Track generative AI costs with Amazon Bedrock inference profiles

Summary

When multiple departments share a single foundation model in Amazon Bedrock under one IAM role, usage aggregates into one billing line item, complicating department-level chargebacks and budget tracking. This issue can be resolved by deploying Amazon Bedrock application inference profiles wrapped around the shared foundation model and tagging each profile for cost allocation. The application layer determines the user department during authentication and passes the corresponding inference profile ARN as the model identifier. Once the designated cost allocation tag is activated in the AWS Billing and Cost Management console, usage separates into per-department line items in AWS Cost Explorer. This method incurs standard per-token model rates without extra charges for cost attribution.

Context

Multiple teams sharing a single foundation model under one IAM role generate aggregate billing that prevents finance from performing chargebacks, setting per-team budgets, or tracking individual department spend.

Approach / What changed

Create department-specific application inference profiles referencing the same foundation model, tag each profile for cost allocation, and configure the application to route user requests to the respective profile ARN.

Takeaways

• Using an application inference profile incurs standard per-token pricing with no additional charges for cost attribution.
• Cost allocation tags take 24 to 48 hours after activation in AWS Billing and Cost Management to display tagged costs in Cost Explorer.
• Deleting an application inference profile immediately breaks applications relying on its ARN, and recreated profiles receive new ARNs.

Related reading

• AWS · Jul 22, 2026

  Architecting offline-first generative AI applications for edge deployments using AWS services

  Industrial edge environments often face significant unplanned downtime yet lack reliable cloud connectivity to run large-scale artificial intelligence models. To address this limitation, an offline-first reference architecture shifts generative inference to edge devices while leveraging cloud services for preparation and lifecycle management. The workflow uses Amazon Bedrock with Amazon Nova Pro to create labeled dataset pairs, Amazon SageMaker AI Pipelines to fine-tune a compact model, and AWS IoT Greengrass to orchestrate edge deployments. Local execution relies on Ollama, Strands Agents, and a CPU-based ChromaDB retrieval pipeline that consumes zero GPU VRAM. In evaluations across three foundation model judges, the fine-tuned gpt-oss-20b model combined with retrieval-augmented generation consistently outperformed the base model configuration across accuracy, completeness, and relevance metrics.

  Ahmed Ewis

• AWS · Aug 26, 2026