Datadog CCM Brings Unified AI Cost Visibility Across All Major Providers
Organizations are spending more on AI than ever, and most have no clear picture of where that money actually goes. Billing dashboards differ by provider. Schemas are inconsistent. Attribution is manual, patchy, and slow. Datadog is now addressing all of this directly with a new AI cost management feature inside Datadog CCM.
The new AI Costs feature in Datadog Cloud Cost Management gives FinOps and engineering teams a unified view of AI spend across providers including OpenAI, Anthropic, Amazon Bedrock, Google Gemini, and Vertex AI. Instead of reconciling separate exports and billing APIs, teams can now analyze AI spend alongside existing infrastructure costs, all in one platform.
The feature includes a unified AI cost landing page that aggregates total spend across providers. It surfaces daily trends, provider-level breakdowns, top cost drivers, and anomalies. From there, provider-specific dashboards go deeper, correlating cost data with usage signals like token consumption, model distribution, and request volume. Engineers can spot exactly when a spike in spend connects to a model switch, a traffic surge, or an inefficient usage pattern.
AI cost management with Datadog CCM also solves a longstanding tagging problem. Each provider exposes billing data differently. One might track by model. Another emphasizes workspace-level data. Without normalization, cross-provider queries require separate logic for every source. Datadog maps all provider data to a consistent tag set, covering provider, project, model, token type, token category, and direction, so teams can query across all sources without rewriting their filters.
Attribution is where things get especially powerful. Native provider billing data typically stops at the model or workspace level. There are no API keys or user names in the bill, so individual usage goes untraced. Datadog’s out-of-the-box allocation rules solve this automatically. They analyze SaaS observability metrics already inside CCM and attribute AI spend proportionally to the API keys and users responsible, without requiring extra instrumentation. For OpenAI and Anthropic, where OOTB rules are currently available, teams can further use Tag Pipelines to map user emails to specific teams, services, or business units.
Once attribution is in place, teams can build CCM reports to roll up AI spend by team, service, or project. Leaders can use those reports to drive budgeting decisions. Engineers can use the same data to reduce waste, without needing a new tool or a manual data export.
Because CCM already aggregates infrastructure costs from AWS, Microsoft Azure, Google Cloud, and Oracle Cloud Infrastructure (OCI), the new AI cost layer simply extends an existing workflow. There is no separate setup. No new platform to learn.
