tapebrief
DDOG · Q3 2025

DDOG · Q4 2025 Earnings

Cautious

Datadog

Reported February 10, 2026

30-second summary

Q4 revenue of $953M (+29.2% YoY) beat the high end of the prior guide by $37M and non-GAAP EPS of $0.59 cleared the high end by $0.03; non-GAAP operating margin held at 24% — clean execution against last quarter's bar. The signal worth pricing is the FY26 setup: $4.06–4.10B in revenue is only 18.7–19.7% growth against FY25's 27.6% actual, and FY26 non-GAAP EPS of $2.08–2.16 implies just 1.5–5.4% growth off $2.05 — a sharp narrative shift from a company that spent FY25 raising the bar every quarter. Either the guide is deeply conservative in the Datadog house style, or AI-cohort velocity is normalizing into 2026.

Headline numbers

EPS

Q4 FY2025

$0.59

Revenue

Q4 FY2025

$0.95B

+29.2% YoY

Gross margin

Q4 FY2025

80.4%

Free cash flow

Q4 FY2025

$0.29B

Operating margin

Q4 FY2025

1.0%

Key financials

Q4 FY2025
MetricQ4 FY2025YoYQ3 FY2025QoQ
Revenue$0.95B+29.2%$0.89B+7.6%
EPS$0.59—$0.55+7.3%
Gross margin80.4%—81.0%-60bps
Operating margin1.0%—23.0%-2200bps
Free cash flow$0.29B—$0.21B+36.0%

Guidance

Company beat Q4 FY2025 guidance across revenue and EPS but issued cautious FY2026 guidance with 18.7-19.7% revenue growth (vs. 27.6% actual FY2025), modest 1.5-5.4% EPS growth, and flat-to-conservative messaging despite strong current-quarter execution.

Guidance is issued for both next quarter and the full year. Both may appear below.

Actuals vs prior guidance

MetricPeriodPrior guideActualΔResult
RevenueQ4 FY2025$912 million to $916 million$953 million+$37-41 million above guideBeat
Non-GAAP EPSQ4 FY2025$0.54 to $0.56$0.59+$0.03-0.05 above guideBeat
Non-GAAP Operating IncomeQ4 FY2025$216 million to $220 million$229 million+$9-13 million above guideBeat

New guidance

MetricPeriodGuideYoY
RevenueFY2026$4.06 billion to $4.10 billion+18.7% to +19.7%
Non-GAAP EPSFY2026$2.08 to $2.16+1.5% to +5.4%
Non-GAAP Operating IncomeFY2026$840 million to $880 million—
RevenueQ1 FY2026$951 million to $961 million—
Non-GAAP EPSQ1 FY2026$0.49 to $0.51—
Non-GAAP Operating IncomeQ1 FY2026$195 million to $205 million—
Revenue Growth (YoY %)FY202618.7% to 19.7%—

Platform metrics

Q4 FY2025
SegmentQ4 FY2025
ARR Customers $1M+603
ARR Customers $100K+4,310
Q4 YoY Revenue Growth29%
FY2025 YoY Revenue Growth28%

Profitability

Q4 FY2025
SegmentQ4 FY2025
Non-GAAP Operating Margin24%
Free Cash Flow Margin31%
Gross Margin80%
Operating Cash Flow$1,050M (FY2025)

Management tone

Q1-25 AI experiments → Q2-25 AI-native cohort named as the accelerant → Q3-25 acceleration broadens to the base business → Q4-25 platform consolidation and AI agents as the operating model, FY26 setup deliberately conservative.

The disclosure framework completed its rotation away from the AI-cohort cut. Two quarters ago management introduced the $1M+ and $100K+ AI-native customer counts; last quarter they telegraphed those would "become less relevant"; this quarter the press release leads with $1M+ ARR customers across the whole base (603, +31%) and $100K+ ARR customers (4,310, +19%). That answers Tapebrief's Q3 watch item — the replacement disclosure is multi-product-driven customer cohort counts, not concentration-specific cuts. Cleaner for management, less granular for investors tracking AI-cohort dependency.

The AI moat argument moved from "we benefit from AI" to "LLMs cannot replicate what we do in-stream." Across the Q&A — Singh, Borges, Coupland — Olivier repeatedly anchored the defensibility thesis on real-time, in-stream analysis at data-plane scale, with specialized embedded models running on "orders of magnitude more data than LLM post-hoc analysis." That is a sharper articulation of category defensibility than the "AI creates more complexity to monitor" framing of Q2 and Q3, and it reads as a direct response to the rising analyst pressure on whether observability gets disintermediated by generalist AI. The eight-figure annualized deal with a leading AI model company — explicitly framed as consolidation from open-source, commercial, hyperscaler, and in-house tools — is the proof point management is leading with.

