The Core Accelerated Again While the Multiple Halved: Upgrading Datadog to Outperform

Results vs. Consensus

Q4 2025 Scorecard

Versus Company Guidance (issued November 6, 2025)

The Q4 revenue beat of 4.1% against the guide ceiling is almost exactly the 4.1% beat delivered in Q3. This is now a settled pattern rather than a surprise, and it is the single most important input to reading the FY26 guide below.

Year-Over-Year Comparisons

The line that changed most versus our Q3 read is stock-based compensation. It grew 29.6% against 29.2% revenue growth, so SBC as a share of revenue was flat at 21.5%. In Q3 that gap was 13 points wide and the ratio deteriorated 204bps. One quarter does not fix the FY25 picture, where SBC still grew 31.6% against 27.7% revenue growth, but the Q4 run-rate is materially better behaved.

Quarter-Over-Quarter Comparisons

Fiscal Year 2025 Versus Fiscal Year 2024

The four reported quarters of FY2025 sum to $3,427.2M, matching the reported full-year figure exactly. The FY picture is the clean statement of the year's trade: 27.7% revenue growth and 18.0% free cash flow growth purchased with 271bps of non-GAAP operating margin and a swing from GAAP operating profit to a $44.4M loss.

Quality of Beat

Platform & Cohort Performance

Revenue Cohort Mix

The core cohort

This is the pillar our initiation said would decide whether Datadog is a durable compounder or an AI-cycle derivative, and it has now posted three straight quarters of higher growth: 18%, then 20%, then 23%. Management said the trend continued into January, which is the second consecutive quarter it has volunteered a post-quarter datapoint rather than leaving the guide to speak for itself.

"Revenue growth accelerated with our broad base of customers, excluding the AI natives to 23% year-over-year, up from 20% in Q3. We saw strong growth across our customer base with broad-based strength across customer size, spending bands and industries. And we have seen this trend of accelerated revenue growth continue in January." — David Obstler, CFO

Assessment: Three sequential acceleration prints in a consumption model, described each time as broad-based across size and spending bands, is no longer a datapoint that can be dismissed as a favourable comparison. In November we declined to extrapolate a single quarter. That caution has been answered, and it is the primary reason for the upgrade.

The AI-native cohort, minus a number

The cohort grew to roughly 650 customers with 19 above $1M in annual spend, and management disclosed for the first time that 14 of the top 20 AI-native companies are customers. What it did not disclose is the cohort's share of revenue, having reported 12% in Q3 and 11% in Q2. Asked directly for the figure, the CFO declined.

Assessment: The qualitative disclosure improved and the quantitative disclosure got worse. On the evidence of this quarter alone the broadening is real, and 14 of the top 20 AI labs as customers is a genuinely strong competitive statement. But withdrawing a metric in the same quarter you introduce four new ones is a choice, and the most likely explanation is that the percentage would have drawn attention to the largest customer's trajectory rather than the cohort's.

Product Line ARR (first-time disclosure)

Assessment: The three-pillar ARR disclosure is the most useful new information in the release and has been absent for years. It establishes that Datadog is three separate billion-dollar franchises rather than one monitoring product with attachments, and it sizes the consolidation opportunity: management said roughly half of customers do not buy all three pillars. Core APM re-accelerating into the mid-30s is the standout, because APM is the pillar most exposed to the argument that AI-generated code needs less human debugging.

Platform Adoption Density

Assessment: The 6-or-more and 8-or-more bands both widened their year-over-year gains versus the Q3 disclosure, and the new 10+ band matches the 8+ band as the fastest-growing cohort in relative terms. The Fortune 500 statistic is the one to keep: half the Fortune 500 are already customers and the median one spends under $500,000, which frames the expansion opportunity inside accounts that have already cleared procurement.

Key Operating Metrics

Assessment: Two things on this table deserve more attention than they will get. Current RPO growth decelerated from the low 50s to about 40% while total RPO growth held at 52% and management noted duration lengthened on a richer multiyear mix. That is a favourable explanation and probably the correct one, but it means a rising share of the commitment book converts to revenue beyond the next twelve months. Separately, security ARR growth went undisclosed after a mid-50s figure in Q3, in a quarter where management described security qualitatively as "actively displacing existing market-leading solutions" without attaching a number to it.

