Join our Newsletter — 33% off our NHI Course
Home› Guides› AWS European Sovereign Cloud Guide: Partition, Identity and…
Guide Cloud Security

AWS European Sovereign Cloud Guide: Partition, Identity and Assurance Design

PDF for newsletter subscribers · free to read online

← All guides
By NHI Mgmt Group, edited by Lalit Choda Updated 8 November 2026 12 min read
On this page

The AWS European Sovereign Cloud is not another AWS Region you can switch on from an existing account: it is a separate cloud with its own accounts, billing, identity and operations, and most of the security work is in designing for that separation. AWS announced general availability on 15 January 2026, with the first Region in Brandenburg, Germany. This guide explains what is actually separate, how identity and privileged access work when nothing is shared with your commercial AWS estate, how to carry logging, keys and machine identities across, what assurance AWS publishes, and the mistakes buyers make. It is for cloud architects, security leads and procurement teams assessing or adopting the service. Facts were checked against AWS sources on 27 October 2026; the service is new and changes quickly, so confirm details in the current AWS documentation. For the wider legal picture, see the Data Sovereignty and Residency Guide and the EU Sovereign Cloud Options Compared Guide.

This is an independent guide from NHI Mgmt Group, based on AWS's public documentation as of November 2026. It is not sponsored or reviewed by AWS. Product features change, so confirm settings against AWS's current documentation.

Key takeaways

  • New accounts, no interoperation. AWS states that commercial accounts cannot be used and that sovereign cloud accounts do not interoperate with other AWS Regions; plan a parallel landing zone, not an extension.
  • Identity is rebuilt, not reused. IAM, billing and metadata such as roles, permissions and configurations stay in the EU, which means federation, roles and machine identities must be created inside the new environment.
  • Operations stay in the EU. AWS says day-to-day operations, including data centre access and support, are run by EU residents in the EU, moving to EU citizens, with no operational access from outside the EU.
  • Assurance is published separately. AWS lists C5, ISO/IEC 27001, 27017, 27018, 27701, ISO 22301 and SOC 2 among assessed programmes, plus its own Sovereign Reference Framework (ESC-SRF); check which services are in scope for each.
  • Sovereignty is shared responsibility too. The provider's separation does not stop you exporting data, logs or keys to non-EU systems; your pipelines, SaaS tools and support processes decide much of the outcome.

How identity and access work in the AWS European Sovereign Cloud

What is separate

ElementWhat AWS statesDesign implication
InfrastructurePhysically and logically separate from other AWS Regions, located wholly in the EU, multiple Availability Zones, built on the Nitro SystemTreat it as a separate failure and trust domain
RegionRegion code eusc-de-east-1, named AWS European Sovereign Cloud (Germany); Sovereign Local Zones planned in Belgium, the Netherlands and PortugalOnly one Region today; plan resilience across its Availability Zones
AccountsNew account required; commercial accounts cannot be used; no interoperation with other RegionsSeparate account vending, organisation and guardrails
Identifiers and endpointsSign-in and ARN references use the aws-eusc prefix; service endpoints follow <service>.eusc-de-east-1.amazonaws.euPolicies, SDK configuration and IaC that hard-code arn:aws: or amazonaws.com break
Identity and metadataSovereign IAM; customer data and metadata (roles, permissions, resource labels, configurations) can stay in the EUNo shared IAM principals with commercial accounts
BillingOwn billing and usage metering, EUR experience; Marketplace needs a sovereign cloud accountSeparate cost governance and procurement path
OperationsEU-resident operators in the EU, transitioning to EU citizens; designed to keep operating if disconnected from the rest of the worldSupport cases and access requests are handled in the EU
ContractingEMEA-billed customers contract with Amazon Web Services EMEA SARL; the Region is operated by German entities including AWS European Sovereign Cloud GmbH; the ESC Addendum applies automaticallyLegal review should read the Addendum, not only the standard customer agreement

Identity architecture

Because accounts and IAM do not interoperate with your commercial estate, expect, as in AWS's other separate partitions, that roles in one cannot be assumed from the other and that organisation-level policies do not span both. Confirm the specifics in the service differences documentation for this Region before you design. Practical consequences:

  • Human access: federate workforce sign-in from your identity provider directly into the sovereign environment, with its own SAML or OIDC configuration, rather than hopping through a commercial account. If data residency is the driver, check where your identity provider itself stores sign-in logs and attributes.
  • Separate root users: every new account has its own root user; protect each with MFA and keep them out of daily use. See the AWS IAM Policies and Roles Guide.
  • Guardrails rebuilt: SCPs, permission boundaries and data perimeter policies written for the commercial partition need new ARNs and organisation identifiers. See the AWS Resource Control Policies and Data Perimeter Guide.

Admin roles and privileged access

Two kinds of privileged access matter: your own administrators, and the provider's operators.

