%. A 5.9% revenue raise and a 13.2% EPS raise on a business whose growth had been assumed to be decelerating is a change in trajectory, and trajectory is what a growth multiple pays for.

The setup amplified it. Even after recovering 23.3% over the prior 30 days, the shares entered the print up only 5.7% year-to-date against the index at 7.6%, so positioning had not yet caught up to the February inflection. The 5.5x volume matches the 5.7x that accompanied the Q3 2025 print, which was the other session where the market had to revise its view of the growth rate rather than the quarter.

The result is that the stock has round-tripped. At $188.73 it sits 5.5% below its 52-week closing high of $199.72, up 65.5% from the $114.01 close ahead of the February print, and 83.9% above the 52-week closing low. The de-rating that created our February upgrade has been fully reversed in three months.

Street Perspective

Debate: Is 32% a New Run-Rate or an AI-Cycle Peak?

Bull view: The acceleration is in the core, not the AI cohort. Growth excluding AI natives has gone 18%, 20%, 23%, mid-20s across four quarters, which is not a cycle, it is a re-rating of the underlying business. Net revenue retention has turned up, current RPO growth has re-accelerated, and new-logo bookings and land sizes both set records and more than doubled. Every leading indicator points the same way.

Bear view: Total growth of 32% requires the AI-native cohort to be growing extremely fast off a base the company will no longer disclose. Training workloads are the newest and most volatile revenue in the business, funded by capital expenditure budgets that are themselves cyclical. Peak second-derivative is exactly when a growth multiple is most dangerous, and the company has removed the disclosure that would let anyone check.

Our take: The bull has the better of the growth argument and the bear has the better of the disclosure argument. Four consecutive core accelerations, with an all-time-record sequential ARR add in a seasonally weak quarter, is not something a cycle manufactures. But withholding the AI-native revenue share for a second quarter means the market is now underwriting a 32% growth rate it cannot decompose, and that is a legitimate reason to demand a lower multiple, not a higher one.

Debate: Does the Training Opportunity Change the TAM?

Bull view: Management said explicitly that training was not a market a year ago and is one now, and backed it with seven-figure and eight-figure landings at two of the largest technology companies in the world. GPU monitoring shipped to general availability in the same quarter. This is a new addressable layer arriving on top of a business already growing 32%, and none of it is in consensus models.

Bear view: Training observability attaches to AI capital expenditure, which is the most cyclical spending in technology, and management could not answer what observability attaches at as a percentage of that spend. Meanwhile gross margin fell 124bps sequentially in the quarter these workloads onboarded, which suggests the new revenue may be structurally lower-margin than the base.

Our take: The TAM addition is real and the option was free, since we were paying nothing for it at initiation when management said GPU monitoring generated no significant revenue. The margin question is the right one to ask and nobody asked it. We would treat training as genuine upside with a lower quality-of-earnings weighting than core observability revenue until the gross margin path through it is visible.

Debate: Does the Price Still Work After a 31% Session?

Bull view: At roughly 14.6x enterprise value to the raised FY26 guide, a business accelerating to 32% growth with rising net revenue retention, 80% gross margins and a Rule of 40 near 60 is not expensive. The multiple is below where this stock traded for most of the prior three years and the growth rate is higher.

Bear view: The stock is up 65.5% in three months and sits 5.5% below its all-time closing high. Free cash flow margin has fallen year-over-year for three straight quarters, capitalized software is up 86%, the FY26 EPS guide implies 78.6x earnings, and two disclosures the market used to rely on have gone dark. Paying a full multiple for a business you can no longer fully decompose is how the November peak happened.

Our take: Both are right about different things, which is what a Hold looks like. The business is materially better than when we upgraded and our fair value estimate rises accordingly. But our estimate rose 17% and the price rose 46%, and at $188.73 the remaining gap to fair value is roughly 9%. That is an in-line expected return, and we do not pay up for a growth rate we cannot decompose.

Model & Valuation Framework

Model Update

Valuation

At the May 7 close of $188.73 and roughly 355M shares outstanding, market capitalization is approximately $67.0B. Netting $4.8B of cash and marketable securities and adding back the $1.0B of 2029 convertible notes gives an enterprise value near $63.2B.

Fair value framework: Applying a 14x to 18x multiple to our FY26 revenue estimate of $4.46B produces an enterprise value range of $62.4B to $80.3B, or $187 to $237 per share after adjusting for net cash and the converts. Our central case of 15.5x gives $205.

Valuation impact: Our 12-month fair value estimate rises to $205 from $175. The increase has two components: our FY26 revenue estimate rose 3.2% to $4.46B from $4.32B on the guidance raise, and we widened the multiple range from 12-to-15x to 14-to-18x because the business demonstrably re-accelerated, net revenue retention turned up, and the agentic-disruption debate that suppressed the prior range has been answered twice with hyperscaler and frontier-lab landings. The estimate rose 17%; the price rose 46%. At $188.73 the remaining gap is roughly 9%.

Thesis Scorecard Post-Earnings

Scored against the standing thesis carried since initiation, not a fresh set of pillars.

Overall: Thesis strengthened on the business and weakened on the entry point. Bull #4 was upgraded on the first hard revenue evidence the AI product line has produced, and Bull #1 confirmed for a fourth time. But the two bear points that have driven every rating decision in this coverage, cash conversion and valuation, both deteriorated, and the disclosure narrowing persisted.

Action: Hold. Take the gain from the February upgrade and wait for either a pullback or the disclosures that would justify paying a full multiple.

Bottom Line: Downgrading to Hold

This is the best quarter Datadog has reported in the period we have covered it. Revenue crossed $1B for the first time, growth accelerated to 32%, the core cohort excluding AI natives accelerated for a fourth consecutive quarter into the mid-20s, sequential ARR added was an all-time record, and new-logo bookings and average land size both set records and more than doubled year-over-year. Net revenue retention rose for the first time since we initiated. Management raised full-year revenue guidance by $240M against a quarterly beat of $50M, which means roughly 79% of the raise is forward-looking.

Two things that were open questions at initiation closed favourably. GPU monitoring, which management described in November as generating no significant revenue, reached general availability and anchored seven-figure and eight-figure landings with the AI research divisions of two of the world's largest technology companies. And the build-versus-buy debate that drove the winter de-rating has now been answered twice: a frontier AI lab last quarter, hyperscalers this quarter. Those are the two most capable possible buyers, and both bought.

The problem is entirely the price. We upgraded on February 11 at $129.67 because the shares had de-rated 32% while the business improved, and we put fair value at $175. The stock closed this print at $188.73, up 31.3% in a single session and 45.5% since that upgrade, 5.5% below its all-time closing high. Our fair value estimate rises to $205 on genuinely better numbers and a wider multiple range, but the estimate moved 17% and the price moved 46%. The remaining gap is roughly 9%, which is an in-line expected return.

Two things also keep us from stretching. Free cash flow margin has now fallen year-over-year for three consecutive quarters, to 28.7%, with capitalized software up 86% and no free cash flow guidance offered. And for a second quarter the company withheld both the AI-native revenue share and security ARR growth, so a 32% growth rate is being underwritten by a market that cannot decompose it. Neither is disqualifying; both argue for a lower multiple rather than a higher one.

We remain constructive on the business and expect to own it again at a better price. The February upgrade did its job. We downgrade to Hold from Outperform at $188.73 against a $205 fair value estimate.