Key Topics & Management Commentary

Overall Management Tone: Confident and noticeably more expansive on disclosure than in prior quarters, volunteering three-pillar ARR figures, a bookings number, a Fortune 500 penetration statistic and a 10-or-more-products band that had never been given before. The defensive energy on the call went almost entirely into the agentic-disruption question, which absorbed the first three analyst turns and drew the most structured answers of the year. The one place management was less than forthcoming was on the two metrics it chose to stop providing, where the response was a flat decline rather than an explanation.

1. A Third Straight Quarter of Higher Core Growth

Revenue excluding AI-native customers grew 23% year-over-year, from 20% in Q3 and 18% in Q2. Total revenue growth of 29.2% is up from 28.4% in Q3. Management described the strength as broad-based across customer size, spending bands, industries and geographies, the same framing used in Q3, and again volunteered that the trend persisted into the following month.

"We saw a continued acceleration of our revenue growth. This acceleration was driven, in large part, by the inflection of our broad-based business outside of the AI-native group of customers we discussed in the past." — Olivier Pomel, Co-Founder and CEO

Assessment: Our initiation graded this pillar Confirmed but explicitly declined to extrapolate from one quarter. Two further quarters of acceleration, with the increment widening from two points to three, converts it from a datapoint into a trend. This is the single most consequential fact in the print.

2. Record Bookings and the Shape of the Deal Book

Q4 bookings reached $1.63B, up 37% year-over-year, the first time the company has disclosed a bookings figure. Inside it were 18 deals above $10M in total contract value, two above $100M, and an eight-figure annualized land described as the largest new-logo deal in company history.

"Our go-to-market teams executed to a record $1.63 billion in bookings, up 37% year-over-year. This included some of the largest deals we have ever made. We signed 18 deals over $10 million in TCV this quarter, of which 2 were over $100 million and 1 was an 8-figure land with a leading AI model company." — Olivier Pomel, Co-Founder and CEO

Assessment: Bookings growing 37% against 29% revenue growth is the forward indicator that matters most, and disclosing it for the first time in a quarter when the stock has de-rated 32% is not a coincidence. The deal-size distribution is the more durable signal: 18 deals above $10M TCV is an enterprise motion operating at a scale this company did not have two years ago.

3. What the FY26 Guide Actually Assumes About the Largest Customer

Management guided FY2026 revenue to $4.06B to $4.10B, characterised on the call as 18% to 20% growth, and then added a second constraint: the guide models the business excluding the largest customer growing at least 20%. Pressed on what that implies for the largest customer itself, the CFO confirmed the arithmetic runs the way it appears to.

"We noted that with the guidance being 18% to 20% and the non-AI or heavily diversified business being 20% plus, that would imply that the growth rate of that core business assumed in the guidance is higher than the growth rate of the large customer. It doesn't mean the large customer is growing any which way. It's just that in our consumption model, we essentially don't control that. And so we took a very conservative assumption there." — David Obstler, CFO

The two statements together bound the largest customer's modelled growth. Working from the FY26 midpoint of $4.08B and a FY25 base of $3,427.2M, the arithmetic below shows the maximum growth the guide leaves for that customer at various assumed sizes. This is our derivation from management's two disclosures, not a company figure.

Assessment: Management has effectively guided the largest customer to roughly flat, and said so in a way that only shows up if you do the arithmetic. Read correctly this is a positive: the FY26 guide asks nothing of the account whose repricing we flagged as the largest unquantified risk at initiation, so the risk is now absorbed into the guide rather than sitting outside it. Read carelessly it is a headline waiting to happen. Investors should be clear which one they are holding.

4. The FY26 Guide Against the FY25 Precedent

The FY26 opening guide of 18% to 20% growth is structurally identical to the FY25 opening guide. On February 13, 2025 the company guided FY25 revenue to $3.175B to $3.195B, or 18% to 19% growth. It delivered $3,427.2M, or 27.7%, exceeding the initial guide midpoint by 7.6%.

