The quarter, in numbers
Datadog reported third-quarter 2026 results for the period ended September 30, with revenue of 1.2 billion dollars, up 35 percent from roughly 886 million dollars a year earlier. Non-GAAP operating income reached 334 million dollars, a 28 percent margin, while GAAP operating income came in at 75 million dollars, a 6 percent margin, the usual wide gap between GAAP and non-GAAP profitability that stock-based compensation creates at a growth-stage infrastructure software company. Operating cash flow was 290 million dollars and free cash flow was 251 million dollars, with cash and marketable securities totaling 5.2 billion dollars on the balance sheet.
Customers paying 100,000 dollars or more in annual recurring revenue grew 20 percent year over year to roughly 4,860, up from about 4,060, a growth rate that, while solid, is slower than Datadog's overall revenue growth, meaning expansion within existing large accounts is contributing more to the total than new large-customer acquisition alone. That pattern, existing customers spending more rather than purely new logos driving growth, is generally read as a healthier signal of product stickiness than new-customer count on its own.
From monitoring AI to building AI
The more consequential news in the release is Datadog's launch of Scout 1.0, described as its first specialized AI model, currently in preview and powering a feature called Bits Investigation. This marks a structural shift in what an observability vendor actually is: Datadog has spent most of its history helping enterprises monitor and troubleshoot other companies' software and infrastructure, including, increasingly, AI systems those customers deploy. Building and shipping its own specialized model moves Datadog from purely an observer of AI workloads to a company developing AI capability of its own, embedded directly into its core product.
CEO Olivier Pomel's comments on the earnings call reinforced this direction, noting the company is developing specialized AI models through its own AI Research team specifically to help customers manage growing complexity, rather than relying solely on general-purpose foundation models licensed from outside labs. That is a meaningful bet: building and maintaining proprietary models is expensive and carries real opportunity cost against simply integrating a leading third-party model, but it also gives Datadog tighter control over how the model behaves inside its specific observability and incident-response workflows.
The industry recognition, and what it is actually worth
Datadog was named a Leader in the 2026 Forrester Wave for AIOps Platforms for the second consecutive year, and Gartner separately named the company its Company to Beat for observability platforms in a 2026 AI Vendor Race report. Analyst recognition of this kind is a lagging indicator more than a leading one, reflecting market position already established rather than predicting future performance, but it is still useful context for procurement teams building a shortlist, since both Forrester and Gartner's methodology weighs real customer deployment data alongside vendor briefings.
The more interesting signal buried in these designations is the category itself: Gartner is explicitly framing competitive dynamics in observability as an AI vendor race now, not a traditional infrastructure monitoring market. That category shift matters for budget owners deciding where observability spend sits organizationally, since a tool being evaluated as part of an AI vendor race is likely to face different procurement scrutiny, different security review requirements, and different budget ownership than a tool still categorized as conventional infrastructure monitoring.
Guidance and what it signals about demand
Datadog guided fourth-quarter revenue to 1.19 to 1.20 billion dollars and full-year 2026 revenue to 4.515 to 4.525 billion dollars, implying continued growth at a pace broadly consistent with the third quarter rather than a deceleration. Full-year non-GAAP operating income guidance of 1.095 to 1.105 billion dollars points to margin expansion continuing alongside revenue growth, a combination that is generally harder to sustain the larger a company gets, making this guidance worth watching against actual results next quarter.
For enterprise technology buyers, consistent guidance at this scale from an observability vendor is itself a demand signal: Datadog's customer base is not pulling back on infrastructure and application monitoring spend even as broader enterprise software budgets face scrutiny elsewhere, which suggests observability specifically is being treated as essential infrastructure rather than discretionary tooling, even during periods when other software categories see budget tightening.
What this means for observability and data strategy
The practical takeaway for data and infrastructure leaders is that observability platforms are no longer a passive layer that reports on what other systems do. Datadog's own specialized model, if it performs as intended inside Bits Investigation, represents a meaningful change in how incident triage and root-cause analysis get done, shifting work from human engineers manually correlating dashboards toward a model doing an initial pass automatically.
Enterprises already running Datadog or evaluating it against alternatives should specifically test Scout's actual investigation output against real incidents in their own environment during the preview period, since the gap between a promising specialized model demo and reliable production performance on your specific infrastructure is exactly where observability tooling has historically disappointed, regardless of how strong the underlying vendor's market position looks on an analyst scorecard.
