Datadog Grew 36% as Bits Agents Expanded Operations

On August 6, 2026, Datadog’s second-quarter financial release put revenue at $1.12 billion, up 36% year over year, and free cash flow at $279 million; it also counted approximately 4,720 customers with annual recurring revenue of at least $100,000, up 23%. Revenue is a GAAP measure, while free cash flow is a non-GAAP liquidity measure and should not be treated as operating profit.

The direct product answer is that Bits Investigation, Bits Security Analyst and Bits Code are commercially available. Datadog’s current agent catalog assigns them distinct jobs: investigating operational alerts, triaging security signals and turning production findings into reviewable code fixes.

A contemporaneous independent market recap also recorded $1.12 billion of revenue, 36% growth and roughly 4,720 large customers. The recap corroborates the headline financial figures, although Datadog’s own release remains the primary source for its accounting results and definitions.

Growth gives the agents a substantial distribution base

The important commercial connection is scale, not evidence that agents produced the quarter’s growth. Datadog can introduce agent workflows to customers already using its infrastructure monitoring, application performance, logging, security and incident-management products. That installed context matters because an operational agent needs more than a single alert to investigate a failure credibly.

The August investor presentation places the agents inside a platform serving approximately 33,400 customers as of June 30, 2026; it shows that 58% used four or more products, 37% used six or more and 22% used eight or more. Those adoption rates do not measure demand for Bits. They indicate how much operational data and how many existing workflows Datadog can potentially connect to agent-led investigation or remediation.

An incident investigation may require metrics to establish when performance changed, traces to locate the affected service, logs to expose an error and ownership information to identify the responsible team. Datadog already collects and relates those inputs. The agent layer changes the output from a collection of dashboards and alerts into a proposed explanation or action.

The available agents cover reliability, security and code

Bits Investigation is the reliability agent. It autonomously examines alerts, correlates telemetry and organizational context, surfaces a likely root cause and suggests remediation. Engineers can inspect the reasoning and supporting evidence, so the product automates evidence gathering and hypothesis testing without making the resulting conclusion inherently infallible.

Bits Security Analyst applies the same operating model to security work. It triages Cloud SIEM signals, examines the assets and users involved, recommends a verdict and supplies response guidance. Its commercial role is therefore closer to an automated first-line analyst than to another security dashboard.

Bits Code carries an operational finding into software remediation. Datadog’s DASH product roundup labels Bits Code generally available and says it can use logs, traces, metrics, profiles, runtime variables and security findings to propose fixes grounded in production behavior. It can create work for code review or run on scheduled and telemetry-triggered workflows, but engineers retain control over what is merged.

The same roundup distinguishes released software from products still in Preview. Bits Release, which follows changes from pull request into production, and Bits Testing Agent, which generates and maintains synthetic tests, were not presented as generally available. They therefore should not be counted with the commercially released agents when assessing what enterprise buyers can adopt on standard availability terms.

Autonomy now extends beyond investigation

Traditional observability ends with visibility: collect signals, display system behavior and notify an operator. Datadog’s agent portfolio adds successive stages—assemble evidence, investigate likely causes, explain a conclusion and, in selected workflows, prepare or trigger remediation. The result can be a root-cause analysis, a security verdict or a proposed code change rather than another notification for a human to investigate from scratch.

That is the defensible basis for describing the direction as autonomous operations. It does not mean people have been removed from production decisions. Investigation can begin without a prompt, but permissions, repository access, approval policies and deployment controls still determine what an agent may inspect and whether a proposed action reaches a live system.

The distinction is especially important for Bits Code. Generating a pull request is an operational action, yet it remains separated from approving and deploying that change. Datadog is automating more of the path between detection and resolution while leaving customers responsible for the governance boundary.

The results do not show how much agents contribute

Datadog has not disclosed separate revenue, ARR, customer adoption or retention figures for Bits Investigation, Bits Security Analyst or Bits Code. The company-wide growth rate therefore cannot be attributed to the agent products. The evidence supports a narrower conclusion: financial growth and large-customer expansion provide distribution capacity while the released agents extend the platform from observation into investigation and governed remediation.

Pricing contribution and execution rates also remain unclear. Datadog has not quantified how often customers accept an agent’s conclusion, merge a proposed fix or permit an automated remediation to run. Those disclosures would be needed to judge whether autonomous operations are becoming a material revenue driver rather than an expanding product capability.

As of the released quarter, the financial and product stories are related but not interchangeable. Datadog has a growing, cash-generative platform and three commercially available agents spanning reliability, security and development; what remains unknown is how much those agents are contributing to growth and how much operational authority customers are prepared to delegate.

Datadog Grew 36% as Observability Turns Into Autonomous Operations

Growth gives the agents a substantial distribution base

The available agents cover reliability, security and code

Autonomy now extends beyond investigation

The results do not show how much agents contribute

Subscribe to our newsletter