Stolen AI credentials are now showing up as attacker infrastructure

Anthropic's September report documents resale, victim-funded compute and attribution cover as separate values of compromised AI access.

A dated change to public knowledge. Follow the affected records for their current evidence.

Published

Sep 10, 2026

DiggingBeagle

The report

Anthropic's September threat report makes stolen AI access a first-class security asset.

The report describes fraudulent resellers harvesting model credentials, financially motivated operators switching attack workloads onto victim API keys, a hacktivist campaign running on stolen access and GTG-50020 taking production AI-provider keys after compromising an evaluation sandbox.

Anthropic summarizes the attacker value of those credentials as resale value, compute paid by the victim and attribution cover.

This changes how the existing DiggingBeagle cyber-operation cases should be read. AI credentials are not merely configuration secrets for a model API. In the reported campaigns they become transferable infrastructure for other offensive work.

Cite this record

DiggingBeagle. “Stolen AI credentials are now showing up as attacker infrastructure.” Published Sep 10, 2026. https://diggingbeagle.com/updates/stolen-ai-credentials-are-now-showing-up-as-attacker-infrastructure/

Citation guidance

The source matters after the headline fades.

DiggingBeagle is a non profit research project documenting AI security incidents, agent failures, vulnerabilities and AI-assisted operations. A case keeps its claims beside the sources that support, contest or limit them. Later updates stay visible, so a reader can see when the account changed.

We publish case reconstructions, dated reporting and analysis across records. Each has a different evidentiary role. About the project and our methodology explain how the work is reviewed.