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

Assessment: The FY26 guide should be read as the same instrument as the FY25 guide, not as a growth forecast. Applying the FY25 beat mechanically to the FY26 midpoint produces roughly $4.39B, or 28% growth. We model something less aggressive because the ex-AI acceleration is now visible to management when setting the guide rather than hidden from it, but the direction of the error is not in doubt.

5. The AI Labs Are Buying Observability, Not Building It

The largest new-logo deal in company history was an eight-figure annualized land with a leading AI foundational model company that was previously running a fragmented stack of more than five open-source, commercial, hyperscaler and in-house tools. Asked why a company with that much engineering talent would buy rather than build, management gave the same answer it gives about any customer.

"Every customer we land has some -- has had some at homegrown. They have some open source. They might still run some open source, like that's typically what we see everywhere. The -- it's cheaper to do it yourself is usually not the case." — Olivier Pomel, Co-Founder and CEO

Assessment: This is the most commercially specific rebuttal available to the thesis that de-rated the stock. If frontier AI labs, the organisations with the most capable engineering staff and the strongest incentive to build, are instead consolidating onto a commercial observability platform at eight-figure annual scale, the argument that AI makes observability self-service is harder to sustain. It is one deal, but it is the right deal.

6. The Agentic Disruption Question and the In-Stream Answer

The first three analyst turns all probed whether large language models erode the observability category, either by letting customers build homegrown solutions or by absorbing anomaly detection directly. Management conceded the premise on capability and then drew a technical distinction that is the closest thing to a structural moat argument the company has articulated.

"Where the world is going is you're going to have many more changes, many more things. You cannot actually afford to have incidents to look at for everything that's happening in your system. So you need to be proactive. You'll need to run analysis in stream as all the data flows through, you'll need to run detection and resolution before you actually have outages materialize. And for that, you'll need to be embedded into the data plane, which is what we run." — Olivier Pomel, Co-Founder and CEO

Management added that the data volumes involved are several orders of magnitude larger than what is typically fed to a general-purpose model, and was explicit that the product is not there yet.

Assessment: The argument is coherent and the honesty about not having built it yet is credible. The distinction between post-hoc summarization, which a general model does well, and in-stream real-time judgment at millions of events per second, which requires ownership of the data plane, is the right axis on which to defend the category. It is also a promise about a product that does not exist, and the market is entitled to discount it until it does.

7. Bits AI SRE Reached General Availability

The agent product that had no monetization answer in Q3 went generally available in December. More than 2,000 trial and paying customers ran investigations in the past month. Datadog MCP server tool calls grew elevenfold from Q3 to Q4, LLM Observability passed 1,000 customers with usage up tenfold over six months, and roughly 5,500 customers now send data through at least one AI integration.

"We launched Bits AI SRE Agent for general availability in December to accelerate root cause analysis and incident response. Over 2,000 trial and paying customers have run investigations in the past month, which indicates significant interest and shows great outcomes with Bits AI SRE." — Olivier Pomel, Co-Founder and CEO

Assessment: Adoption metrics moved from anecdote to counts, which is progress, but the phrase "trial and paying" does the heavy lifting and no split was given. There is still no revenue figure and no packaging disclosure for a product that shipped two months ago. This pillar stays unproven.

8. Margin Guidance Steps Down Again, Deliberately

FY26 non-GAAP operating income is guided to $840M to $880M on revenue of $4.06B to $4.10B, an implied margin near 21% against 22.4% delivered in FY25 and 25.1% in FY24. Q1 is guided to a 21% margin against 24.1% just reported. Asked where margin ultimately settles, the CFO described the sequencing rather than a target.

"we try to plan with more conservative revenues, understanding that if the revenues exceed above the targets that we give, it's difficult in the short term to invest incrementally. So what we're trying to do is invest first in the revenue growth and then layer in additional investment as we see -- if we see excess of target." — David Obstler, CFO

Assessment: This is internally consistent with the guidance philosophy and with what actually happened in FY25, where the initial 20.9% implied margin became 22.4% delivered. The guided margin step-down is therefore an artifact of the conservative revenue guide, not a signal of cost pressure. The genuine issue is that FY25 came in below FY24 and FY26 is guided lower again, which is a choice management keeps making and has not yet been asked to defend on a multi-year basis.