  • Your administrators: apply the same multi-account pattern as elsewhere, with a management account used only for organisation tasks and a security tooling account holding delegated administration, but build it inside the sovereign environment. Use short-lived role sessions, not IAM users with keys. See the AWS Multi-Account Security Reference Architecture Guide.
  • Who may administer: if your sovereignty requirement covers your own staff (for example, only EU-located administrators), enforce it through identity provider conditions and device posture, since AWS controls apply to its operators, not to yours.
  • Provider operators: AWS states that operations are performed by EU residents located in the EU and that there is no operational access from outside the EU borders; in exceptional cases EU-resident authorised staff can access a replica of source code needed to maintain services. AWS also lists AWS KMS among services designed with zero operator access. Record these commitments in your risk assessment and map them to the ESC Addendum. See the Cloud Provider Support and Operator Access Controls Guide.
  • Break-glass: keep break-glass credentials for the sovereign organisation separate from commercial ones, stored and tested under the same controls.

Integrations, API tokens and service accounts

The separation is most visible for machine identities. Every pipeline, agent and SaaS tool that touches the sovereign environment needs its own identity there, and every one of them is a path data can take out of it.

IntegrationTypical commercial patternSovereign cloud approach
CI/CD deploymentOIDC federation from the CI platform into a commercial deploy roleSeparate OIDC provider and deploy roles in sovereign accounts; check where the CI platform and its runners operate
Security tooling (CSPM, SIEM, EDR)Vendor assumes a cross-account role from its own AWS accountThe vendor needs a presence in the sovereign environment, or you accept that findings and metadata leave it; confirm before buying
Log shippingOrganisation trail to a central log archiveA log archive inside the sovereign environment; forward to a SIEM only after deciding what may leave
Encryption keysAWS KMS keys in the workload RegionAWS KMS customer managed keys in eusc-de-east-1; check FIPS endpoint availability (for example kms-fips) if required
Marketplace softwareSubscribed from the commercial accountRequires a sovereign cloud account and the listing's availability there
Static access keysSometimes tolerated for legacy toolsAvoid: a key copied to a non-EU system defeats the residency design
  • Inventory every integration: before migration, list every NHI that touches the target workloads (roles, keys, OAuth apps, tokens) and decide for each whether its counterpart can run in the EU. See the Non-Human Identity (NHI) Guide.
  • Prefer federation: use OIDC or IAM Roles Anywhere style short-lived credentials over long-lived keys. See the AWS Access Keys Security Guide.
  • Update endpoints and ARNs: SDKs and tools need the Region code and the amazonaws.eu endpoints; hard-coded arn:aws: strings in policies and code are a common migration defect.

Logging and detection

AWS describes the sovereign environment as able to keep running with no connection to the rest of the world, so your detection stack should work the same way. Design logging so that evidence is collected, stored and analysed in the EU, and decide explicitly what may be exported.

  • Collect: CloudTrail management and data events, VPC flow logs, GuardDuty and posture findings in the sovereign organisation, aggregated into a log archive account there. Check service availability for each before relying on it. See the AWS CloudTrail Guide.
  • Analyse: if your SOC platform runs outside the EU, decide whether to forward full logs, filtered alerts, or nothing. Many organisations forward alerts with minimal personal data and keep raw logs in the EU.
  • Watch for leakage: alert on new cross-account trust policies naming principals outside your sovereign organisation, resource policies granting external access, S3 replication or exports to destinations outside the environment, and new access keys (MITRE ATT&CK T1537, Transfer Data to Cloud Account; T1098.001, Additional Cloud Credentials).

Metrics

  • Integration residency: share of integrations touching sovereign workloads whose counterpart operates in the EU.
  • Long-lived credentials: number of IAM users with access keys in sovereign accounts. Target zero outside exceptions.
  • Log containment: percentage of log sources stored in a sovereign log archive with retention set.
  • Export events: count of approved and unapproved data exports out of the environment per month.
  • Assurance coverage: share of services in use that are in scope of the C5, ISO and SOC 2 reports you rely on.

Hardening baseline

Assurance to collect

EvidenceWhat it showsCheck
C5 (BSI Cloud Computing Compliance Criteria Catalogue)German federal criteria for cloud securityServices in scope and report period
ISO/IEC 27001, 27017, 27018, 27701Information security, cloud controls, PII protection in cloud, privacy managementCertificate scope names the sovereign environment
ISO 22301, ISO 9001, ISO 20000Business continuity, quality, service managementRelevance to your resilience obligations
SOC 2Controls over security and related criteriaComplementary user entity controls you must operate
ESC-SRF (Sovereign Reference Framework)AWS criteria on governance independence, operational control, data residency and technical isolation, with a third-party validated reportRead the report itself; the framework is AWS's own
ESC AddendumContractual commitments specific to the sovereign cloudLegal review against your requirements

AWS's compliance page lists these programmes as assessed by an independent auditor (page last updated August 2026). In NIST CSF 2.0 terms, this work sits in GV.OC-03 (legal, regulatory and contractual requirements understood and managed) and GV.SC-07 (supplier risks assessed and monitored); in ISO/IEC 27001:2022 Annex A, in A.5.23 Information security for use of cloud services and A.5.34 Privacy and protection of PII.