FY26 guide tone is meaningfully more conservative than the FY25 trajectory taught investors to expect. Datadog spent FY25 raising the revenue growth bar from initial 19% guidance to 27.6% actual — three consecutive raises. The initial FY26 print of 18.7–19.7% growth, paired with 1.5–5.4% EPS growth, fits the Datadog house pattern of starting low and raising through the year, but the EPS-growth compression specifically is harder to dismiss as pure conservatism. The Q4 operating margin landing exactly at the 24% guide (vs Q3's 200bps beat) and gross margin slipping 60bps QoQ add to the texture that operating leverage is not flowing as freely heading into FY26.

Competitive posture remained dismissive but specifics thinned. Asked directly by Coupland on competitive share shifts, Olivier described "no material change" and characterized recent M&A targets as "non-winning companies." That is the same answer management has given for four quarters. The lack of fresh competitive specifics is consistent with confidence — but also consistent with a thinning of disclosed evidence as the AI-cohort cut retires.

Recurring themes management leaned on this quarter:

AI-native platform capabilitiesCustomer expansion and land-and-expandPlatform consolidation driving retention

Q&A highlights

Sanjit Singh · Morgan Stanley

How will agentic frameworks and frontier AI models impact observability as a category? Can customers build homegrown solutions? What is Datadog's defensibility?

Olivier explained that agentic AI will accelerate developer productivity and create more complexity requiring greater observability. The key differentiator is Datadog's ability to embed intelligence in the data plane for real-time detection and preemptive resolution, not just post-hoc analysis. LLMs are good at broad analysis but lack the specialized, in-stream capabilities Datadog provides.

More applications and complexity will be created as developers build faster with AIValue shifts from code writing to validation, testing, and ensuring safety in productionObservability needed for understanding and aligning AI agents with expectationsMCP server exposure for agent functionality seeing explosive growth

Remo Linschau · Barclays

What drove the eight-figure AI model company deal to consolidate from open-source and homegrown tools to Datadog? Is the cheap-DIY narrative invalid?

Olivier stated the situation mirrors all customer acquisitions: companies start with homegrown/open-source solutions but switch to Datadog due to ROI. Engineer compensation and velocity are primary cost drivers, not tooling costs. The AI cohort is a high-growth, mission-critical segment with the same consolidation logic as other customers.

Eight-figure annualized deal with leading AI model companyCustomer consolidating from open-source, commercial, hyperscaler, and in-house observability toolsEngineer compensation is larger cost component than toolingVelocity and productivity are primary business drivers

Gabriella Borges · Goldman Sachs

Where is the line between LLM capability and Datadog's domain expertise in observability? Can LLMs be a moat threat for anomaly detection and vulnerability management?

Olivier acknowledged LLMs are improving rapidly but distinguished Datadog's moat in two ways: (1) ability to assemble context and aggregate data to feed into intelligence engines (exposed via MCP); (2) future requirement for proactive, in-stream analysis running on embedded, specialized models at data-plane scale, not just post-hoc summarization. This requires scale and specialization LLM providers cannot match.

LLMs very good at broad data analysis when given large datasetsDatadog's moat: context assembly and data aggregation for intelligenceMCP server exposes functionality for customer recombination with different toolsFuture observability requires in-stream, proactive analysis before outages

Mark Murphy · JP Morgan

Given $500B+ hyperscaler CapEx (growing 40-60%), can Datadog estimate how much is training vs. inferencing and predict when LLM observability revenue will ramp predictably?

Olivier reframed the question beyond pure LLM observability, noting that massive CapEx points to vastly more applications, intelligence, and system complexity across the economy. Direct mapping of CapEx to product mix is difficult and timing uncertain, but the trend clearly benefits Datadog's overall business through increased complexity and scale.

$500B+ annual hyperscaler CapEx growing 40-60%Points to way more applications, intelligence, and system complexityDifficult to directly map CapEx to infrastructure usage 2-4 years forwardOverall business will benefit from increased system complexity and reach

Todd Coupland · CIBC

How is the rise of LLMs impacting competitive share shifts and Datadog's competitive positioning?

Olivier stated competitive dynamics have not materially changed; same competitors, Datadog pulling share from all scale players. Recent M&A dealt with non-winning companies not encountered in deals. For LLM observability, still early and undifferentiated. Datadog's advantage: integrated observability across entire system, as LLMs don't operate in isolation but depend on applications and tools.