9. FY26 Earnings Growth Is Guided at Almost Nothing, and the Reason Is Below the Line

FY26 non-GAAP EPS is guided to $2.08 to $2.16 against $2.05 delivered, a midpoint increase of 3.4%, despite operating income guided up roughly 12% at the midpoint. Three below-the-line items explain the gap: net interest and other income guided to approximately $140M against $171.4M in FY25, cash taxes rising to $30M to $40M from $10M to $20M, and the diluted share count rising to approximately 372M from 363.5M.

Assessment: The headline EPS guide will read as a disappointment to anyone anchoring on it, and it is almost entirely a falling-rates and share-count story rather than an operating story. Anyone modelling this business on EPS growth is measuring the wrong thing; the operating income and free cash flow lines are where the FY26 investment case lives.

10. Cash Conversion Repaired

Q4 free cash flow of $291.0M at a 30.5% margin recovered 637bps sequentially from Q3's 24.2%, and operating cash flow grew 23.3% year-over-year. FY25 free cash flow reached $914.7M on a 26.7% margin. Capital expenditures plus capitalized software are guided to 4% to 5% of revenue in FY26, up from 4% in FY25.

Assessment: The Q3 deterioration we flagged as an emerging bear point was substantially working capital and has largely reversed, which is the ordinary resolution. What has not reversed is the capitalized software line, which rose to $27.2M in Q4 and is guided structurally higher for FY26. Free cash flow remains heavily supported by stock-based compensation, which at $750.7M for FY25 is 82% of reported free cash flow.

11. The Consolidation Motion Against Legacy Logging

Management disclosed roughly 100 deals in 2025 that displaced a large legacy logging vendor, worth tens of millions of dollars of new revenue, and noted Flex Logs is approaching $100M in ARR. Several of the named Q4 wins involved replacing between five and thirty separate tools.

"During 2025, we saw an increasing demand to replace a large legacy vendor with takeout in nearly 100 deals for tens of millions of dollars of new revenue." — Olivier Pomel, Co-Founder and CEO

Assessment: Consolidation revenue is the highest-quality growth available to a platform vendor because it is funded by a budget line that already exists and it raises switching costs on the way in. That this is happening at the same time as the AI-disruption narrative peaked is the practical counterpoint: enterprises are actively reducing their observability vendor count, not fragmenting it toward agent-built alternatives.

Guidance & Outlook

Implied Q-over-Q ramp: The Q1 2026 midpoint of $956M requires just 0.3% sequential growth off Q4's $953.2M. For comparison, Q1 2025 grew 3.2% sequentially off Q4 2024. Extending the FY26 guide across the year at an even sequential rate, the midpoint implies fourth-quarter 2026 revenue near $1.09B, or roughly 14% year-over-year growth against the 29% just reported. That is the deceleration curve embedded in the guide, and it is steeper than anything in the company's recent operating history.

Street at: The FY26 revenue guide of $4.06B to $4.10B, together with the Q4 beat, is what produced the 13.7% reaction. The FY26 EPS guide of $2.08 to $2.16 is the weaker half of the package for the reasons set out above.

Guidance style: Unchanged and now well documented. The initial FY25 guide called for 18% to 19% growth and the year delivered 27.7%, beating the initial midpoint by 7.6%. Q3 and Q4 each beat their quarterly guide ceiling by 4.1%. The FY26 guide opens from the same posture.

Other guidance notes: Net interest and other income of approximately $140M for FY26, cash taxes of $30M to $40M, a 21% non-GAAP tax rate maintained for 2026 and beyond, and capital expenditures plus capitalized software of 4% to 5% of revenue.

Analyst Q&A Highlights

Whether the Observability Category Survives the Agentic Era

The call opened on the question that has driven the stock since November: whether agentic frameworks let customers build homegrown observability and collapse the category. Management reframed the argument around where value migrates when code becomes cheap to write.