Adoption in phases

  1. Decide why: write down the specific requirement (customer contract, sector regulator, public sector rules, internal policy) and what it requires: data location, operator location, metadata location, legal control, or continuity if disconnected. Different drivers need different evidence.
  2. Check service fit: compare the services you use against availability in eusc-de-east-1 and the service differences documentation, including FIPS endpoints and the security services your design depends on.
  3. Build the landing zone: organisation, management account, security tooling and log archive accounts, SCPs, identity federation and break-glass, all inside the sovereign environment.
  4. Rebuild integrations: CI/CD federation, security tools and support processes, each with an EU-operated counterpart or an approved exception.
  5. Pilot: migrate one non-critical workload, run detection and incident response exercises end to end in the EU, and test what happens when the commercial estate is unreachable.
  6. Roll back: keep the source workload running in the commercial environment until the pilot is accepted; rollback is pointing DNS and clients back. Plan data synchronisation so that rollback does not itself move regulated data out of the EU.

Decision guidance

  • Choose it when a contract or regulator requires EU operational control and EU-located metadata, not only EU data location, or when continuity during loss of connection to non-EU infrastructure matters.
  • A commercial EU Region may be enough when the requirement is data residency only and can be met with Region restrictions, customer managed keys and contractual terms. It is simpler and offers more Regions.
  • Small organisations should weigh the cost of running a second landing zone, toolchain and identity integration; a single sovereign workload often still needs the full foundation.
  • Large organisations should treat it as a separate platform with its own owners, budget and assurance cycle, and decide early which shared services (SOC, IAM, CI/CD) must be duplicated.

Common misconfigurations

  • Assuming an existing account works: teams plan to "enable the Region" in a commercial account and discover late that a new account, organisation and billing setup are required.
  • Exporting metadata through tooling: a CSPM or SIEM in a non-EU location reads configurations and logs, undoing the metadata residency the service provides.
  • Copying long-lived keys: a deploy key created in the sovereign account and stored in a CI system outside the EU gives that system standing access.
  • Hard-coded partition strings: policies with arn:aws: and endpoints with amazonaws.com fail, and the "fix" is sometimes a wildcard that broadens access.
  • Relying on marketing statements: procurement records "sovereign" without reading the ESC-SRF report, the ESC Addendum or the scope of each certificate.
  • Forgetting your own people: provider operators are EU-based, but your administrators, support partners and managed service providers may not be.

Practitioner checklist

  • The sovereignty requirement is written down with the specific evidence it needs.
  • Services, FIPS endpoints and security services needed are confirmed available in eusc-de-east-1.
  • A separate organisation, management, security tooling and log archive account structure exists in the sovereign environment.
  • Workforce federation, root user MFA and break-glass are set up independently of the commercial estate.
  • Every integration has an EU-operated counterpart or a documented exception; no long-lived access keys are exported.
  • Policies and code use the correct ARN prefix and endpoints; no wildcards were added during migration.
  • Logs are stored and analysed in the EU, with an explicit rule for what may be exported.
  • C5, ISO, SOC 2 and ESC-SRF reports are on file, with service scope checked, and the ESC Addendum has had legal review.

Standards and references

Related NHI Mgmt Group resources: Data Sovereignty and Residency Guide · EU Sovereign Cloud Options Compared Guide · AWS Security Guide · CLOUD Act Guide · Azure Government and Sovereign Cloud Guide

PDF for subscribers

Take this guide with you

Every section, table and checklist in one PDF, with links to the sources and related guides. Subscribe to our weekly NHI & AI Identity Journal and we’ll email you a download link.

Free weekly newsletter

Subscribe to the NHI & AI Identity Journal

The latest on NHI and Agentic AI security – articles, research, breaches, news and events every week.

Bonus 33% off our NHI Course when you subscribe.

NHIMG Editorial Note
Researched and written by the NHI Mgmt Group editorial team using AI tools, under the editorial direction of Lalit Choda. Facts are drawn from the sources listed on this page and checked by our automated quality process. Last updated 8 November 2026. Spotted an error? Use our contact page.
NHI Mgmt Group, the independent authority on Non-Human Identity, IAM and Agentic AI security. nhimg.org

PDF Downloads for Newsletter Subscribers

Get the guide

Enter your details and we’ll email you a download link.

The full guide is free to read online. The PDF is a thank you for subscribing.

Fill in all fields marked * and tick the PDF box to continue.