Skip to content
worth noting Tools and apps

Amazon Bedrock introduces automatic cost allocation for AI tools by individual user

only one source so far

AWS Bedrock now allocates AI inference costs to individual IAM identities without code changes and links them to AWS Cost and Usage Reports. It also works for Claude Code, Claude Desktop, and OpenAI Codex, enabling per-user spending limits.

According to AWS, Amazon Bedrock offers a new cost allocation mechanism that requires no code instrumentation — every inference call is automatically tied to the IAM identity of the user who made it. This identity, together with organizational tags, propagates into AWS Cost and Usage Reports (CUR 2.0) and AWS Cost Explorer, while the data remains inside the AWS security perimeter. The mechanism is meant to answer questions such as: what is the actual cost per user, which teams generate the most spending, and how to set a budget based on headcount and adoption.

According to AWS, the approach works the same regardless of which AI tool makes the call — Claude Code, Claude Desktop, OpenAI Codex, ChatGPT Work, and OpenCode all communicate with Bedrock through one of two endpoints (bedrock-runtime or bedrock-mantle), and cost allocation operates at the level of this endpoint, beneath the logic of the specific tool. Identity and session tags thus propagate automatically even when the tool changes, and once an identity has been configured, it is inherited by all subsequent calls.

Deployment requires configuring AWS IAM Identity Center with an external identity provider (SAML 2.0 or SCIM), enabling SCIM provisioning, and mapping organizational attributes (team, cost center) to session tags, which then appear in CUR 2.0 with the iamPrincipal/ prefix and as filterable dimensions in Cost Explorer. CUR does not contain the cost of an individual request, only aggregation by principal and usage type within the export interval (hour or day); for detail at the level of an individual prompt, you need to use model invocation logs or request metadata tagging.

A different model applies to Claude Platform on AWS and Claude for Enterprise — there, usage by workspace and API key is visible in the Anthropic admin console, and spending limits are enforced directly by Anthropic, not by the AWS IAM mechanism. The source text is incomplete; further implementation details can be found in the source article.

What changed

Why it matters

For FinOps and platform teams, this is a tool for determining the actual cost per user or team without touching code, and for budgeting and setting spending limits for AI tools accordingly. For individual users of tools like Claude Code or OpenAI Codex, it means that in organizations with this configuration, their usage will be tracked by name and may be subject to a personal spending limit.

Two audiences, two different impacts

What this means

01

For individuals

Anyone using Claude Code, Claude Desktop, or OpenAI Codex within an organization running Amazon Bedrock may now have their usage tracked by name, and the company may set a personal spending limit for them.

What to do Find out whether your organization tracks per-user costs for the AI tools it uses on Bedrock and whether it has set spending limits.
More practical updates →
02

For a business

FinOps and platform teams gain visibility into AI inference costs by individual user, team, or cost center in AWS Cost and Usage Reports and Cost Explorer, without any application code changes, which the company says enables budgeting for AI tools and setting spending limits.

Processes
What to decide Consider implementing IAM Identity Center with session tags (team, cost center) to allocate AI tool costs to individual users on Amazon Bedrock.
More business impacts →
AWS Bedrock AWS Cost and Usage Reports Claude Code Cost attribution IAM OpenAI Codex

Check the original

Event sources

only one source so far · 1 publisher, 0 independent. We count feeds from the same owner only once.

1
AWS Machine Learning Blog primary source · first detected Cost attribution and controls for AI developer and productivity tools on Amazon Bedrock