Q: "just in terms of like what this means for observability as a category, defensibility of it in terms of can customers use these tools to build homegrown solutions for observability? So just give your latest comments on defensibility of the category, and how Datadog may potentially have to evolve in this new sort of a agentic era?"
— Sanjit Singh, Morgan Stanley

A: "you move a lot of the value from the act of writing the code, which now you actually don't do yourself anymore to validating, testing, making sure it works in production, making sure it's safe, making sure it interacts well with the rest of the world, with end users, make sure it does what it's supposed to do for the business, which is what we do with observability."
— Olivier Pomel, Co-Founder and CEO

Assessment: The argument is that agentic development increases the ratio of code in production to human understanding of that code, which is the demand driver for observability rather than a threat to it. It is a good argument and it is consistent with the 23% core acceleration actually printed. It is not proof, and management has an obvious interest in it, but the empirical evidence in this quarter points the same way as the rhetoric.

Where the Line Sits Between a Frontier Model and an Observability Platform

A more precise version of the same challenge asked what stops a general-purpose model from absorbing anomaly detection outright. Management conceded that models are excellent at analysing large data sets handed to them, then located the defensible ground somewhere else entirely.

Q: "How do you think about the limiting factor to using LLMs as an anomaly detection tool that could potentially take share from observability over time in the category? And how do you think about the moat that Datadog has that offers customers a better solution relative to where the road map in LLMs can go long term?"
— Gabriela Borges, Goldman Sachs

A: "So if you feed a lot of data to an LLM and ask for an analysis, you're very likely to get something that is very good and that is going to get even better. So when you think of what we have that is fundamentally our moat here, there's 2 parts. One is how we are able to assemble that contact, so we can feed it into those intelligence engines."
— Olivier Pomel, Co-Founder and CEO

Assessment: Conceding that models will keep getting better at analysis, and relocating the moat to context assembly plus in-stream real-time processing at the data plane, is a more durable position than claiming analytical superiority. The candid admission that the in-stream product is not built yet is what makes the rest of the answer believable, and it also means this defence is a roadmap rather than a shipped capability.

How Much Conservatism Sits Inside the FY26 Guide

The most consequential exchange of the call asked management to decompose the FY26 guide, and specifically what it assumes for the largest customer. The answer confirmed that the guide models the core growing faster than the largest account.

Q: "So maybe you could talk about the level of conservatism that you've built into the guide for the year? And also, you've talked about at least 20% growth for the core, excluding the largest customer, but what is it that we should assume for the large customer? And now when you look at the AI cohort, excluding this large customer, are there any concentrations evolving over there given your strong success there?"
— Ittai Kidron, Oppenheimer & Co.

A: "for the overall business, which is quite diversified, we talked about diversification by industry, by geography, by SMB, mid-market and enterprise, we took the same approach. We noted that with the guidance being 18% to 20% and the non-AI or heavily diversified business being 20% plus, that would imply that the growth rate of that core business assumed in the guidance is higher than the growth rate of the large customer."
— David Obstler, CFO

Assessment: This retires the largest open question from our initiation. The concentration risk we flagged in November is now inside the guide rather than outside it, which converts an unquantified overhang into a modelled assumption. The remaining exposure is that the account shrinks by more than the guide allows, which would still be a headline even though the arithmetic already contemplates roughly flat.

The AI Cohort Revenue Percentage That Was Not Given

Immediately after the guidance exchange came a one-line request for the metric disclosed in each of the prior two quarters. It was declined without elaboration, and no analyst followed up.

Q: "And can you give us the percent of revenue of the AI cohort this quarter?"
— Ittai Kidron, Oppenheimer & Co.

A: "We didn't -- have not put it in there."
— David Obstler, CFO

Assessment: A five-word answer to a metric given at 11% and 12% in the two prior quarters. Management had signalled in Q3 that it expected the metric to lose relevance over time, so this is not unheralded, but retiring it in the same quarter the guide embeds a flat largest customer is unfortunate sequencing. The disclosure trend elsewhere in this release was strongly positive, which makes the two withdrawals stand out rather than blend in.

Whether AI Customer Concentration Can Be Diversified Away

A question on whether the emergence of credible competitors to the current leading model providers creates a path to diversifying the AI cohort drew an unusually direct statement of intent about the company's own structure.

