The business
Datadog collects metrics, traces and logs from a company's servers, containers, databases and applications and turns them into dashboards, alerts and investigations. Engineers install its agent and libraries once and then add products: infrastructure monitoring, application performance monitoring, log management, security, real-user monitoring and observability for AI applications.
Pricing is mostly usage-based. Revenue in FY2025 was $3.4B, up 27.7%.
Moat and switching costs
Instrumentation is the moat: once Datadog's agent and libraries are in every service, its dashboards are how on-call engineers think, and the history is in Datadog. Adding a product is a click; replacing the platform is a project.
Usage pricing cuts both ways. Customers optimise their bills (2022 to 2023 was a hard period), and open-source tooling (OpenTelemetry, Grafana) makes the collection layer more portable over time.
How the technology shows up in the numbers
Gross margin of 80% and free cash flow of $1.0B in FY2025 (about 30% of revenue). Stock pay was $751M, so owner earnings were about $250M and GAAP operating margin was near zero. The share count rose 1.3% in FY2025. Net cash of $3.5B.
Thesis
Every new application, and especially AI applications, needs observability, and Datadog wins most of the cloud-native market. The case rests on growth staying above 20% while stock pay falls as a share of revenue.
We are watching for the owner economics to catch up with the business quality.
Risks
- Customers optimising usage when cloud budgets tighten.
- OpenTelemetry and cheaper open-source back ends commoditising collection.
- Stock pay staying above 20% of revenue.
What would change our mind
- Owner earnings rising toward 15% of revenue (a reason to own).
- Net revenue retention falling well below 110%.
- Large customers moving telemetry to cheaper back ends.
