Why the Street Is Wrong About Datadog's Observability Multiple

Datadog operates the observability and security platform that cloud-native enterprises use to monitor the health of their applications. The product stack includes infrastructure monitoring, application performance monitoring, log management, security monitoring, synthetic testing, and more recently AI-specific observability. The gross revenue retention rates have been above 95%, and the net dollar retention rate (NRR) has stayed above 120% through 2024-2025.

The customer count of approximately 29,000 enterprises at end of 2025 compares to roughly 22,000 two years earlier. The customers with annual run-rate revenue above $100,000 numbered approximately 3,600 at end of 2025, up from 2,800. The customer with ARR above $1 million numbered roughly 425, up from 320. These cohort metrics describe a platform that is still widening and deepening its customer base at an attractive pace.

The competitive set has consolidated around three major platforms: Datadog, New Relic (taken private in 2023), and Dynatrace. Cisco's AppDynamics continues to serve large legacy accounts but has lost share to Datadog in cloud-native workloads. Observability is one of the few enterprise software categories where the dominant vendor has extended its lead over the last two years, not lost ground.

The 2025 GAAP operating income came in at negative $44 million against positive $54 million in 2024. The reported swing is driven almost entirely by stock-based compensation (SBC) expense that accelerated through the year as a new grant cycle landed and prior-year vesting schedules overlapped. SBC as a percentage of revenue moved from approximately 24% in 2024 to approximately 28% in 2025.

Stock-based compensation is a genuine economic cost (shares transferred to employees are dilutive and represent real compensation). But it is not a cash economic cost in the period it is recognised. The free cash flow line captures the actual operating cash generation of the business. At $1.0 billion of FCF on $3.43 billion of revenue, the underlying cash margin is approximately 29%, up from 31% in 2024 on a like-for-like basis after normalising for working capital timing. The business is generating cash at a healthy pace.

The market's reaction has been to treat the GAAP swing as a fundamental deterioration. The Signals Desk reads it as an accounting-driven optical effect. The distinction matters because it implies very different forward fair values.

On cash flow terms, Datadog at $44 billion market cap and $1.0 billion trailing FCF yields 2.3% cash flow yield with 20%+ annual growth. That is a reasonable valuation for a 20%+ compounder. On GAAP earnings terms, the stock looks expensive at a 59x forward multiple. The contradiction is the mispricing. Cash flow is the right lens.

Build the forward earnings model around cash flow. FY2026 revenue consensus sits at $4.3 billion (25% growth). Assume gross margin holds at 81% and operating expense grows at 18% off a $2.2 billion base. That gives FY2026 operating income (GAAP) of approximately $350 million, or roughly 8% operating margin, depending on the SBC trajectory.

Adjusted operating income (excluding SBC) lands at approximately $1.3 billion, a 30% margin. Free cash flow, based on the historical ratio, lands at $1.2 billion. At a 30x forward FCF multiple (reasonable for a 25%+ compounder at scale), implied market cap is $36 billion, which is lower than current $44 billion. But the right multiple to apply depends on the quality of the FCF and the compounding runway.

Observability as a category is roughly 20% penetrated against a long-term total addressable market of $60-80 billion by 2030. Datadog has roughly 20% share within that penetrated base. The compounding runway is 10+ years. A 30x forward FCF multiple is on the low end of what comparable-growth compounders have earned in similar cycle positioning. A 35-40x multiple is defensible, which takes fair value toward $50-55 billion market cap, or $180 per share.

The carry through the waiting period is modest: no dividend, modest buyback executed opportunistically, and capital structure discipline that avoids dilutive raises. The economic return depends on the multiple re-rate, which depends on the Q1 FY2026 print and the management commentary on SBC normalisation.

Return on invested capital through cycles for Datadog has been hampered by the SBC charge but the cash-basis ROIC is in the 35-45% range. That is the profile of a structurally high-quality business that the market is currently mispricing on an accounting metric rather than on underlying economics.

One additional framing on the multiple debate. The median forward FCF multiple across the enterprise software compounders (Microsoft, Salesforce, ServiceNow, Adobe) over the last decade has been approximately 32x. Datadog has historically traded in line with that median during normal operating periods. The current multiple of roughly 28-29x trailing FCF sits below the peer median, despite Datadog's faster revenue growth and comparable cash margin profile. The reversion to median is itself worth 10-15% of re-rate.

The Signals Desk reads cohort data across enterprise software and Datadog's pattern has been consistently strong. Net dollar retention has held above 120% for multiple years, which places Datadog in the top decile of the SaaS peer set. Gross retention has held above 95%. Customer count at the $1M+ ARR tier grew 33% year over year in 2025, accelerating from 25% in 2024.

These cohort metrics are the forward leading indicators for revenue growth. The current revenue growth rate of 28% is trailing the cohort-implied forward rate; a reacceleration of reported revenue growth in FY2026 toward 30%+ is plausible based on the cohort data. The Street models are currently sitting at 25% for FY2026. The Signals Desk sees upside against those models.

Historically, when enterprise software leaders have produced cohort metrics at this level with a depressed multiple, the forward 12-month total return has averaged 25-40%. The pattern has played out repeatedly: Salesforce in 2008, ServiceNow in 2022, CrowdStrike in early 2024. Each case featured a market punishment for a short-term accounting or margin signal, followed by a revenue re-acceleration that reset the multiple.

The historical base rate supports the contrarian view. The immediate catalyst is the Q1 FY2026 print, which lands in early May and will confirm whether the cohort data is translating into reported revenue as the Signals Desk expects.

The cohort-based projection methodology produces a forward revenue estimate that sits meaningfully above consensus. Taking the Q4 2025 net dollar retention of 121% and applying it to the run-rate base while adjusting for new customer additions suggests a FY2026 revenue of $4.4-4.5 billion, against a consensus of $4.3 billion. That is a 3-5% upside to the reported revenue estimate, which typically translates to a 5-8% upside to the share price when it becomes visible in actual reporting.

One dynamic that the current market pricing has discounted heavily is Datadog's positioning in AI-specific observability. As enterprises deploy large language model applications into production, they need monitoring tooling for model performance, token-usage economics, hallucination rates, latency across inference endpoints, and data pipeline integrity. Datadog has launched dedicated AI observability products over the last 12-18 months and has begun disclosing customer uptake metrics.

The AI observability sub-category is still very early. Revenue contribution to Datadog in 2025 was modest, probably under $100 million on a $3.43 billion base. The forward growth rate, however, is exceptional; AI observability revenue is tracking at 3-5x the corporate average growth rate. On a five-year horizon, AI observability could represent 15-25% of Datadog's revenue mix, at gross margins above the corporate average.

The competitive landscape in AI observability is less consolidated than in traditional observability. Startups like LangSmith, Helicone and Arize are building specialised tooling. The advantage for Datadog is the enterprise sales motion, the existing customer relationships, and the integration with the broader observability stack. Customers would rather buy AI observability from their existing observability vendor than add a new platform, provided the feature set is competitive.

The Signals Desk reads AI observability as a free option in the Datadog thesis that is not priced in the current multiple. If the sub-category plays out at the pace the early adoption data suggests, it adds another 3-5 years of compounding runway to an already attractive story.