Q: "I'm wondering, if we step back, do you see an opportunity to diversify that AI customer concentration, whether sometimes it might be a direct customer relationship there? Or it could be some of the products like Claude Code being adopted globally, just kind of creating more surface area to drive business to Datadog."
— Mark Murphy, JPMorgan

A: "we've never been -- we're not built as a business to be concentrated on a couple of customers. That's not how we've become successful. That's probably not how we'll be successful in the long term. So yes, I mean we -- at the end of the day, it should be irrational for customers -- for all customers in the AI cohort not to use our product."
— Olivier Pomel, Co-Founder and CEO

Assessment: The supporting evidence is better than the rhetoric: 14 of the top 20 AI-native companies are already customers, 19 now spend above $1M, and the largest new logo of the quarter was a frontier model company. Concentration is falling in shape even without a disclosed percentage. The answer also implicitly concedes that concentration exists today, which management has never said quite so plainly.

Whether the Agent Products Can Capture Their Own Value

With Bits AI SRE now generally available, the pricing question from last quarter became a value-capture question, framed against both startup competition and the frontier labs. Management answered on selling method rather than packaging, then extended into where it thinks the durable advantage sits.

Q: "How does Datadog really capture this value and protect it for the business?"
— Peter Weed, Bernstein Research

A: "if the alternative is in 5 minutes, you have the answer, you only get 3 people looking that are the right folks and you have a fix within 10 minutes, shorter impact on the customer, many, many, many less folks internally involved, lower cost. So it's fairly easy to make that case."
— Olivier Pomel, Co-Founder and CEO

Assessment: A clear articulation of the value proposition and still not an answer on pricing, one quarter after general availability. The follow-on argument, that the durable advantage is in-stream analysis at millions of judgments per second rather than post-hoc summarization, is the same moat claim made elsewhere on the call and rests on unshipped product. This remains the weakest-evidenced pillar in the thesis.

Where Operating Margin Ultimately Settles

With the FY26 margin guided near 21% against a mid-20s level achieved two years ago, the question asked for a destination rather than a year. The answer described a process instead, and in doing so explained why the guided margin is systematically understated.

Q: "I think many gravitate back to that mid-20% margin you put up a couple of years ago. And I know the last couple of years, including the guide are looking at low 20%. Can you talk to maybe your true north, how you're thinking about that, obviously, growth being #1, but how you're thinking about the framework on the bottom line?"
— Brent Thill, Jefferies

A: "as we've seen over the years in our beat and raise, we've tended to have some of that flow through into the margin line and then re-up again for the next phase of growth."
— David Obstler, CFO

Assessment: The mechanic is confirmed by FY25, where an initially implied margin near 20.9% became 22.4% delivered, so the guided 21% for FY26 should be read as a floor. What the answer does not provide is any multi-year commitment, which means the margin question resolves one year at a time and shareholders have no stated destination to hold management to.

What They're NOT Saying

1. The AI-native share of revenue: Disclosed at 11% in Q2 and 12% in Q3, declined outright this quarter. The most likely reason is that the figure would direct attention at the largest customer's trajectory in the same quarter the guide models it flat.
2. Security ARR growth: Given as mid-50s percent in Q3 and mid-40s in Q2, absent here. Security was discussed at length qualitatively, including "actively displacing existing market-leading solutions," with no growth rate attached.
3. The size of the largest customer: Still unquantified after four quarters of questions. Management will now describe what the guide assumes about its growth but not what it contributes in dollars.
4. Why current RPO growth decelerated: From low 50s to roughly 40% while total RPO held at 52%. Management attributed lengthening duration to multiyear mix but was not asked to reconcile the cRPO deceleration, and did not volunteer it.
5. Bits AI revenue or packaging: Generally available since December with 2,000-plus customers running investigations, and still no split between trial and paying, no price, and no FY26 contribution.
6. FY2026 free cash flow: No guidance given, despite capital expenditures plus capitalized software being guided up to 4% to 5% of revenue from 4%. The FCF line is where the investment case sits and it is the one line without an outlook.
7. Any multi-year margin destination: Non-GAAP operating margin fell from 25.1% in FY24 to 22.4% in FY25 and is guided to roughly 21% for FY26, with no stated endpoint.
8. The path to GAAP profitability: FY25 closed with a $44.4M GAAP operating loss against $54.3M of GAAP operating income in FY24. Not raised by management, not asked by any analyst, for a second consecutive quarter.