No material change in competitive dynamicsM&A targets were non-winning companies with limited market impactLLM observability market still early and undifferentiatedCompetitive advantage: integration with production applications and existing infrastructure

Answers to last quarter's watch list

Whether Q4 revenue lands inside or above the $912–916M guide. Q4 came in at $953M — $37M above the high end, a 4.0–4.5% beat that materially exceeds the "Q3 beat by $35M" precedent. Q4 YoY growth of 29.2% essentially matched Q3's 28.4%; there is no in-quarter deceleration signal. The "imply conservatism" framing held.
Resolved positively
Whether Q4 non-GAAP operating margin clears the 24% guide. Operating margin landed exactly at 24%, with $229M of non-GAAP operating income beating the dollar guide by $9–13M but with the percentage in line. A meaningful change from Q3's 200bps beat — the percentage upside compressed even as the revenue beat persisted, consistent with hiring catching up to revenue.
Continue monitoring
What replaces the "AI cohort" disclosure. The replacement is $1M+ ARR customer count (603, +31% YoY) and $100K+ ARR customer count (4,310, +19% YoY) for the entire customer base. That is meaningful disclosure and answers the question — but it explicitly retires the AI-cohort point-contribution and ex-largest-AI-customer cuts. Cleaner narrative, less concentration visibility.
Resolved positively
New logo land size and bookings on the Q4 call. Without a transcript, the press release does not contain a new-logo land size or bookings disclosure comparable to Q3's "more than doubled YoY." The Q&A focused on AI cohort consolidation deals rather than aggregate new-logo metrics.
Not resolved
Whether platform-adoption tiers (6+ products at 31%, 8+ at 16%) continue stepping up in Q4. Updated tier percentages were not disclosed in the press release excerpt. The +31% YoY growth in $1M+ ARR customers is a positive proxy for multi-product depth, but the specific tier framework that ran in lockstep through FY25 was not refreshed.
Not resolved
First quantified monetization signal on Bits.ai SRE or any AI agent product. No paid customer count, pricing model, or revenue contribution disclosed. Olivier discussed agentic capabilities at the architectural level in Q&A and pointed to MCP server adoption, but pricing/monetization details for Bits AI agents were not provided.
Continue monitoring

What to watch into next quarter

Whether Q1 FY26 revenue clears the $961M high end and prints above $980M. Q4 beat its guide by 4–4.5%; a similar beat on the Q1 guide of $951–961M would imply roughly $993–1,003M. Anything inside the range would mark the first quarter in over a year where Datadog did not materially exceed the high end — and would validate the cautious FY26 setup.

Whether FY26 revenue is raised on the Q1 call. Datadog raised FY25 guidance every quarter; the FY26 starting point of 18.7–19.7% growth has obvious room to move up if AI cohort velocity sustains. A Q1 raise into the low 20s% would frame the initial guide as standard Datadog conservatism; a Q1 reaffirm would suggest the moderation is structural.

Whether Q1 non-GAAP operating margin runs above the guided ~21% implied by $195–205M on $951–961M revenue. Q4's exactly-met operating margin and gross margin slip to 80.4% are the first FY25 datapoints suggesting drop-through is compressing. A Q1 operating margin print at 21% or below would confirm the FY26 EPS guide of 1.5–5.4% growth is not pure conservatism.

Whether the $1M+ ARR customer count maintains +25%+ YoY growth. The new headline disclosure is 603 customers at +31%. A step-down toward +20% in Q1 would suggest the large-deal motion is cooling; sustained +25%+ would validate the replacement metric as a credible AI-cohort proxy.

First quantified monetization datapoint on Bits AI agents. Three quarters of architectural commentary without a paid customer count, ARR contribution, or pricing model. By the Q1 FY26 call, any of those datapoints would mark commercialization; continued silence raises the question of whether agents are a paid product or a sales-cycle enabler.

Whether share count guidance of 372M implies buyback absence in FY26. The 372M FY26 average share count vs ~367M Q4 FY25 weighted average is ~1.4% dilution and is the primary contributor to the EPS-growth-vs-operating-income gap. Any commentary on a buyback authorization on the Q1 call would directly reframe the FY26 EPS optics.

Sources

  1. Datadog Q4 2025 press release (8-K exhibit 99.1, filed with SEC): https://www.sec.gov/Archives/edgar/data/1561550/000162828026006645/ex-991x20251231x8k.htm
  2. Datadog Q4 2025 earnings call Q&A (analyst exchanges; prepared remarks not available at time of brief).

Get the next brief, free.

We publish analyst-grade earnings briefs the same day or morning after every call — headline numbers, segment KPIs, Q&A highlights, and tone analysis. Free during beta.

This is not investment advice.