Market Reaction

• Pre-print setup: DDOG closed at $114.01 on February 9. The stock was down 16.2% year-to-date against the S&P 500's +1.7%, down 22.5% over trailing twelve months, and down 9.1% over the trailing 30 days. The 52-week closing range going in was $87.00 to $199.72, so the shares entered the print 42.9% below their own 52-week closing high.
• Reaction session (February 10, before the open): Opened at $124.18, an 8.9% gap, traded between $123.31 and $133.44, and closed at $129.67. That is a 13.7% single-session gain, or $15.66 per share.
• Volume: 18.8M shares against a 30-day average of 5.3M, or 3.5 times normal.
• Relative move: The S&P 500 fell 0.3% on the same session, making the relative move roughly 14 points.

The setup is the story here, not the reaction. From its 52-week closing high the stock fell 42.9% into this print, to $114.01, without a single negative earnings revision from the company. Contemporaneous coverage attributes the decline to multiple compression driven by sector-wide concern that autonomous agents displace conventional software, amplified by a downgrade with a target cut that alone produced a 7.6% one-day decline, and by a broad rotation out of high-growth technology.

What the print did was supply evidence against precisely that thesis. Core growth excluding AI natives accelerated to 23%, bookings grew 37% to a record, 18 deals cleared $10M in total contract value, and the largest new logo in company history was a frontier AI lab consolidating five-plus tools onto the platform. A 13.7% recovery is a reasonable reaction to a print that contradicts the narrative that caused the decline, and it still leaves the stock 35.1% below its 52-week closing high.

The 3.5x volume is lighter than the 5.7x that accompanied the November move, which suggests repositioning rather than wholesale re-rating. The market has taken the evidence seriously without yet restoring the multiple it removed.

Street Perspective

Debate: Was the 32% De-Rating a Correct Anticipation or a Mistake?

Bull view: The de-rating was a sector-level multiple compression applied indiscriminately, and this print is the counter-evidence. Core growth accelerated for a third quarter, bookings grew 37%, and the AI labs themselves are buying rather than building. Nothing in the fundamentals justified a 32% decline, and the fundamentals are the only thing that eventually settles the argument.

Bear view: Markets discount forward, and observability is structurally exposed to a world where agents write, test and fix code with less human involvement. Current results tell you about workloads provisioned two years ago. The category could look fine for eight more quarters and still be terminally impaired, and paying 12x sales for that risk is not obviously cheap.

Our take: The bear argument is genuinely unfalsifiable on this quarter's data, which is exactly why it produced a 32% de-rating rather than a 10% one. But the specific mechanism proposed, that sophisticated engineering organisations will build rather than buy, has now been tested at the hardest possible case and failed. That does not settle the long-run question, and it does move the burden of proof.

Debate: Is the Flat Largest-Customer Assumption Prudence or Warning?

Bull view: This is a company removing its largest known risk from the forecast. The guide asks nothing of the account, so any contribution is upside, and the core at 20%-plus carries the year on its own. Management has been conservative in exactly this way for years and it has been right to be.

Bear view: Companies do not model their largest customer at flat unless they have a reason. Combine that with withdrawing the AI-cohort revenue percentage in the same quarter and the pattern reads as management managing down an exposure it can see and investors cannot. The contract was repriced in Q3 and the disclosure went dark in Q4.

Our take: The sequencing is unfortunate and the bear reading is not unreasonable, but the arithmetic favours the bull. If the largest customer were deteriorating faster than flat, guiding the core to 20%-plus would be a strange way to hide it, because the core number is the one that gets checked every quarter. Our reading is genuine conservatism, held with less confidence than we would like given the withdrawn disclosure.

Debate: Does 12x Sales Compensate for the Uncertainty?

Bull view: A business compounding revenue at 28% with 81% gross margins, mid-to-high-90s gross retention, 27% free cash flow margins and a record bookings quarter is trading at 12.3x enterprise value to trailing revenue and about 10.3x the FY26 guide, which the company's own history says is too low. That combination has not been available in this name in years.

Bear view: Twelve times sales is only cheap relative to what this stock used to trade at, and the anchor is doing the work. On free cash flow it is still 46x trailing, GAAP operating income is negative, stock-based compensation is 82% of free cash flow, and the FY26 EPS guide implies 3.4% growth. Cheap-versus-history is the weakest form of cheap.

Our take: The bear is right that anchoring on the old multiple proves nothing, and right that the GAAP and SBC criticisms have not been answered. But the FY26 EPS guide is a rates-and-share-count artifact rather than an operating one, and 36x our FY26 free cash flow estimate for a business growing revenue in the mid-20s with this retention profile is a reasonable price rather than a demanding one. We think the risk/reward has genuinely turned.

Model & Valuation Framework

Model Update

Valuation

At the February 10 close of $129.67 and roughly 352M shares outstanding (Q4 weighted-average basic was 350.9M), market capitalization is approximately $45.6B. Netting $4.47B of cash and marketable securities and adding back the $1.0B of 2029 convertible notes gives an enterprise value near $42.2B.

The November column is restated here on today's actual denominators (FY25 revenue of $3.43B, FY25 EPS of $2.05, FY25 free cash flow of $914.7M) so the two columns are like-for-like. Our November note quoted 18.8x and roughly 75x against the FY25 guidance midpoint and our then-estimate of free cash flow, which were the only figures available at the time.

Fair value framework: Applying a 12x to 15x multiple to our FY26 revenue estimate of $4.32B produces an enterprise value range of $51.8B to $64.8B, or $157 to $194 per share after adjusting for net cash and the converts. The midpoint of roughly $175 implies about 35% upside from the $129.67 close.

Valuation impact: Our 12-month fair value estimate moves to $175 from $197. The estimate falls because the FY26 revenue base is now anchored to a guide rather than to our prior extrapolation and because the share count guide steps up 2.3%, while the multiple range we are willing to pay steps down from 14-to-17x to reflect the unresolved agentic-disruption debate. The stock fell considerably more than the estimate did, which is the entire basis of the upgrade.

Thesis Scorecard Post-Earnings

Scored against the standing thesis established at initiation in November, not a fresh set of pillars.

Overall: Thesis strengthened. The pillar that mattered most confirmed for a third time, the two bear points that drove the Hold rating have both been contained, and the largest unquantified risk from the initiation is now absorbed into company guidance. Against that, one bull pillar was downgraded and one new bear point opened, both on disclosure rather than on results.

Action: Buy. The business improved and the price fell by roughly a third.

Bottom Line: Upgrading to Outperform

We initiated coverage of Datadog at Hold in November at $190.82, arguing that an excellent quarter had been fully paid for in a single 23% session and that we would upgrade on a second confirming quarter or a better price, whichever came first. Both arrived together.

The confirming evidence is unambiguous. Core revenue growth excluding AI-native customers rose to 23% from 20%, a third straight quarter of higher growth, and management said the trend continued into January. Bookings reached a record $1.63B, up 37%, with 18 deals above $10M in total contract value and two above $100M. Customers at $1M or more in ARR grew 31% to 603. Three separate product pillars now exceed $1B in ARR, disclosed for the first time, with core APM re-accelerating into the mid-30s. And the largest new logo in company history was a frontier AI lab consolidating more than five tools onto the platform, which is the most direct available answer to the argument that de-rated this stock.

The price came to us as well. The shares fell roughly 32% between mid-November and this print, on multiple compression rather than on any company-driven revision, and closed the reaction session at $129.67, still 35% below the 52-week closing high. Enterprise value is 12.3x trailing revenue against 18.6x when we initiated, measured on the same actual FY25 denominator, and about 36x our FY26 free cash flow estimate.

The upgrade is not unqualified. Management stopped disclosing the AI-native share of revenue and stopped disclosing security ARR growth, both in a quarter that otherwise expanded disclosure considerably, and the FY26 guide quietly models the largest customer at roughly flat. Those facts are compatible with prudence and they are also compatible with something management can see that we cannot. The FY26 non-GAAP operating margin guide of about 21% would be a second consecutive annual decline, with no multi-year destination offered, and the GAAP business posted a $44.4M operating loss for the year.

Weighing it, the evidence in this print runs against the narrative that caused the de-rating, and the de-rating was larger than any reasonable reading of the risk. Our fair value estimate moves to $175 from $197, about 35% above the close. We upgrade to Outperform from Hold.